The rise and fall of the standard user interface

In the early days of microcomputers, everyone just invented their own user interfaces, until an Apple-influenced IBM standard brought about harmony. Then, sadly, the world forgot. In 1981, the IBM PC arrived and legitimized microcomputers as business tools, not just home playthings. The PC largely created the industry that the …

[John Riddoch]

Efficient interface

The history of vi tracks from this. Once you get the hang of its obscure commands, vi is very powerful, allowing very quick editing with minimal keypresses. When you consider it developed from "ed" (the old single line text editor) run over a slow serial line, it makes sense. You didn't want to send complicated commands to the expensive server from your dumb terminal, because that took limited bandwidth on your serial line and used expensive CPU cycles.

Does vi follow any UI guidelines? No. Is it user friendly? No. Is it efficient once you get to know it? Yes. Given that it was the only editor guaranteed to be available on Solaris, AIX & HP-UX servers, it became my default editor to use when managing them. With Linux, you're far more likely to have nano or something available but I still just use vi/vim.

[jake]

Re: Efficient interface

"Is it user friendly?"

Yes. But it's very picky about who its friends are.

"it was the only editor guaranteed to be available"

No. That would be ed, the UNIX text editor. However, in middle-early days, vi was often the only screen editor available.

ed is still available on every serious example of a *nix. Handy in emergencies. One cannot be considered a true systems administrator unless one has a good working knowledge of ed.

[Anonymous Coward]

Re: Efficient interface

Yep, vi often depended on libraries in /usr. If you couldn't mount /usr then ed was your only real option for fixing things (other than "cat > /etc/rc0.d/S99hackmount" and reboot with crossed fingers, in extreme cases).

[Peter Gathercole]

Re: Efficient interface

Before the days of shared libraries and dynamic linking, vi was compiled as a fully static linked binary, and would work as long as you had access to the executable file, and a copy of termcap or terminfo.

But shared libraries with dynamic linking were seen as the way forward, so vi stopped working if /usr/lib was not available. I remember back in the day recovering AIX systems booted from three floppy disks while using nothing other than what would fit on a 1.44MB floppy, which only just contained a shell and ed, plus a handful of other commands.

I think that ed is still linked as a single fully static linked binary on real UNIX systems, although on RH Linux it appears to be dynamically linked.

[CrazyOldCatMan]

Re: Efficient interface

But shared libraries with dynamic linking were seen as the way forward, so vi stopped working if /usr/lib was not available

I vagely remember that the Makefile (or was it the configure script?) for vi had a --static option which would build a static binary. Or is that just on the FreeBSD ports tree?

[Anonymous Coward]

Re: Efficient interface

termcap and terminfo are often in /usr/lib or /usr/share anyway, so the same problem exists.

[munnoch]

Re: Efficient interface

Yep, SunOS in single user mode wouldn't mount /usr so it was ed or nothing.

Even with vi available if your TERM env var wasn't just right it was unusable and you might as well go back to ed.

Going further back I remember using ed on a teletype when all the green screens in the lab were in use.

I'm minimally functional in vi mainly because I can hit colon and do ed type things... obviously real work requires emacs (dons asbestos overalls...)

[Doctor Syntax]

Re: Efficient interface

"Even with vi available if your TERM env var wasn't just right "

You've raised another thing that came in alongside vi: termcap and, later, terminfo (was terminfo the result of a fit of NIH by Bell Labs?) so that, providing you had the correct description for it you could hang whatever ASCII terminal you liked on the end of your serial line. No need for an ADM3A - VT100 or VT220 were where it was at as far as I was concerned.

[Peter Gathercole]

Re: Efficient interface

If I remember correctly, termcap (and Curses) was effectively the code that allowed vi to function on different terminals ripped out to make a more generic terminal handling library using termcap and the libcurses library. I remember reading about this, and doing a bit of testing when I got a BSD 2.something release for Edition 6 and 7 (unfortunately on a PDP11/34, which was too small to run vi), obtained because it had the Ingres RDB system on it, which was wanted to teach Relational Database. to students.

I do not know whether terminfo was a NIH thing that AT&T did, but I do remember that it was more functional than terminfo, and because it used compiled, hierarchical files containing the essential features needed to drive a terminal, it loaded faster than parsing the large, single file of terminal descriptions that was termcap. I think I remember some comments in the shipped version of termcap from BSD which suggested that you create a cut-down version of the termcap file just containing the terminals that you had on your system for the sake of speed.

Part of the code shipped with AT&T UNIX releases included a captoinfo command which would take a termcap entry for a terminal, and create a terminfo source file for tic.

Another thing that BSD included was a shell called vsh, or the Visual Shell. This provided an environment a bit like Midnight Commander (itself heavily influenced by Norton Commander) on text only terminals. I think that vsh might have actually pre-dated Norton Commander, so it is possible that vsh actually influenced Norton Commander, rather than the other way round!

I'm sure the code for vsh is available in the tuhs archives, so it may be an interesting project to ressurect it to see how it compared.

[Michael Wojcik]

Re: Efficient interface

vi, at least Joy's original implementation, did not use curses, and in fact predates it. vi does use termcap (and later terminfo), but has its own output routines. Legend has it that Arnold took the curses code for optimizing cursor movement from vi.

[Liam Proven]

Re: Efficient interface

[Author here]

> VT100 or VT220 were where it was at as far as I was concerned.

I used them at university, but when I first deployed xNix in production, it was SCO Xenix and my employers' customers mostly used Wyse 60 and the like terminals, which were something like 1/4 of the price DEC terminals.

This was why the ICE.TEN and later ICE.TCP terminal emulation software for PCs was so important: it was more or less the first terminal emulator that emulated the real terminals real systems actually used, meaning Wyse kit, not the expensive stuff that only subsidised places like higher education could afford.

Where SCO sold up against early PC networks, or before that DR CDOS, it was on the basis of price:performance. An early 80386DX with 4MB of RAM could support 4 users on an single ordinary PC host, as opposed to giving them a £2000 PC each. Those customers bought Wyse because a Wyse terminal was £500:

https://terminals-wiki.org/wiki/index.php/Wyse_WY-60

Whereas a genuine VT was a lot more. E.g. the VT220 launched a few years before the Wyse and was over twice the price.

https://en.wikipedia.org/wiki/VT220

And that was a later, cheaper DEC terminal. I see Wikipedia claims they were much more successful than Wyse, but I think that is regional bias. Americans always were richer and could afford fancy expensive kit. That's why Apple sold so well Statesside and in Europe you almost never saw Apple kit, because local computers were 1/4 of the price or less.

[Peter Gathercole]

Re: Efficient interface

In the UK, Plessey sold a VT100 look-and-work-a-like which at first glance looked a lot like a VT10x, but was a lot cheaper. There were very common in UK education. DEC terminals, like most computer manufacturer's own terminals, were quite pricey. But this was partly because of the way they were constructed. The cheaper terminals were effectively programmed single board Intel 8085 computers, with the programming in ROM, whereas DEC VT52 and VT10X terminals, although they became microprocessor based, tended to be multi-board implementations in a backplane. This made sense for maintainability and upgradeability (things like the Advanced Video Option AVO, and Regis graphics, were optional plug-in boards to a base VT100), but not necessarily for price. Later DEC terminals like the price reduced VT102 and VT220 onward followed the single board model, to bring the price down (a bit).

I was never very fond of Wyse terminals, because although they were cheap and functional, I seem to remember them having a bad reputation for reliability (and their native command set was awkward to write termcap/terminfo entries for). Of course, for Wyse 75's and later terminals, you normally operated them in emulation mode, not native.

Prior to cheap PC's like the Amstrad 1512, unless someone had a PC on their desk for another reason, providing one just for a terminal emulator seemed a lot of money for a terminal replacement. Once PCs could be bought for less than a decent terminal, the serial terminal industry died a death. Wyse and some others tried to keep it going by diversifying to X-termnals, Winterms and Thin Clients, but these could not really compete either.

But it was not just IBM PC's that got terminal emulation software. I wrote my own 98% VT52 compatible terminal emulator (it was not timing-compatible, and only implemented XON/XOFF, not hardware flow control, but was otherwise complete) for the BBC micro before the IBM PC was common, (and I also did a minimal Tek4010 emulation as well) when I worked at Newcastle Polytechnic. I stopped short of doing an ANSI/VT100 emulation, partly because commercial emulations came along, such as Computer Concepts Communicator, and Acornsoft Termulator, and partly because it would be much more complex. And of course Kermit was freely available.

I did sketch out an MSc proposal for myself to implement a WIMP managed serial terminal on the BBC micro using an AMS mouse, but I realised things were moving so fast that it would have been obsoleted by commercial offerings before I finished it!

[Peter Gathercole]

Re: Efficient interface

Hmm. It seems that 8085 was not as common as I thought. 8051 and the later variants were used by Wyse, and probably other manufacturers as well.

It just happened that I saw 8085 in a couple of terminals I saw in bits in the '80s, so I assumed this was the common processor.

[JoeCool]

Re: Efficient interface

uh huh. ed (or ex mode) is also the answer to "this shared modem pool connection over is taking 60 seconds or more to update my vi screen. I need a 2400 baud upgrade."

[Doctor Syntax]

Re: Efficient interface

"Yes. But it's very picky about who its friends are."

I'd express is more along the lines of "you've got to make friends with it".

[Liam Proven]

Re: Efficient interface

[Author here]

> That would be ed, the UNIX text editor.

Got to say, I am with the person to whom you're replying here.

Vi is on virtually every xNix box. Yes, we know the gag about "ed is the standard editor" but we're talking about real machines those of us still working for a living might have to maintain: mid-to-late 1980s and 1990s xNix, not the very early minicomputer versions of the 1970s. I am an olde pharte of 56 who was working for a living at the end of the 1980s, and I didn't maintain minicomputer xNix because when those boxes were current, I was in primary school busy learning to read and write.

[CrazyOldCatMan]

Re: Efficient interface

I am an olde pharte of 56 who was working for a living at the end of the 1980s

Ah - a mere youth then - almost 3 years younger than me!

My first job that actively used a computer was CAD - drawing location cases for railway signalling kit. Got bored with that, moved down to Plymouth to get married.. Then we both got jobs at the new airline CRS in Swindon and became TPF programmers (I had a spectacularly bad interview, partly because the interviewer had a really strong Swiss/German accent and, being really nervous, I had difficulty understanding him. They ended up offering my wife a job but not me. Her response was to say "well, I'm not moving to Swindon if he doesn't have a job" so they eventually offered me a job as well :-)

My first unix/linux was Slackware SLS 0.99pl15 (I think) - I had a friend who worked somewhere with Internet access (1993-4? Somewhere around there) and could download (and re-download when the floppies failed as happened with every 3rd or 4th floppy) it for me. This was on a 386sx25 with 4mb? RAM (tiny amount anyway) and, initally an 80MB hard drive, connected via an SB16 sound card..

Can't remember why I did it - probably because I was bored at work. I was with Demon at the time and they had really good FAQs on how to do stuff like dial-on-demand and fetchmail

So I guess my vi knowledge starts then - I certainly would count myself an expert - I know how to do what I need to do (add lines, delete lines, replace text, search and replace etc etc). I've never knowingly used emacs.. I do vaguely remember ed (from my auto-borking of my linux box) but I can't remember any of the keystrokes.

By the time I officially had a job herding unix boxes (1999 - post y2k projects - involved looking after the packet-switched networks, administering the BT Meridian PABX [1], doing user admin with NIS, building new servers and workstations etc etc - they initally gave me a Windows PC but, with my manager permission I dual-booted it into linux (can't remember which flavour)) I had 5-6 years of doing stuff on linux and, fortunately, managed to do enough reading in the days before my interview to talk like I knew what I was doing :-)

Sadly, I don't herd linux at work (we have very little of it) but do use MacOS as my main work machine and herd the 100 or so Macs that we have (and do a bunch of stuff on the Windows side as well).

[CRConrad]

Heh, youngsters nowadays...

Ah - a mere youth then - almost 3 years younger than me!

Bah, you whippersnapper, the kid is almost

four years younger than me.

I only got into vi around a decade ago, PuTTYing into our DB / ETL servers at work to run, modify, and in the end create the shell scripts that ran our jobs. Think I saw ed in use for real once, in about 2001-02, over the shoulder of a coworker as I assisted him in installing Oracle on AIX at a client's facilities. Could be misremembering, could have been vi.

[Peter Gathercole]

Re: Heh, youngsters nowadays...

I really don't want to say how much younger than me he is, but I was using video terminals on UNIX systems (and also MTS via PDP11 terminal concentrators) in 1978 at Uni. Back then they were really glass-teletypes, although we did have Queen Mary's Editor for Mortals (em) working on UNIX V6/V7 on Newbury pretty dumb terminals that allowed you to open a line and do insert/delete processing using cursor movement on a line-by-line basis. But I still remember how to use ed even today.

But I know that there even older people than me knocking around here.

[4mc]

Re: Heh, youngsters nowadays...

I wrote the main SAA REXX Article for the UK PC User Magazine back in August 1987 while Jerry Sanders was still editor. I was hired by IBM UK a couple of months later, I'm 67.

I started as a relational database in 1974, sorting punch cards using rows and columns to produce a result set based on an sql-like query on a piece of paper.

I started to learn to program a few months later when running programmer compiles overnight on the IBM 360/40 while nothing else was running. I would check the compile output and correct obvious errors on the punch cards and resubmit, writing notes on the listings produced by the compiler.

In June 1997 I taught the first ever Java programming class in the UK. Funny old world. Old?

++Mark.

https://cathcam.wordpress.com

[doublelayer]

Re: Efficient interface

That claim requires some evidence. I'm sure you can use vi quite efficiently, but is it automatically the case that it's more efficient than an experienced user in a different editor? I think that depends a lot on what the user is trying to edit with it.

I know the basics of using vi for systems that have nothing else, but I came along when text editors could do more things and so I've never taken the time to use vi instinctively rather than deliberately. It's perhaps unsurprising that I find it to be slower to use than editors I've spent lots of time in. However, I also think that may be true if people find certain features to be useful. A typical GUI editor tends to have more builtin features than vi does, partially because there are other programs that can be used to get the same effect on the CLI. The more of those the user finds useful, the faster the editor with them included will be to them.

[Peter Gathercole]

Re: Efficient interface

The difference here is that with vi, you tend to look for ways to do things globally often using regular expressions, so for example global find and replace is trivial in vi if you know how, but doing it from a GUI is often much more convoluted.

Maybe this is because I still think in terms of ed commands myself, In GUI editors, I've seen people doing a find string, move the cursor within the string, delete, type the replacement and repeat, even when the GUI actually does allow global find and replace, but just makes it more difficult to apply. In vi, you can even do a "repeat the last complex operation", rather akin to an automatic macro facility that remembers the last operation you did.

And vi (original vi, not any of the re-implementations) has many quite obscure features like tagging, block moves, multiple named buffers for text moves. It is much more than the simple move a cursor, make a change.

And, even though back in the 16-bit minicomputer days vi was too big to run on some systems (PDP11s without separate I&D like the 11/34 and 11/40), in this day, it's absolutely tiny compared to most things on other OS's.

If you run the version of vi that came with BSD, rather than vim or any of the other implementations, you're using what was effectively a prototype. Anecdotally, Bill Joy lost the code for a later version he was working on and had to include his working copy in the BSD releases.. I often wonder how much better than the version we use that lost version would have been.

[Doctor Syntax]

Re: Efficient interface

"I often wonder how much better than the version we use that lost version would have been."

Maybe it would have been a victim of Brooks' second system effect.

[doublelayer]

Re: Efficient interface

That's actually a feature I value a lot, but I value it in Notepad++. One of the major reasons I like that editor when I'm on Windows is that it has a good regex system in its find and replace system. The GUI is still helpful, but I like those facilities as well.

[Phil O'Sophical]

Re: Efficient interface

I often wonder how much better than the version we use that lost version would have been.

Allegedly it had cut & paste.

[jake]

Re: Efficient interface

You mean cut buffers?

[Phil O'Sophical]

Re: Efficient interface

No, it has those. I vaguely remembered it having better copy/cut/paste, but perhaps I'm confusing it with something else.

There's an interview with Bill Joy from 1984 that details what was lost - multiple windows mainly. http://xahlee.info/comp/interview_with_bill_joy.html "vi would have multiple windows, and I might have put in some programmability"

[jake]

Re: Efficient interface

Here's the original interview, as printed in Unix Review in August of 1984:

https://archive.org/details/Unix_Review_1984_Aug.pdf/page/n59/mode/1up

I actually called Bill after that one to argue with a couple of his opinions. 40 years ago. Yeesh.

Of interest to this entire batch of commentardery is the article by Bill Tuthlll on screen handling from C programs, which directly follows the Joy interview. Worth a read, helps put some of this stuff into perspective. The advertising in the issue might open a few eyes, too.

Thanks for the memories. My round, I think.

[jake]

Re: Efficient interface

"perhaps I'm confusing it with something else."

Probably. Everybody and their mother had their own prototype screen editor back then.

[katrinab]

Re: Efficient interface

The difference here is that with vi, you tend to look for ways to do things globally often using regular expressions, so for example global find and replace is trivial in vi if you know how, but doing it from a GUI is often much more convoluted.

You can do that pretty efficiently on vscode.

I don't know if it is quicker than vi, but it is certainly a lot easier to discover how to do it.

[CRConrad]

Efficient interface or efficient user?

I came along when text editors could do more things

I kind of doubt that. My guess is, vi even at the time could do at least as much as any editor you were familiar with; it was just that you didn't know how to do that in vi.

[bitwise]

Walk up and use software

It used to be very easy to learn new windows software

Hold down alt to see the shortcuts to get into the menus, go into the first menu, navigate and see all the options.

Hold down alt, and mouse over all the icons and see the shortcuts.

As well as this, the menus listed the keyboard shortcuts and corresponding icons, so you could build a mental model of what is where.

[CRConrad]

Re: Walk up and use software

As well as this, the menus listed the keyboard shortcuts and corresponding icons, so you could build a mental model of what is where.

AFAICR icons started showing up in menus only with Windows 95 / NT 4, and only got widespread in the XP era. But floating tooltips on toolbar icons, with a bit of luck including the keystroke (and not requiring Alt to be held down IIRC) were pretty much everywhere from 1995.

[CrazyOldCatMan]

Re: Efficient interface

The history of vi tracks from this. Once you get the hang of its obscure commands, vi is very powerful, allowing very quick editing with minimal keypresses

The first 'proper' editor I ever really used was xedit on the IBM mainframe which was similarly obscure but powerful (a metaphor for the whole VM/CMS itself and TPF - the OS I was programming for). Designed for use on 3270 terminals that didn't have the concept of interactive screen updates (essentially just a list of screen updates got sent so, in those days of scarce bandwidth and CPU/RAM, you wanted to send as little info as possible which made short commands important).The whole screen got repainted every time there was an update although only the changed fields needed transmitting to the terminal.

We didn't have physical terminals - we ran the IBM 3270 emulator on our PS/2 50/z PCs.. (and not a lot of people realised that they allowed a DOS shell in the background or that they could run quite happily under Desqview..) I still remember that the right-hand CTRL key was repurposed as the enter key so that people who were used to the layout of the hardware terminals could use their muscle memory.

Also the mainframe had Rexx for doing scripting - not that we used it much.

[Bebu]

First time I have heard of the CUA for decades

The CUA guide was included in documentation shipped with the Window 2 SDK. Probably the most useful part of the whole box load. I think it was actually an IBM publication. I kept that guide for years after discarding the rest. Windows 2 was pretty dismal and even with the Whitewater Group's Actor language and development environment Windows 2 was just too restrictive and difficult to develop anything useful.

[bombastic bob]

Re: First time I have heard of the CUA for decades

I still have the book. It came with the Windows 3.0 SDK.

Article mentions how nobody liked Windows 1 and 2.x as they were considered 'ugly', and of course 3.0's OS/2-like (CUA) appearance became VERY popular.

So, then, WHY did Micros~1 *ABANDON* something NICE for a FLATSO 1.0-like appearance,,,??? (i.e. TIFKAM/Ape/Win-10-nic/1.1)

Micros~1 should have STUCK WITH THE CUA. You know, like Mate and Cinnamon (etc.)

[Someone Else]

Re: First time I have heard of the CUA for decades

So, then, WHY did Micros~1 *ABANDON* something NICE for a FLATSO 1.0-like appearance,,,??? (i.e. TIFKAM/Ape/Win-10-nic/1.1)

That's easy! By that time, Micros~1's marketing department (the tail that wags that dog) was slowly (or maybe not so slowly) being infiltrated by a crop pf new, young Millennial marketing bucks. One thing about Millenials (and Zots, too -- but I digress) is that they absolutely believe that anything created before they were born is of no importance¹. And since anyone fresh out of school (especially Marketing school) knows everything, they felt free to redesign everything in their own image. Couple that with a corresponding crop of new, young Millennial dev bucks who couldn't spell 'skeuomorphic interface' with a crayon, much less implement one, and you have a perfect (shit)storm of indifference and incompetence resulting in FLATSO.

_{¹ I've actually had Millennials and Zots tell me this to my face; you can't make this stuff up.}

[Dave559]

Re: First time I have heard of the CUA for decades

"anything created before they were born is of no importance"

To be honest, I'm sure a lot of us thought a lot of fairly similar things when we were also that same age…

The trick is for any bunch of overly-exuberant recent grads to be supervised and mentored by older, wiser (but not yet wearier) professionals who can help to remind them that, while new ideas are certainly welcome, sometimes there are very good reasons for certain things to have evolved to be the way they are, and that often only minor changes, if any, are necessary.

[Liam Proven]

Re: First time I have heard of the CUA for decades

[Author here]

> To be honest, I'm sure a lot of us thought a lot of fairly similar things when we were also that same age…

Douglas Adams:

«

Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.

Anything that’s invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.

Anything invented after you’re thirty-five is against the natural order of things.

Apply this list to movies, rock music, word processors and mobile phones to work out how old you are.

»

‘How to Stop Worrying and Learn to Love the Internet’, 1999

[Michael Wojcik]

Re: First time I have heard of the CUA for decades

Personally, I rather prefer a lot of contemporary music to what was being released when I was in my teens. Movies then were mostly pretty dreadful too; yes, they're mostly pretty dreadful now, but (assuming we include direct-to-streaming ones, and the readier availability of imports) there are a lot more of them, so there's a greater absolute quantity of decent ones.

As far as I'm concerned, the entire concept of the "word processor" is a mistake. WordPerfect 5.1 was probably the least wrong of the bunch, but it's still the Wrong Idea.

Mobile phones, I feel, peaked when I was in my 40's; but while today's devices are a pain in the ass to use (because of a lack of physical buttons and the ubiquity of idiotic touchscreens), I do find it convenient that I can read a book on mine — saves me carrying a paperback everywhere I go — and once in a while a feature such as Google Maps is useful.

[disgruntled yank]

Re: First time I have heard of the CUA for decades

@Someone else

< "believe that anything created before they were born is of no importance."

Boomer here, but it sure sounds as if you're talkin' 'bout my generation.

Well, actually, we probably pushed "before they were born" up to "before we hit puberty."

[Fruit and Nutcase]

Re: First time I have heard of the CUA for decades

I still have the book. Which, in my case, it came with the Microsoft OS/2 1.x SDK

[ldo]

Re: First time I have heard of the CUA for decades

The reason for GUI churn is, I think very simple: GUIs don’t scale.

A computer system would start out with some unified GUI design, with carefully-thought-out conventions catering for the tasks that were commonly performed at the time the product came out.

Then new applications would develop, pushing the boundaries of those GUI conventions. To cater for all the functionality, you would need menus upon menus, dialogs upon dialogs, widgets upon widgets and so on.

So at some point you have to redesign everything to cater for the new normal: the new range of tasks that are now considered to be commonly performed. But this annoys users who are accustomed to finding things in the old places, having to look in new places for the same old functionality. So maybe the GUI designers have to compromise on the purity of the redesign, to cater to some extent to backward compatibility. Which means having to decide where to let in these compromises, and where not to.

And so it goes.

[Liam Proven]

Re: First time I have heard of the CUA for decades

[Author here]

> GUIs don’t scale.

I think that is totally false and bordering on absurd.

What is happening now is that mobile phone and tablet UIs are being imported to desktop UIs and it doesn't work well. The original designs are now 40 years old and are used by hundreds of millions of people with great efficiency, and the problem is that people who don't really know how to use a keyboard and mouse effectively and only know finger gestures can't drive them, so the powerful aspects are being stripped out leaving a crippled shell that lacks much of the functionality.

[Doctor Syntax]

Re: First time I have heard of the CUA for decades

What's even worse, there is little agreement on mobile apps as to where functionality is to be found. One might have some functionality under a menu at the top of the screen, another might have something similar behind an icon at the bottom.

A pox on the lot of them for bringing their own chaos to the desktop.

[Michael Wojcik]

Re: First time I have heard of the CUA for decades

And where there is agreement, it has often been ruined by designers. My favorite example is that slider control used for binary options in settings. It's an astoundingly terrible way to represent an on/off state. Is left on or off? Checkboxes are unambiguous and intuitive. The slider control could only have been conceived of by some febrile, self-indulgent fool of a designer determined to change things for the sake of change.

[ldo]

Re: I think that is totally false and bordering on absurd.

Amusing, isn’t it? First you strenuously deny, even try to ridicule, my point, then go on to admit its validity by detailing how entirely different applications required entirely different GUI designs, and attempts to adapt existing ones did not work well.

[Nick Ryan]

Re: I think that is totally false and bordering on absurd.

It's analogous to the interface in a car.

There is a wheel in front of the user that steers the car. There are three pedals, arranged in the same order of Clutch (for non-automatic gearbox vehicles), the Brake and the Accelerator. Depending on where the steering wheel is in the vehicle there will be a hand brake either to the left or to the right of the driver.

These are all standard and an individual that drives one vehicle can quite reasonably expect to be able to get into another vehicle and drive it. Very specialised vehicles will stray from these standards but usually as little as possible.

Back to computer user interfaces... we have inexperienced, clueless user interfaces developers who are pathologically unable to understand that just because the came up with a hair brained re-arrangement of standard controls and through doing this are used to them and therefore declare this "better" that they have to trash establish standards and make computers harder to use for everyone else. If these lackwits were to design a car the steering wheel would likely be replaced by a side-to-side leaver hidden under the seat, one pedal would make the car go backwards and the other would make the car go forwards but only when both were pressed at the same time. The braking system would be replaced by a unified brake wheel moved into the door of the vehicle where once a certain, impossible to determine rotation spot was reached, would apply the hand brake instead of the normal brakes.

[ldo]

Re: It's analogous to the interface in a car.

You mean, the one where you need multiple lessons and to pass a test to get a licence before you’re allowed to use it in outside of a tightly-controlled environment? The one where, if you make any kind of mistake using it, the results could lead to serious injury or death, for yourself or others?

[Doctor Syntax]

Re: First time I have heard of the CUA for decades

"A computer system would start out with some unified GUI design, with carefully-thought-out conventions catering for the tasks that were commonly performed at the time the product came out."

We're dealing not with a specific design but with design principles. I'm sitting here in front of two windows, browser and mail/news opened by Seamonkey. The entire application is based on code for Firefox, as is the Thunderbird email/news client.

SM sticks to the CUA guidelines. It has the expected menu bar descended from what you'd have encountered on a text-based terminal application built to CUA guidelines. But the contents of those menu bars are somewhat different. They both have the File, Edit, and Help options you'ld find on pretty well any such application, View and Tools found on quite a lot and Window found on applications capable of displaying multiple windows. They both have Go menu options which reflect specific needs of what they do. Between Go and Tools they have window specific options - Bookmarks on one, Message and Events & Tasks on the other. If I select a drop-down menu of any of the apparently common options I get a selection which I get s somewhat different selection of new options - some are common to both windows some are not. Only the Windows option shows the same entries although event there a different box is shown ticked.

If I open a LibreOffice tool I see the File, Edit, View, Tools, Window and Help options and, between View and Tools, a set of options relating to that particular office function.

On all these different windows, despite the differences I'll expect quite a few commonalities. Under File, for instance, I'll find whatever New, Open, Close, Save, Print and Quit functionality that applies to the particular window, that despite SM & LO being quite unrelated to each other. I can open Pinta and again find File, Edit, View, Window and Help along with some image editing related options between View and Window. I can run through the KDE-related tool set and find much the same thing, It doesn't matter what the application is, even if I'm unfamiliar with it, I can start to find by way around because the corresponding areas of functionality will lurk under File etc. as they do under the menu bar of the old friends. The guide lines do not limit the functionality, they provide a level of agreement between different developers and all users as to where similar things are to be found.

Now I open Firefox or Thunderbird. Where is that familiar menu bar? All the basic functionality of the appropriate SeaMonkey window is there - after all SM uses the same underpinnings - but I have to go and hunt for it simply because the developers decided to do their own thing. There is now no agreement, there is chaos.

[keithpeter]

Re: First time I have heard of the CUA for decades

"Now I open Firefox or Thunderbird. Where is that familiar menu bar? "

Press ALT.

If you want to keep the menu bar displayed then right click over it and tick the box.

But I take your main point. Chrome / Chromium is a good example of the absence of a menu bar.

[Martin-73]

Re: First time I have heard of the CUA for decades

Never understood why the chrome developers were incapable of developing a menu bar

[steelpillow]

Yes, but

Another great article, Liam. Just a couple of thoughts:

Back in the 8-bit text-based era, there should be room to mention the CP/M operating system which arrived in 1974. It was the one MS-DOS had to beat when it eventually came along.

Honourable mention must also go to the Amstrad PCW series, which came with CP/M bundled alongside the Locoscript killer app running under its own OS. Bypassing the expensive IBM PC, it brought next-generation 3" floppy disks, cheap printing and proper long lines of text to the mass market. And a decidedly nonstandard keyboard, with many extra keys for common wordprocessor functions. But still usable for those other office apps that ran under CP/M.

"The first version of Ubuntu was released in 2004, arguably the first decent desktop Linux distro that was free of charge"

Certainly not for me. I tried it for all of five minutes before retreating back to Debian. Tried the latest a couple of years later, same result. Both shared GNOME2, which was a wise decision. I still use its MATE fork today (on the Devuan fork of guess who).

"the Linux desktop lacks diversity of design"

A lot of that is down to Darwinian evolution. Motif never caught on, though it was familiar from Solaris and other UNIX systems. KDE was always GNOME's great rival, each jockeying for its own quirks to push it ahead of the other's. ROX - RISC OS On Linux - was another bold initiative that left an indelible mark on the scene before falling away, though few recognise its influence today. Then again, how can anybody suggest that the hideously unmanageable GNOME3 is remotely in the same ballpark as its predecessor. And don't forget that Android is a Linux desktop too. That the modern Linux desktop still resembles CUA is a testament to those original designers, to the "millions of dollars and years of R&D" which followed the same evolutionary imperative.

[John Brown (no body)]

Re: Yes, but

"killer app"

Speaking of killer apps (and I'd also like to upvote the article!), Lotus 1-2-3 became the killer app on DOS after following in the footsteps of SuperCalc, which, speaking of CP/M, was born there and ported to MS-DOS a year before anyone had even heard of Lotus 1-2-3. Ditto with WordStar being the killer app before WordPerfect, although to be fair, WS did get an honourable mention in the article.

I never did get the hang of Lotus 1-2-3 having come from a CP/M background where SuperCalc was the king of spreadsheets and initially it was the only useful one available for MS-DOS and used all the same familiar commands.

[IvyKing]

Re: Yes, but

Supercalc? Lotus 1-2-3? Not to mention Viscalc. All pale in comparison with the mighty Multiplan!

Yes, I did have my tongue in cheek when typing the above.

[John Brown (no body)]

Re: Yes, but

Did VisiCalc make it to DOS? It was the first spreadsheet I used, on a TRS-80 no less! :-)

[Jou (Mxyzptlk)]

Re: Yes, but

> Did VisiCalc make it to DOS?

Yes, even before making it to apple.

[John Brown (no body)]

Re: Yes, but

Careful what you say when referring to Apple :-)

From your link "It was originally created by Dan Bricklin and Bob Frankston of Software Arts and first released in 1979 for the Apple II....It was one of the first commercial packages available when the IBM PC shipped in 1981.

[Elongated Muskrat]

Re: Yes, but

Amstrad's choice of 3" disks (not 3.5" ones) was a bit odd, though. They had a lower capacity than a 1.44Mb 3.5" disk (IIRC, 180K per side), and the drives were single-sided, so you had to flip the disk to read the other side. They certainly weren't "next gen", because the next gen (after 5.25" floppy disks) was definitely the 3.5" disk, the 3" disk ended up being the Betamax of the removable media world.

Amstrad used them because they were cheap (the same reason why the printer port in the CPCs was 7-bit, because the 7-bit controller chip was cheaper than an 8-bit one), and because Alan Sugar was (and still is) essentially just a spiv.

[steelpillow]

Re: Yes, but

At the time, the 3" and 3.5" were emerging competitors for the next-gen market and both had similar dismal capacity to start with. There were also early concerns about the reliability of the 3.5", as the 3" had a more robust cassette.

Yes, absolutely about Amstrad. They just launched with the drive they could get hold of cheaply and in enough volume. But in the end, the 3.5" proved reliable enough, the high-quality 3" cassettes lost out on cost per disk, and the volume market made its decision. Alan Sugar, you're fired!

[John Brown (no body)]

Re: Yes, but

"They had a lower capacity than a 1.44Mb 3.5" disk (IIRC, 180K per side), "

The Amstrad PCW came out in 1985, two years before the 1.44MB 3.5" floppy. It was, however, contemporaneous with the 720KB 3.5" floppy disks. By 1987, when the 1.44MB 3.5" disks arrived, Amstrad had updated to double density, 360KB per side or 720KB per disk. A few years later, they switched to 3.5" disks on the PCW, probably because they had now became cheaper than the now obsolete 3" drives. Bearing in mind it was built to a price, I wonder if an 8-bit Z80 CPU could manage with an HD 3.5" 1.44MB floppy or if they simply didn't see the point in changing again. After all, the target market wasn't power users :-)

Sometimes people just don't remember (or weren't there to know) just how fast computers were evolving back then and the various routes hardware makers went down only to find they chose the wrong route. There was no real reason those 3" drives could not have been double sided, they just got beat out by the Matsushita 3.5" format. It could have been one of the other competing formats that were also aiming to out compete and out perform the 5.25" floppies that had reigned supreme. At least we never standardised on the IBM 4" floppy!

[CrazyOldCatMan]

Re: Yes, but

Motif never caught on, though it was familiar from Solaris and other UNIX systems

To which me response, as someone that had to use it in work is "good". It was very much and MVP desktop/framework..

I think my first linux desktop was KDE on Mandrake/Mandriva (before it disappeared up its own backside).

Speaking of CUA, it was endlessly amusing (well - for about 5 seconds) that IBMs' flagship desktop application (for a while anyway) Lotus Notes was *utterly* non-CUA. CUA defines F5 as "Refresh", Bloats defines it as "force end user to reauthenticate". Yes, I know it was something that they bought in and they didn't want to alienate existing users but surely it wouldn't have been too difficult to implement a CUA-Mode switch? It seemed to have switches for pretty much everything else..

I only used/ran Notes for a couple of years and that was a couple of years too long.

[Neil Barnes]

Mint, which methodically put that UI back

Which is the major reason I use Mint.

It's not that it's intrinsically better or worse than other interfaces (for some value of better or worse) but that it's better for _me_.

(Though there are still some holes in the Mint world, particularly around the operation of menu bars).

[Neil Barnes]

Re: Mint, which methodically put that UI back

* too late to edit: I meant scroll bars.

[Doctor Syntax]

Re: Mint, which methodically put that UI back

That's Gnome/GTK's constant attempts to fix what wasn't broken.

[Mage]

Re: Mint, which methodically put that UI back

Mint + Mate is excellent and unlike Win10 and Win11 the themes etc can genuinely be changed.

[Liam Proven]

Re: Mint, which methodically put that UI back

[Author here]

> Mint + Mate is excellent and unlike Win10 and Win11 the themes etc can genuinely be changed.

The problem is that like Win11, Mate's taskbar does not work in vertical orientation.

The GNOME 2 team did not fully and properly understand the UI they were copying, and so their copy doesn't do all the things the original did. Mate merely keeps that broken design alive.

[John Brown (no body)]

Re: Mint, which methodically put that UI back

"(Though there are still some holes in the Mint world, particularly around the operation of menu bars)."

Mint with holes are usually called Polos and don't normally feature menus never mind bars although they do come in [sc]rolls :-)

[Dan 55]

Motif?

A great article but it was a shame the one of the first Unix widget toolkits, Motif, was missed out. Sun's CDE used Motif and Windows 3.0 based more than a passing resemblance to both.

Also, I approve of the retro tech week lasting more than a week.

[Phil O'Sophical]

Re: Motif?

Yes, the article missed out much of the 1970s and early 80s changes.

One of the very useful UI changes from the 70s was the standardisation of ANSI escape sequences, for cursor control and screen positioning. The archetypal embodiment of those is the DEC VT100 from 1978, an update to the previous VT52, and almost every terminal and terminal emulator today still has a "VT100 mode". Even early DOS-based PC terminal emulators offered that for cursor positioning and basic line-oriented graphics.

Before CDE came along, Sun had introduced the NeWS network-capable windowing system based on Display PostScript in the late 80s, and it was MIT's Project Athena that gave birth to X-Windows in the early 80s. I had a Sun desktop that had both NeWS and X installed (adding X required another 100MB of disk space, so it was usually NFS-mounted from a server!). X-Windows had (still has) the advantage of being usable even on slow connections, although initial display takes a while over a 9600bps modem once the window is open it can be remarkably responsive.

Sun developed OPEN LOOK as a UI standard, while other competing industry players designed Motif which was the more successful of the two. It was adopted by DEC for DECWindows, and many of the elements from it will be, as you say, familiar to anyone using today's MS Windows. Unfortunately that also seems to have led to a proliferation of UI designs whose main aim is NOT to be like Windows, which hardly helps simplification & standardisation. Most of the current Linux window managers are derivatives (at least in look and feel) of those early kits.

If someone's not even aware there is a standard, then the tools they build won't follow it.

I think https://xkcd.com/927/ is the appropriate comment here...

As for "But its influence never reached one part of the software world: the Linux (and Unix) shell", I think that's a little unfair. The shell was never intended to be much of a UI component, it was primarily a way to let users run programs and (if required) chain them together. UI design, if it was considered at all, was assumed to be the under remit of the program designer. The addition of getopts to the Unix standard definition did add some degree of standardisation to shells, but they were never really intended to be the sort of rich command interface provided by things like DEC's DCL. That too has changed, with bash, tcsh, csh, fish, zsh, ksh, etc. See the XKCD above...

[Doctor Syntax]

Re: Motif?

"Unfortunately that also seems to have led to a proliferation of UI designs whose main aim is NOT to be like Windows"

Including, of course, Windows where the object of every iteration is not to be like any previous version.

[Zack Mollusc]

Re: Motif?

That is not actially an objective, it is just the by-product of each new Microsoft Windows version being coded by programmers who have never seen previous versions of Windows. Or, indeed, any operating systems.

[DJV]

Re: Motif?

Or, indeed, code...

I am beginning to suspect it's just monkeys and typewriters all the way down...

[Doctor Syntax]

Re: Motif?

I can't accept that excuse. They must start with the previous version as their workbench. They have to have seen it. It's sheer wilfulness at work, like any badly behaved toddler.

[CrazyOldCatMan]

Re: Motif?

Including, of course, Windows where the object of every iteration is not to be like any previous version

If it wasn't the Marketdroids would be able to splash the "ALL NEW" crap over it. And the list of "new exciting features" would be a lot shorter and, as we all know, marketing is judged by verbosity, not ability..

[LionelB]

Re: Motif?

"Unfortunately that also seems to have led to a proliferation of UI designs whose main aim is NOT to be like Windows, ..."

There are, however, popular UIs like Xfce and Mate, which arguably are trying to be like the better iterations of the Windows UI (let's say Windows 2000, Windows 7).

"... which hardly helps simplification & standardisation."

Well, at least if you feel that simplification is always a good thing - and if you feel that Windows is worthy of setting the standard.

[Doctor Syntax]

Re: Motif?

The 95 - W2K era did bring together a lot of good ideas that had been around in the previous few years (given the previous art they built on Microsoft were just taking the piss in patenting any of it). W2K was a good jumping-off point for further refinement. Instead Windows just went down the tubes, what with wanting to phone home and visual coarseness (Windows for Tellytubbies) and others, particularly, IMV, KDE got it right - although these days even KDE are showing worrying signs of going off piste.

[Liam Proven]

Re: Motif?

[Author here]

Interesting comment. I have to take issue with some parts, though.

> Yes, the article missed out much of the 1970s and early 80s changes.

Well, [1] I wasn't using Unix back then, I was at primary school. And [2]...

> the standardisation of ANSI escape sequences, for cursor control and screen positioning.

But that is precisely my point. This enabled CUA to come to Unix terminal apps. I installed and supported both MS Word for Xenix and WordPerfect 5.1 for Xenix.

WordPerfect, especially, looked and worked near identically on a Wyse 60 terminal to how it worked on a DOS PC with VGA.

The very thing you point out is _how_ what I asked for could have been done.

> DECWindows [...] will be [...] familiar to anyone using today's MS Windows

Indeed. It was not concidental. The Open Group actually paid MS to licence the Win 2.0 UI design elements for Motif: the bevelled window-control buttons and so on.

Just as around 2000 Corel paid MS to licence the Office 2000 look and feel, the toolbars and the buttons and so on, for use in WordPerfect Office. It also agreed to cancel Corel LinuxOS, WordPerfect Office for Linux, and the NetWinder ARM desktop.

Microsoft, being the fair dealing honest sort of outfit it is, promptly replaced all that stuff with new different versions in Office XP.

> a proliferation of UI designs whose main aim is NOT to be like Windows

Because Microsoft said it was going to sue.

https://www.theregister.com/Print/2013/06/03/thank_microsoft_for_linux_desktop_fail/

And then didn't.

> The shell was never intended to be much of a UI component

Oh come on. It was the sole UI for decades.

But the shell is how you access Vi, Emacs, pico, nano, joe, ee... or MC... or dvtm, or tmux, and any number of other full-screen console apps.

My core point is that thanks to tools such as termcap, and ncurses, there is absolutely no reason why the legions of existing text-only apps that take over the whole console shouldn't support the same standard UI as the graphical desktop, even including mouse support thanks to `gpm`, even over ssh or whatever.

It would make life easier for everyone, and need not hurt or inconvenience anyone in any way.

[Doctor Syntax]

Re: Motif?

Oh, come off it, Liam. The primary school is no excuser. When I'm not wasting time here I'm looking at medieval history and before that I've done research that went back to the late glacial. And even i am not that old.

But as to "there is absolutely no reason why the legions of existing text-only apps that take over the whole console shouldn't support the same standard UI as the graphical desktop" there are two reasons. The first is that they belong to the character oriented world, not the pixel oriented one. Screen space is valuable and not to be wasted with on-screen menus that are not needed. The other is that they're key-board oriented and we ancient curmudgeons are well aware that if your UI is designed to work solely with the key-board taking your hands off that to faff about with a mouse or put greasy fingermarks on the screen ruins your productivity.

By and large these character-based editors were designed - and honed to perfection even - before CUA and, much as i support the latter in its rightful place, there's no reason why they should have chucked away their essence to adapt to an entirely different UI approach.

[CRConrad]

Re: Motif?

Screen space is valuable and not to be wasted with on-screen menus that are not needed.

Easily solved, and AIUI right there in the CUA standard (otherwise, an easy addition, as witnessed by the many applications that already do it): Keep menus hidden until you press Alt.

they're key-board oriented and we ancient curmudgeons are well aware that if your UI is designed to work solely with the key-board taking your hands off that to faff about with a mouse

Tell me you didn't read the article without telling me you didn't read the article: The author goes on a bit about how being able to use applications without having a mouse is a vital and integral part of the CUA standard. Freaking Windows still does that! (At least up to version 10; only used 11 for a couple months, about a year ago.) You really never grokked what the underlined letters in the menus mean?!?

Sheesh.

[Martin-73]

Re: Motif?

can confirm win 11 does this too

[John Brown (no body)]

Re: Motif?

"One of the very useful UI changes from the 70s was the standardisation of ANSI escape sequences, for cursor control and screen positioning. "

Ah, the days when programmes not only came with proper, detailed manuals, they had dedicated patch areas in the actual executable (and sometimes an actual patcher program!) so you could customise the screen and printer ESC sequences to match the hardware you were using. :-)

[Peter Gathercole]

Re: Motif?

If you look into the command sequences for VT52 and VT100, you will see that VT100 was not an evolution from the VT52. It was a completely different terminal, not even sharing the same form (the VT52 did not have a separate keyboard). It did have a VT52 mode and understood some of the commands of the older terminal in VT100 mode, but the structure of the command set was completely different. The command set for VT52's and ADM-31's looked much more similar.

But the thing that made the VT100 (well, actually probably the VT102 which was taken as the base for xterm) the reference was that it was the first commonly available terminal that implemented the ANSI X3.64 terminal specification. This, along with the fact that the VAX (and PDP-11) had a large market share meant that most other terminal manufacturers made their terminals so they would work under VMS, RSTS and RSX to maximise their potential sales. As far as I am aware, VMS did not have a terminal abstraction layer like termcap/curses (and I'm certain that RSX-11M didn't), so if you wanted to use a terminal from another manufacturer, you had to get one that worked like a DEC VT terminal.

Other mini-computer manufacturers tried to do similar things (I shudder whenever I think of an IBM 3151 in native mode), but DEC won the day to become the one everyone copied.

[Phil O'Sophical]

Re: Motif?

If you look into the command sequences for VT52 and VT100, you will see that VT100 was not an evolution from the VT52. It was a completely different terminal, not even sharing the same form (the VT52 did not have a separate keyboard). It did have a VT52 mode and understood some of the commands of the older terminal in VT100 mode, but the structure of the command set was completely different.

Agreed, I really meant that it was an evolution in the sense that it was the next step, not a direct technical update. I never had much chance to play with a real VT52, most of my experience was VT100/102/13x and the VT2xx series. The VT52 escape sequences were much simpler than the ANSI ones, and the VT2xx ones with <CSI> prefixes extended that even further, especially the colour ones on the VT240.

VT100 ... was that it was the first commonly available terminal that implemented the ANSI X3.64 terminal specification

I rather think it was the other way around. X3.64 dates from 1979, the VT100 was publicly released at least a year before that.

As far as I am aware, VMS did not have a terminal abstraction layer like termcap/curses (and I'm certain that RSX-11M didn't)

I think that VMS eventually got termcap, when they got POSIX certification, but certainly things like the EDT editor didn't have any terminal independence. The escape sequences were hardcoded into it, on VMS as with RSX. VMS got a certain level of terminal independence when the Text Processing Utility (TPU) was added in the mid-80s. That was essentially a language for writing editors, and it came with EVE, an EDT-lookalike written in TPU. I remember getting it mostly working on a Wyse terminal that had non-ANSI cursor positioning sequences, but the non-VT100 keyboard meant that many of the EDT sequences that were coded into my fingers' muscle menu couldn't be used, so it was more of an academic exercise than a practical one.

[Peter Gathercole]

Re: Motif?

Your comment about the timing of VT100 and ANSI X3.64 is interesting. I had not looked into the dates, but I just thought that the "<ESC> [" CSI was just so bizarre that it must have come from a committee (the ANSI X3L2 committee, to be precise) rather than a single manufacturer. This committee seems to have been working before the introduction of the VT100.

At the time that the VT100 came out, I was still at university, and did not actually use real VT100 terminals at all. Most of my experience came later, and from VT compatibles. The reason why I wrote a VT52 emulator was that the SYSTIME IV terminals that were on my (as in "the one I looked after") SYSTIME 5000E which was the system I was responsible for, running RSX-11M and UNIX Edition 6 and 7, were merely VT52 supersets (they had additional function, some of which I did also implement). This enabled me to keep track of what was going on from my office (where I had both a BBC Micro and a serial connection to a CAMTEC PAD which had a line to the system) without tying up one of the small number of terminals available for students in the lab itself.

The 'killer apps" that needed compatibility were EDT (as you pointed out) and a transaction monitor called SYSTEL which was being used to teach commercial application design.

In my next job, I came across all sorts of different terminals, from Wyse, Lear Segler, HP, AT&T, IBM and several cheaper VT2x0 compatibles (as well as real VT220's themselves), and I became the go-to person for anything terminal related, and I've sort of kept that reputation wherever I've been since.

[fromxyzzy]

Re: Motif?

I didn't really experience the era firsthand, but you (or someone else) may know: why have I read people hating on OpenLook? I've used it on Solaris 2, along with DecWindows on VaxStations and Alphas and Motif on a variety of systems, I grew up using Gnome 2 and KDE back in the 90s, and of all of them, OpenLook was the nicest and most coherent GUI on *nix in the 90s. Was it really awful under the hood or something?

[Michael Wojcik]

Re: Motif?

I don't know why it was unpopular. Open Look (I guess the official rendering is "OPEN LOOK", because who doesn't love block capitals?) was actually just the look-and-feel specification, and it was implemented under both NeWS and X11. There are detractors for both windowing systems, of course; some people really didn't like the whole Display Postscript concept underpinning NeWS, while others took issue with Xlib.¹

Sun eventually used the X11 version of Open Look for OpenWindows, which might be what people recall disliking. Don't think I ever used it myself; my preference was minimal window managers and more basic toolkits.

IIRC, Open Look invented the "pin" control for making non-modal dialogs persist rather than being unmapped when an action button was pressed. That's since appeared in some Windows applications, such as Visual Studio.²

¹IBM attempted to have it both ways in AIX 3, by adding a Display Postscript extension to X11. I don't think it saw a lot of use. I recall playing with it in a couple of experimental programs, but didn't really see any compelling use case.

²Ugh.

[Anonymous Coward]

Re: Motif?

Sun eventually used the X11 version of Open Look for OpenWindows, which might be what people recall disliking.

That may have been down to the xnews window server, which handled both X11 and Display PostScript, and was very slow as a result. That said, I knew a system admin who still ran Open Windows with an OPEN LOOK interface well into the 2000s, he hated CDE & Gnome so much.

[munnoch]

Re: Motif?

Indeed, was about to comment the same. Kids nowadays seem to think Unix history began with Linux...

Motif/OSF was the goto widget toolkit for any X application we built. We'd occasionally fall back to the odd Xt function call but mainly it was Xm all the way.

Still have the row of O'Reilly "Volumes" on my bookshelf. Utterly useless now but so much nostalgia. They were well thumbed at the time.

[Handy Plough]

Re: Motif?

Couldn't agree more with your first sentence. I saw one recently lamenting, completely unironically, the fact that the BSD's don't use "standard Unix tools..."!

[munnoch]

Re: Motif?

Ah BSD, you never forget your first love... Took me a long time to forgive Sun for switching to System V with Solaris 2... At least you could still put /usr/ucb first in your PATH.

[ldo]

Re: BSD's don't use "standard Unix tools..."!

My first encounter with trying to use networking on a BSD machine, I discovered that the “route” command had no option to display the routing table! Instead, that was an option on the “netstat” command.

Even before iproute2 came along, at least Linux was more consistent than that.

[IvyKing]

Re: Motif?

Motif was an HP creation and the basis for Visual User Environment from which CDE was largely derived from. Unlike Window$, there was an effort for coherent design in the UI as described in the book about the visual design of Motif. One thing I miss is the emphasis on using as little of the display's area for the windows. As for the supposed complexity of using VUE or CDE with a three button mouse, my then 2.5 year old daughter had no problem in figuring out how to click on an icon to bring up a picture just after saying "daddy's 'puter".

The other thing I miss about VUE/CDE was that the default application for opening a file was a text editor - files with registered magic numbers or extensions would be opened up by the appropriate application.

[Doctor Syntax]

Re: Motif?

I think that oversimplifies the history. OTOH W95 took a lot from HP's overlay for Windows 3 (memory says it was called New Era but I may be confusing it with something else). The early W95 and beyond included a copyright declaration for HP. New Era (or whatever) had a huge number of tiny files with character salad names defining odd bits and pieces which, I think, became the foundation of the registry when gathered into one place. It also have a better text editor than W3.

[Michael Wojcik]

Re: Motif?

one of the first Unix widget toolkits, Motif

I'm not sure I'd call Motif "one of the first" X11 widget toolkits. Its first release was in 1989, two years after X11R1 shipped with the "new" Xtk (and the "old" Xtk which was earlier, obviously), and under contrib the InterViews toolkit. X11's distribution eventually also included Xaw (the Athena Widgets, from Project Athena) and the Xt toolkit it was based on, and some others — memory says Xum (from DEC?), though I'm not finding any hits for that.

There was also a toolkit that emulated the look-and-feel of one of the Sun GUIs. OpenLook, maybe?

Motif was early, but Xtk and Xaw were widely used before OSF released Motif.

[Rafael #872397]

Thanks for the Microsoft Word for DOS

... and for the PTSD it brought :-)

[Liam Proven]

Re: Thanks for the Microsoft Word for DOS

[Author here]

Re MS Word for DOS PTSD:

To quote Douglas Adams:

"Actually, I quite liked it."

No, really. I knew and worked with and supported pretty much every DOS word processor there was, from WordStar (classic, Express, 2000, 1512, all of them) to Multimate to DisplayWrite. Some real ugly little spuds they were too.

The only one that defeated me was Samna Executive. Horrible app. It's amazing that Samna went on to write the lovely Amí for Windows.

Of them all, my two favourites were MS Word -- v6 was the best -- and LocoScript PC, which I was fond of because I owned an Amstrad PCW.

[IvyKing]

Re: Thanks for the Microsoft Word for DOS

One neat trick for Word for DOS was putting different default stylesheets in different different directories, so one could have differing formatting for different recipients. The other nice thing was that it was very easy to attach a different stylesheet provided that one was consistent in setting up the style sheets.

The "escape" fr triggering menus was very easy to work, as it rarely required a two key (e.g. ctrl-C) combination. Certainly much faster than having to use a mouse.

[CrazyOldCatMan]

Re: Thanks for the Microsoft Word for DOS

"DisplayWrite"

AAAAAAAAARRRRRRRRGGGGGHHHHH!

In my programmer days (where, on the DOS side we had to use DW4) we used to call it "DeathWish4" because that's what happened if you had to use it for any length of time..

Most people just used SGML in xedit.

[Jou (Mxyzptlk)]

Flash from the past

Nice article! I played with those programs when I was below and around 10 years old on the computers of my aunts and uncles...

BTW: What is "kestrokes" mentioned in the article? It was the slang back then :D.

[Alan Bourke]

This far down in the comments

without someone claiming that WordPerfect 5.1 is better than all modern word processors?

Frankly amazing.

[BinkyTheMagicPaperclip]

Re: This far down in the comments

Better is definitely subjective!

Faster - quite possibly. Flexible, most certainly.

No WYSIWYG though (preview doesn't count), no easy modern integration of images etc - in fact I can't remember how everything was combined, but certainly remember the equation editor being an external program shelled from WordPerfect.

The best that can be said is that it was a very capable word processor for a 286. As soon as computing power increased on a 386, and definitely a 486, it was possible to create a capable WYSIWYG word processor and the world changed.

[doublerot13]

Re: This far down in the comments

Wow, I remember preview simply blowing me away. I was absolutely staggered.

[A Non e-mouse]

Re: This far down in the comments

Wordperfect 5.2 for Windows was the pinacle of word processors. Since then it's been downhill.

[BinkyTheMagicPaperclip]

Re: This far down in the comments

Oh I don't know, there was WordPerfect 6 for Windows and DOS too. Did 5.2 use Windows printer drivers, rather than that or the earlier Windows version which still required WordPerfect specific drivers? They should never have released it with that requirement

[Liam Proven]

Re: This far down in the comments

[Author here]

> WordPerfect 6 for Windows and DOS too.

True... and the DOS version is a wonder on 21st century hardware.

WordPerfect 6.2 in text mode on DOS is blisteringly fast on even a 15YO Core 2 Duo. It's quicker than WP 4.2 was on a 286. A lovely experience.

Corel should hire some DOS guru for a year, buy DeviceLogics and DR-DOS, and sell a bootable USB key with WordPerfect 6.2 preinstalled. Beautiful lightning-fast rich word processing, and 100% distraction free.

I even have a plan for how to write a 32-bit DOS memory manager that lets DOS boot on UEFI machines.

[Jou (Mxyzptlk)]

Re: This far down in the comments

> I even have a plan for how to write a 32-bit DOS memory manager that lets DOS boot on UEFI machines.

I thought FreeDOS + Grub already managed that? Well, at least booting on pure UEFI machine with "Compatibility Mode Disabled".

Good luck to improving that for FreeDOS!

[Liam Proven]

Re: This far down in the comments

[Author here]

> I thought FreeDOS + Grub already managed that? Well, at least booting on pure UEFI machine with "Compatibility Mode Disabled".

No. FreeDOS, like all other forms of DOS, requires a BIOS and will not boot on a UEFI-only machine.

This is a FAQ on the FreeDOS mailing lists and on their website:

http://wiki.freedos.org/wiki/index.php/UEFI

[Someone Else]

Re: This far down in the comments

Corel should hire some DOS guru for a year, buy DeviceLogics and DR-DOS, and sell a bootable USB key with WordPerfect 6.2 preinstalled. Beautiful lightning-fast rich word processing, and 100% distraction free.

I can tell you I'd buy it in a heartbeat!

[JoeCool]

Re: This far down in the comments

Word 6 is the first commercial app I actually paid money for, and was happy to do so. Still have the Corel package CD(?) somewhere.

[Liam Proven]

Re: This far down in the comments

[Author here]

> Word 6 is the first commercial app I actually paid money for, and was happy to do so. Still have the Corel package CD(?) somewhere.

Hang on, what?

Corel never sold Word. Word is a MS product and the DOS versions came on floppies, not CD. Even Word 6 for Windows came on floppies; I have a set.

[Anonymous Coward]

Alas for history

Nitpick: Word 6 on its own may have come on floppy only, but it also came on CD as part of Microsoft Office 4.3 - the poster may be confusing Corel Office and Microsoft Office.

That said, there was a floppy option. Office 4.3 came on a *lot* of floppies... I was really glad to have the CD...

[JoeCool]

Re: This far down in the comments

my bad, meant to type WP.6

[RAMChYLD]

Re: This far down in the comments

"I even have a plan for how to write a 32-bit DOS memory manager that lets DOS boot on UEFI machines."

Do it and donate it to the bods running FreeDOS.

[I ain't Spartacus]

Re: This far down in the comments

Microsoft Excel 97 was the only Wordprocessor anyone ever needed - and has not been improved upon since!

[Jou (Mxyzptlk)]

Re: This far down in the comments

No way to give yourself the troll icon here, it is way too close to the truth. Though the newer "xlsx" and "docx" file formats which came a few years later have their advantage. Since they are just glorified zip files with a specific file structure I was able to rescue or repair quite a few documents by removing wrong stuff. And correcting an image got easier: Get it out there, correct it, put it back in -> works.

[J. Cook]

Re: This far down in the comments

Along with stripping out _some_ forms of password protection, too, unless the chuckleheads went full paranoia with the "let word/Excel encrypt the document" button. I'm sorry Karen, but your grandmother's chocolate cake recipe does not require military grade encryption, and unless you've remembered the password on your file, you'll have to type it back in from the magazine clipping your grandmother got it from originally. /sarcasm

[Doctor Syntax]

Re: This far down in the comments

Whoosh!

[Someone Else]

Re: This far down in the comments

Excel as a word processor?!? Shirley, you jest.

[The answer is 42]

Re: This far down in the comments

If you think that was bad, have a look at Lotus Symphony. Word Processor, spreadsheet, database and PowerPoint all in one program, were all a spreadsheet with peculiar formatting. It worked by magic.

[John Brown (no body)]

Re: This far down in the comments

"Microsoft Excel 97 was the only Wordprocessor anyone ever needed - and has not been improved upon since!"

Wait...what?!?!? All this time I thought it was a database!!!

[Anonymous Coward]

Re: This far down in the comments

When I were a lad we had roff and troff ...

The typefaces were rubbish but you could do almost everything else ...

But VAX Document was the bees-knees. ...

[Peter Gathercole]

Re: This far down in the comments

It depended on your printer. For nroff, it generally used whatever the fonts the printer provided. Troff, in particular Device Independent Troff (di-troff) actually was quite flexible. It was only the original troff that would only drive a Linotype phototypesetter that was limited, and I think even that had different fonts available.

I used di-troff on Xerox 9700 laser printers (although that di-troff backend may only have been in AT&Ts R&D UNIX), and that had different font sets you could load, although you did have to have a font metrics files for each of the fonts.

And later, I also used a di-troff Postscript backend driving a DEC LN03 laser printer, with all of the standard Adobe fonts available.

It was possible to drive printers with nroff doing such things as different sized and proportionally spaced fonts, but it was a pain in the neck to create the font metrics file (I did it for an OKI laser printer once, and it took weeks for me to be happy with the result).

But back then, every word processor relied on the fonts that the printer provided. It was only with the advent of GDI for windows and GS/Gutenprint for Linux and possibly Postscript as an intermediate graphics language that the possibility of rendering-to-pixmap in the PC became a thing, allowing the application to use any font that the rendering engine knows about.

[JoeCool]

Re: This far down in the comments

Is there a need to restate the obvious ?

[ldo]

Re: This far down in the comments

And this far down in the replies, and nobody has yet said how they wish that modern word processors had a “reveal codes” function, too?

Maybe those who didn’t know why we don’t need one have gone.

[JimC]

Re: This far down in the comments

Personally I claim that there is yet to be a really good Windows Word processor.

[Ian Johnston]

I use XFCE with Linux Mint, which is generally nice but even then the abomination of the new GNOME interface is creeping in. Thunderbird has to be asked nicely to display a menu bar )(for how long will that be an option) and far too many things think that having two "gearwheel" icons top right doing completely different things is a Good Idea.

OK, Marketing Division of the Sirius Cybernetics Corporation first, but the GNOME developers should be next against the wall.

[Anonymous Coward]

Everyone raves about XFCE, but when I tried it, it was impossible to resize a window just using a mouse. The drag is not the usual triangle in the bottom right hand corner, but a one pixel sized point, which is almost impossible to hit. WTF? When I queried this I was flamed and told use the keyboard and mouse to drag and resize a window. Like just using a mouse was some weird unreasonable UI request.

Ended up going for Budgie as my Linux desktop of choice. Just nice and intuitive to use.

[corb]

Those nonsensical one-pixel window borders in XFCE were very likely thanks to the Greybird theme, which is very often the default in XFCE distributions, and very often puts one-pixel borders on its windows. Other XFCE themes use more practical borders that work well with a mouse.

That problem is compounded by using XFCE on a 4k/HiDPI screen. GTK components can be scaled at 200% but only the default XFCE window theme (providing the header bar and borders) honors that. Other XFCE themes shrink to an unusable size. And, of course, it makes the problem of chasing a one-pixel window border twice as fun. Various workarounds exist that attempt to remedy this built-in XFCE weakness. I've found them incomplete and not worth the annoyance.

[Liam Proven]

[Author here]

> That problem is compounded by using XFCE on a 4k/HiDPI screen.

Although I've not tried that -- only my Macs have such screens, where Xfce is a little moot -- but I believe you. Xfce's HiDPI support is poor. It does have a scaling factor, but under the hood it works by reducing the screen resolution. Some choices require entering scaling factors of less than one. Both, IMHO, are ugly hacks.

I am hoping that Xfce 4.20 works well under Wayland -- then, I might finally be willing to switch to a Wayland-based desktop. It looks as if soon most of us may have to move to Wayland, which saddens me.

[Liam Proven]

[Author here]

> When I queried this I was flamed and told use the keyboard and mouse to drag and resize a window

Really sorry to hear that. Queried it where?

Yes, this is a problem, but it's a problem of the default theme many distros put on Xfce and not of the desktop itself. Your flamers were ignorant. It's easily fixed: either just pick a theme with thicker window borders, or slightly more complicated but better if you like the look, edit the theme and make the window borders a few pixels thick.

[Mike 137]

"Your flamers were ignorant"

Flamers usually are -- flaming's symptomatic of Dunning Kruger syndrome.

[Dave559]

Window borders

Yes, this current fad for window themes with no borders (or only a barely perceivable ~1 px border), except at the top (and, even there, all these abominations of program controls eating into the titlebar as well [1]) really annoys me.

Is it really too much to ask for a desktop theme where you can have ~2 - 4 px (configurable) borders, perhaps even with corner grab handles, and where there is a clear colour differentiation between the titlebar and borders of the currently active window versus the other, inactive, windows, so that I can quickly spot which one I am currently interacting with? (And, yes, of course focus follows mouse…)

[1] Invading the titlebar is not the answer to the problem that this is trying to address: refusing to buy crappy low-res 'shortscreen' laptops is. 16:10 aspect ratio (or better) or nothing!

[druck]

Re: Window borders

Bring back CDE, it had borders you could really wrestle with.

[John Brown (no body)]

"When I queried this I was flamed and told use the keyboard and mouse to drag and resize a window."

I had the same issues until I change the theme. On a slightly more positive note though the "standard" (CUA compliant??) keyboard shortcuts do work, eg alt-space for the window menu :-)

[Anonymous Coward]

Could you recommend a XFCE theme please? The reddit mxlinux forum, that seemed to be full of keyboard shortcut fans, the idea of alternative theme was finally suggested but no one could actually name a specific one. The ones I tried that were supposed to have thicker borders (seemed line 2px!) were equally tricky to resize.

Would like to give XFCE another try.

[Doctor Syntax]

"But its[CUA] influence never reached one part of the software world: the Linux (and Unix) shell."

It would be more accurate to put this the other way around. The Unix shell was in place well before CUA. Command line interfaces in general, VMS, CP/M, MSDOS or whatever, don't belong to the screen oriented world of CUA.

Like you, I decry the declining influence of the CUA in UIs such as KDE and especially in broswers. Consistency is a good thing and should be preserved wherever possible, both from one generation of S/W to the next and between applications. Life is easier if GUI applications such as browser, email client, word processor etc. all share a common UI approach.

But in the world of 80x24 editing text requires usage of as much of the screen space as possible and CUA is just so much clutter. That, I think, is why CUA editors didn't display vi for so many of us.

Recollections are rather dim now but at the time I thought the UCSD p-System UI rather slick. IIRC the menu was just a single line across the screen top. Sub-menus were just another single line over-writing the previous. when text editing the menu line would eventually scroll off the screen to make maximum use of space and return when you quit editing mode (Ctrl-C I think). It also had another neat trick that I've never seen elsewhere: enter editing mode with the cursor mid-line and, if there was any space at the end of the line it would shift the line contents right to open up some free space into which you could type. If there wasn't any space at the end of the line or you'd used up the space it has created it would split the line and move the right hand portion onto the next line. If you needed more space than the line you started on it would make it by moving the following text down, eventually clearing it off the bottom of the screen. No shuffling text down a word at a time. Quit editing mode and it closed everything up again. There was a lot of good stuff in there which has been largely lost in later UIs.

[Liam Proven]

[Author here]

In /. terms I'd mod you +1 Insightful.

But we're not on /. so...

Yes and no.

> The Unix shell was in place well before CUA.

Yes it was. I said as much. *But* so was DOS, and most DOS apps changed to accomodate the new standard. I feel *nix apps should have too.

> Command line interfaces in general [...] don't belong to the screen oriented world of CUA.

I disagree. Almost all CLIs involve occasionally switching to full-screen apps. I think that the Right Thing was to update those full-screen editors, viewers, file managers, terminal multiplexers, whatever, so that they were CUA-compliant -- wherever possible preserving existing UI but making them conform.

> CUA is just so much clutter.

I could not disagree more. It's a few lines, and it's a price entirely worth paying. There are plenty of efficient examples such as all the DOS Borland languages, and it enables efficiently scaling to large text resolutions, such as 80*50 and 132*43 etc. The bigger the res, the more use repositionable subwindows are. Equally, as a counter-example, look at the inefficient screen space usage of Pico's 2-line menu, and Joe's huge WordStar screen.

> It also had another neat trick that I've never seen elsewhere

That is how Locoscript on the Amstrad PCW handled insertion as well. It's a lot lighter on CPU than trying to move all the existing text after the insertion point up, and rewrapping, character by character.

[Doctor Syntax]

"That is how Locoscript on the Amstrad PCW handled insertion as well. It's a lot lighter on CPU than trying to move all the existing text after the insertion point up, and rewrapping, character by character."

Ah, we agree on something - on the rest we agree to differ. I never used Locoscript but it sounds as if they had the right idea.

My own trajectory was punched cards to Z-80 basically as an instrument controller to IBM-compatible as a Z-80 replacement. Overlapping that was VT100/220 talking to Unix boxes or PCs running Kermit as terminal emulation. Windows initially came along as a vehicle for running multiple terminal emulations to Unix boxes. IBM only entered that world through its Unix offerings - and without CUA.

[CRConrad]

Rather ironic given your user name.

"But its[CUA] influence never reached one part of the software world: the Linux (and Unix) shell."

It would be more accurate to put this the other way around. The Unix shell was in place well before CUA.

So what? Just because something is there before doesn't mean it can't be reached by outside influences. Like, “Notwithstanding all his other achievements, Napoleon Bonaparte never reached the peak of Mount Everest”. Perfectly true, even though Everest was there long before Napoleon was born.

Likewise, “CUA’s influence never reached the Linux and Unix shell” has nothing to do with whether the shell existed before CUA.

[Uncle Slacky]

helloSystem

I'm hopeful that the helloSystem folks will be successful in reviving (or at least sensitively adapting) the Macintosh Human Interface Guidelines:

https://hellosystem.github.io/docs/developer/ux-guidelines.html

[Roopee]

Excellent!

An excellent and very comprehensive summary Liam, thank you!

It brought back a lot of memories and reminded me why I quickly became a Windows convert and evangelist (I'm not any more I hasten to add). When I started training as an accountant in 1989 (CIMA) the progressive first year IT syllabus included WIMP UIs - because nobody in the accounts world (neither clerks nor professionals/managers) had any idea what one was; green screens were the norm.

Now I'm just an 'old reactionary' bemoaning the loss of sensibly-designed UIs (like our friend @bombastic bob)!

Incidentally I think you've nailed the key points of the historical progress from the technical, business, economic and legal perspectives as far as I am aware (and you've clearly done a lot of research too); the extra details other commentards have added are just that - details. Here on El Reg we all like details, but adding more would cloud the message...

[BinkyTheMagicPaperclip]

It's been worse than ignorance for a time

Whilst I will grudgingly accept that an interface being pretty is useful to a lot of users[1], the pendulum swung away from an interface being usable above all else some time ago.

[1] One of the complains about the CUA compliant OS/2 2.0 and 2.1 was that by default it's extremely drab. Grey features heavily. The reason for this is to prevent eye strain, especially on a CRT monitor with less than 75Hz refresh, changing to a grey background over white leads to a more comfortable experience even when you're young. Windows 3.x decided on a more appealing colour palette.

Reduced consideration for users with impaired facilities, changing functionality on a whim or worse A/B testing for subsets of users, using users as unpaid testers as a result, poor documentation (usually only on the Internet), removing functionality because it was only used by a minority of users. Basically considering business first and end users second. That's modern development.

Note I do not necessarily include the ribbon here. Whereas it rightly annoys many 'power users' if the overall effect is to improve the user experience for inexperienced users that's a greater win than catering mostly for a smaller group of seasoned users who are less likely to upgrade as they know how to get the most out of the current release.

Accessibility, like backup and security, is not appreciated by most people because they don't use or see it. As a result it's less likely to be an emphasis in a commercial product, and vastly less likely in open source, where the emphasis is either offering functionality and solving problems, or creating something for your CV.

I still remember the 'delights' of Raspbian, where a specific key combination does an unconditional reboot of the system and enables text to speech on interface access. Disabling this functionality is extremely awkward, and whoever designed it should hang their head in shame.

[Anonymous Coward]

Re: It's been worse than ignorance for a time

I have to apologize for that Raspbian behavior because I may be partially responsible for it. I didn't write any code, but I think I might have given whoever wrote it the idea. The problem is that, on many operating systems, there is a key command that can be used to turn on the builtin screen reader. This has been important for us blind users because, without it, we can't interact with any computer unless someone sighted follows our instructions to turn it on first, which gets really annoying. Windows and Mac OS have one, Windows only since Windows 10, admittedly, and some Linux distros have one as well.

Raspbian did not have one. There were a couple options for blind users of Raspberry Pis who wanted to use the desktop on them. One was to get a custom image that had the screen reader enabled by default, but that one wasn't kept in sync very often and eventually stopped existing, so it wasn't a great option. The other one, and the one I used, was to mount the card after imaging it and write a systemd service to enable it for me. This obviously doesn't work well for people with less experience. At one point, I wrote this down on a Raspberry Pi forum, without giving them any suggestion on what to do instead, and the next release happened to include a key command to start the screen reader. From what you've said, it sounds like the key command doesn't work like the ones on other systems. On most systems, the same key command that turned on the screen reader can also turn it off, but if it doesn't do that on Raspbian, then I can see why it would be annoying to you. I'm not sure why it's rebooting either, as the program should function without that. I can't remember if it does that because it has been a while since I used Raspbian as a desktop.

[mfalcon]

Vi and WordStar key commands are all I've ever needed

Late 50's former Unix admin now Linux admin here.

Having lived through the history described in the article I'd say that in recent years user interfaces have gone backwards in a big way. Just today I was struggling with Windows 11's minimalist interface where you can't clearly see the scrollbars anymore.

Modern Linux GUIs are equally bad. I refuse to touch Gnome 3. I've always blamed Apple for the current minimalist fad where you can't easily work out what parts of the interface are the active components.

I started editing text using WordStar on a Visual 1050 CP/M computer back in the mid eighties. Once I got my hands on a terminal connected to a Unix computer I learned the way of Vi. Having learned the key commands for these two programs I've never really had to learn anything else. This post was written using Joe.

In the MS-DOS world I used the Q editor and the Borland Turbo-C IDE both of which were very WordStar like.

For Unix and Linux sys admin I use Vi for editing system files and Joe for code.

When I have to use Microsoft Word I prefer to write the text using Joe and paste into Word, its just faster.

I access the Linux world via a VNC session into a MATE GUI configured with many virtual workspaces filled with xterm sessions, Firefox and RDP session or two.

[Liam Proven]

Re: Vi and WordStar key commands are all I've ever needed

[Author here]

Thanks for the comments!

FWIW, this:

> Late 50's former Unix admin now Linux admin here.

... more or less describes me, as well. I am happy I moved away from that sort of work, though.

> Having lived through the history described in the article I'd say that in recent years user interfaces have gone backwards in a big way.

I agree.

However, as is probably clear from the article, I never liked Vi or WordStar much.

Very oddly for a short contract some 30Y ago I was introduced as the team's new WordPerfect and Lotus 1-2-3 guru, despite the fact that I never liked either of them much, either.

[LionelB]

Shell

"But [IBM's UI standards] never reached one part of the software world: the Linux (and Unix) shell."

There's arguably a good reason for this. The shell effectively grants complete access to control of all computer operations (including, e.g., file management and data manipulation, network and peripherals access, and a vast array of OS and userland configuration options) - which may not necessarily be at all menu-friendly. It's the difference between a passive "Computer, show me what you'll allow me to do" paradigm and an imperative "Computer, do this" paradigm.

Of course the latter requires that you develop an extensive knowledge of what the computer can do and how to tell it to do the things you want - but once that knowledge is accumulated the imperative mode is way more powerful and flexible than digging through mazes of (limited) options to find (or not) the thing you want to do. Due to the historically more techy user base of Linux/Unix, it is unsurprising, perhaps, that an imperative, semantically-rich command-driven paradigm became the norm; users have traditionally been capable and motivated to develop the requisite knowledge.

Sure, that is changing - so of course modern Linux distributions ship with perfectly okay graphical interfaces (or at least comparable to other OSes) for pretty much everything, if that's what you want/need; but to this day, for example, I find it way more efficient to handle all file management (including remote management) from the command line rather than a GUI file manager - because I know how to do that. (Not that I'm zealous about these things - I do, for instance, mostly use graphical editors/IDEs rather than vi :-))

[Anonymous Coward]

GUI Standards?

Nice article. That said, I've often wondered about the fundamental assumption behind "GUI design" - namely that all applications should have a similar interface.

Many of the commentards here have said that they learned the interface of a small number of applications -- and stuck with those applications for a long time -- even if they were "non-standard"!!

Surely the reason for this is that users want two things, and two things only:

(1) To get work accomplished

(2) Using a tool that is comfortable for them

I'm not sure that there are many users who actually get upset if their tool(s) of choice fail to "meet the standard"!!!

.....or if the next useful tool fails to "meet the standard"!!!

[Signed: XFCE user, gedit user, LibreOffice user, Thunderbird user, Chromium user, gcc user, Python3 user, K3B user, Thunar user, XFCE terminal user (ssh, sftp, find, zip, gunzip, tar.......) ]

[Roland6]

Re: GUI Standards?

The fundamental reason is user learning: learn once use often.

At a very simple level the standardisation of car controls means (mostly) someone can get into another car and drive off. Okay things can get a little confusing if you drive cars from say two vendors who have decided to place the indicators and windscreen wipers on opposite sides of the steering wheel…

The key back then was a desire to make computers more accessible to the masses and reduce the cost of training, something MS lost after XP.

[Someone Else]

Re: GUI Standards?

Okay things can get a little confusing if you drive cars from say two vendors who have decided to place the indicators and windscreen wipers on opposite sides of the steering wheel…

I'm looking at you, Peugeot!

[Anonymous Coward]

Re: GUI Standards?

Mainland European cars tended to be designed by default for LHD, UK and many Asian ones for RHD, hence the wiper/indicator variants. Just occasionally a manufacturer would swap the switch positions between LHD and RHD versions of a car, but I haven't seen one that still does it in years.

[Doctor Syntax]

Re: GUI Standards?

None of which explains the fact that my car end my wife's have the wiper stalks on the same side but mine moves anti-clockwise to turn them on and hers moves clockwise. It's just sheer contrariness.

[BinkyTheMagicPaperclip]

Re: GUI Standards?

What you're missing is that 1) To get work accomplished is followed by 2) 'with minimum effort'

There is unlikely to be a lower effort than using an application you already know well, unless the other application really is that much of an improvement.

Having a standard means the amount of retraining needed to learn newer applications is lower, and it's more likely they can be picked up and used with minimal effort.

I think in fact you would be rather upset if appilcations failed to meet the standard : 'standard' use of mouse buttons[1], 'standard' use of clipboard, standard use of command line parameters of '-v' '-V' or '--version' to get an application version (if it used '-vers' I bet you would roll your eyes)

[1] Mostly standard these days. In the 90s? MacOS : one button with keyboard modifiiers. Windows : left mouse button drag with keyboard modifiers. OS/2 : right mouse button drag with keyboard modifiers, X Window System : Frankly could be anything, quite possibly middle mouse button for drag (plus modifiers)

Not to forget truly horrid design such as one obscure Unix system that hosted a powerful CAD program where it was necessary to hold the mouse button down, scroll all the way down the menus and not at any time slip off the cascading menus or release the mouse button, because that activated functionality

[doublelayer]

Re: GUI Standards?

It's about what you can do after learning to use one tool. If everything is different, you know how to do the one task that tool is for. If things are standard, then you can probably figure out how to use other ones. So, for example, if you need to use a new piece of software for a bit, you already know the basics of how to open it, how to close it, where the settings are, what stuff you can click on and what stuff you can't, and you skip right to figuring out what you need to do inside it. If you might have to relearn the concepts, it's a lot more frustrating. For example, if you're used to ctrl+c copying and ctrl+v pasting and you also have options in the edit menu for those functions, but your new program has no keyboard shortcuts and the buttons are revealed by clicking a toolbar button with an unfamiliar icon, you spend more time learning that.

People who frequently have to pick up new applications will find this annoying. I imagine you've experienced it at least a couple times yourself. The worse situation is for those who don't do this frequently. How many people have you met who resist any change. An update might move one button and they panic? That's often because they've experienced something which completely changes their workflow and they need to spend a while re-learning things for no reason. There are times when that is necessary, but generally, if people already know some paradigm and yours doesn't offer a specific benefit that you can concisely explain and prove correct, it's best to stick with the working one. This is a lesson that many companies, especially Microsoft, would do well to follow. I wonder whether people would stick less to old Windows versions, even as they go out of support, if the next one looked familiar.

[FF22]

Got history completely wrong

This article seems to get history wrong from the very first paragraph.

Like it attributes common GUI to the SAA, which however was only published in 1987 first. That's several years later than the Apple Lisa and the first Windows came to the market, which all had standard UI elements (like menu bar, scroll bars, buttons, etc.) that virtually didn't change in the last 40 years. They all also had APIs for GUI management, which not only meant that apps running on them didn't have to invent their own GUI concepts, but were actually encouareged to use said services, which made them all use common looks and controls.

But even they all just mostly copied the Xerox Alto, which was the actual system that established the basic GUI elements.

[Liam Proven]

Re: Got history completely wrong

[Author here]

> This article seems to get history wrong from the very first paragraph.

I think you did not read it closely enough, then.

I tried to spell out that CUA was IBM's _response_ to the Mac standardising UI design, and CUA adapted the Mac's ideas and concepts to a model which worked equally well on a text-only display, and with keyboard-only controls, in other words which did not require a mouse.

Lisa: 1983, introduces most of the concepts.

Mac: 1984, refines them and simplifies the conceptual model to programs and documents.

CUA: defines a standard UI clearly based on the Mac and the Apple HIG, but which also scales down to terminals and text consoles, and works well without a pointing device.

> But even they all just mostly copied the Xerox Alto

Also wrong.

The Alto had no standard menus, no menu bar, no pull-down menus, no window control buttons, no dialog boxes, no standard action buttons, no desktop icons, indeed no desktop: no graphical file management at all. Every Alto app looked differently and worked differently.

Apple invented almost all of this, _de novo_.

The Alto introduced overlapping windows... and not much else.

[Handy Plough]

Re: Got history completely wrong

> "Apple invented almost all of this, _de novo_."

Sacrilege!

[Someone Else]

Re: Got history completely wrong

The Alto had no standard menus, no menu bar, no pull-down menus, no window control buttons, no dialog boxes, no standard action buttons, no desktop icons, indeed no desktop: no graphical file management at all. Every Alto app looked differently and worked differently.

I dunno, Liam. My recollection of the Xerox 850 and 860 word processors include such things as menu bars, dropdown menus (not cascading menus or context menus), dialogs, and the clipboard. This was in '81. You may well be correct that "Alto" did not have these things, but the devices made by Xerox (at least, these two specific things) did support these keystones of CUA.

It is true that the 850 interface was somewhat different than the 860. These differences were largely due to styling, but the elements were there in both.

_{And blinding monochrome -- black and white only. Gad!}

[Liam Proven]

Re: Got history completely wrong

[Author here]

> the Xerox 850 and 860 word processors

Yes, but those weren't the Alto, were they? AFAIK those were derived from the Star workstations, which were the 2nd generation, and Xerox's attempt to commercialise the tech.

Here are some screenshots:

https://guidebookgallery.org/guis/xerox/screenshots

Here's a closeup from the docs:

https://guidebookgallery.org/articles/thexeroxstararetrospective/pics/fig1

Note, no menu bars, no window control buttons in the title bar, and so on. This is an interface really quite unlike the standard one that we all know today... because when the Star series came out (1978) that stuff had not been invented yet.

[Someone Else]

Re: Got history completely wrong

Thank you Liam for the clarification, and the links. Not familiar with "Alto" but spent many hours in front of the 850 and 860....

_{BTW, 860 far superior to the 850...}

[Dave559]

Hostname

I'm just very disappointed that Liam hasn't bothered to give his Linux laptop a proper hostname… ;-P

That's definitely a loss of one geek point there! Choosing a naming scheme for all of your boxen is part of the fun…

[Liam Proven]

Re: Hostname

Ha!

Well, true. But then again I am decades past the point where there got to be so many machines on the home LAN that I started using descriptive names for them, rather than trying to memorise a mapping from fun-name-to-actual-box.

My bikes all had whimsical names. Sadly last year's smashed arm may spell the end of my bike-riding days. :'(

But I plan to retrieve my recumbent from Prague and see if I can comfortably ride that with the now-knackered right arm...

[Dave559]

Re: Hostname

Ooof, really sorry to hear that you have badly injured your arm. Hopefully with time it will heal as well as it can do.

Recumbent sounds fun, and hopefully less motion required for steering and controls, as you suggest.

I once met someone on a big mass bike ride, who had sadly lost one arm but who had set up their bike to have a back-pedal brake for the rear and a brake lever for the front wheel, along with both gear levers (old-skool thumbshifters, not the complex combo-gadgets common nowadays) mounted on the same side, and they had enough strength in their remaining arm to be able to control these and steer, and maintain balance/direction, carefully. I was very impressed that they had successfully found a way to keep doing what they loved.

[42656e4d203239]

Re: Hostname

>>Recumbent sounds fun,

Recumbents are fun - and bloody fast if you are crazy enough.... in my youth¹ I had the opportunity to ride down a newly opened bit of road decending the side of chalk downland. After the cycle home, GPS reckoned I reached 50Mph.

The other advantage they have is that one has a very comfy² chair. I once rode 60 miles from north east somerset homewards and got nothing more than tired legs for my effort, despite my distinct lack of training!

I have never had a serious spill from the 'bent - just slow speed nonsense on ice - for those it is much less painful than falling from a standard bike.

Hope the arm fixes properly, Liam, at our age these things are less than certain!

¹for special values of youth

²for special values of comfy

[karlkarl]

Since CDE / Motif is the last GUI system that was ever standardized by IEEE. We should all just keep using that.

Or we could... wait for another to become standardised *snigger*.

[Liam Proven]

[Author here]

> Or we could... wait for another to become standardised *snigger*.

Well, you know what, if there _was_ one and it was any good, and had comparable R&D, I'd be OK with that.

It depends what the tool is and how you use it.

For instance, for me, as a Windows user, I went from Windows 2 and the MS-DOS Executive, to Windows 3.x and NT 3.x and Program Manager and File Manager, to Win9x and NT4 with the Explorer. I was quite happy with all of them from 3.0 onwards, and set some hotkeys to get stuff done and I was away.

I must as a result be one of the only people who was actually perfectly happy with Windows 8.0. To run stuff I tended to hit the Windows key and type a few letters, press Win+R to run a command, Win-D to see the desktop, Win+E for Explorer, etc. It all still worked fine. I barely use the Start menu anyway so I didn't miss it.

But, in contrast, as someone who trained people in word processing and spreadsheets from before MS Office existed, I find Office 2007 and later totally unusable. I despise the ribbon and find it nothing but an encumbrance. I memorised the menu tree -- all of it -- decades ago and can navigate it very fast. Searching visually for little symbolic buttons in random tabs is painfully slow, like searching for a product you want in an unfamiliar supermarket.

For me, the Ribbon is a terrible UI. It doesn't even have a vertical arrangement. It's broken junk and it forced me to switch to Libreoffice, but sadly, that has no outliner.

[captain veg]

Re: the Ribbon is a terrible UI

It certainly is.

For me it was a revelation in WIMP interfaces that all the things you could do were discoverable somewhere in the menus.

The rot set in with dialog boxes and menu entries ending ...

-A.

[John Brown (no body)]

Re: the Ribbon is a terrible UI

"For me it was a revelation in WIMP interfaces that all the things you could do were discoverable somewhere in the menus."

Today, I had to actually print out an email from Outlook. I don't often use Windows, let alone Outlook. I must have spent 10 minutes looking for a way to print that bloody email before I finally noticed an extra (pale gray) search box in the top of the menu ribbon. So I typed "print" into it and it immediately gave me the print options I was looking for. I still have no idea if there is a print option you can actually find and click on in the menu.

In retrospect, I should have tried hitting Ctrl-P, maybe that would have worked, but it was end of day, I was in a hurry and I was a bit overwhelmed by the 6 million menu options in the menu bar that I've never used or needed before ;-)

[Jou (Mxyzptlk)]

Re: the Ribbon is a terrible UI

> Today, I had to actually print out an email from Outlook. I don't often use Windows, let alone Outlook.

CTRL+P for print. At least since Winword 2.0, possibly before as well. The "quick print" icon is actually quite visible too. But even then: "File" "Print", like you are used from the times before Windows 1.0, i.e. even before Micros~9 had a graphical user interface.

It was clearly the end of your day, possibly overworked way beyond human capacity. Some countries drive their workers like that, and this is the result: Brain shutdown, you clock in hours, but you are not more productive.

[CRConrad]

Print...

...has been under the File menu, alternatively accessible via the Ctrl-P shortcut, since time immemorial. And still is.

[CRConrad]

I must as a result be one of the only people who was actually perfectly happy with Windows 8.0.

You and my mother – weirdos, both of you. Oh well, maybe there's just something about you that fits in with the octogenarian crowd. ;-D

[Mage]

Xerox

Jobs and Bill both visted Xerox.

Xerox originated the WIMP and essentially turned a tracker-ball upside down as the mouse because light-pen was rubbish and desktop pen (like Wacom, MPP/NTrig etc) far in future.

Apple, IBM, Digital Research, Microsoft and even some DOS and CP/M text mode programs copied the Xerox GUI ideas.

Wirth built Lilith and Logitech did the mouse.

[biddibiddibiddibiddi]

This went from "everyone forgot" to "Linux abandoned it, possibly for legitimate reasons". No mention of Windows apps that have gotten rid of things like standard menus and layouts.

[martinusher]

Not quite right

The notion of 'expensive minicomputers' versus 'cheap microprocessors' gets things totally wrong. Yes, its true that early systems first had teletypes and which were later emulated in glass but this was due initially to electromechanical devices being the only things available, interfaces were designed to suit and then early 'glass' teletypes emulating them because the interface was already in wide use. (But terminals quickly got cursor movement commands etc.) Microprocessors emulated their bigger brethren, that's all. The early operating environments emulated OS boot loaders -- you'd come up to a simple command prompt that gave you diagnostics and other simple command access to the hardware (or load the 'real' OS).

Graphical interfaces arose out of a lack of operating systems -- the trendy thing in code in the mid-80s was 'event driven'. Here instead of collecting characters from a command line and decoding and executing commands each keypress or other event such as a timer (or, later, mouse movements) was executed in a giant case statement. If you add a set of windows on this -- a stack of windows internally -- then you get an early version of a GUI. It actually does a pretty useful job of emulating an operating system but gets really unwieldy once the system has to deal with external events not related to the user and their interface such as a local area network.

There's really no magic in any of this. The problem is really that people keep trying to tweak stuff to make it their own. The CUA was a well written document and really was all you needed to make a half decent GUI but it had the problem of being a standard -- once you've implemented it you have to move on to something new for the most part. This is, unfortunately, not how the software industry works.

[Anonymous Coward]

Re: Not quite right...so never did any MacOS / Win16 programming then..

Its obvious from your ramblings that you never ever wrote any MacOS / Win16 end user software back then. Because pretty much every single statement you wrote is technically wrong. Both MacOS and Win16 had a full complement of OS Managers API's (process, memory, device, file etc) as well as a whole layer of GUI managers API's (Window, Control, Dialog, Event etc). Just because it did not look like a Unix CLI shell just meant you did not understand how any of this stuff actually worked.Thats all. And if you wanted something like a Unix CLI shell on MacOS you just fired up MPWShell. It was all in there. In 1986. Plus a lot of great features that almost 40 years later still have not been replicated on any on Unix / Linux shell.

And I guess you never wrote any AppleDOS / MS/DOS / CP/M end user software either. Because every non trivial application had a polling loop of some sort inside. Usually waiting see what key the user pressed in the current menu screen to see which code should be called next. Or if a code overlay had to be loaded. etc. etc. Most of us wrote defacto event loops and event handlers for any non trivial CLI application software back then. When I first sat down with the loose leaf Inside Mac in early 1984 there was nothing novel or new about the MacOS Event Manager. Just that the events were now directly supported by the OS and this and a lot of other stuff made life for us end user application programmers back in 1984 a lot easier.

Plus when I was writing software on both micros and minis in the early 1980's the micros cost a couple of thousand quid (for a typical business software setup) and the mini's (DEC and DG) started off at around 20K / 30K quid and quickly hit the mid / high five digits very quickly when yous started adding useful peripherals to the setup. Which is why when the first Unix workstations started coming out in the early 1980's priced at just under $10K they looked like such a bargain to people running minis. And why they were such a great success. With the people who ran mini's. Last time I did any mini hosted software development the total cost of the VAX's etc was over 200K quid. And we ran the final target applications on the target platform micros that cost about 1200 quid.

Guess which platform survived and prospered.

So wrong again.

[Dan 55]

Re: Not quite right...so never did any MacOS / Win16 programming then..

The OP is entirely right. This is the classic GUI event queue vs blocking call problem that both Windows and Mac have had since time immemorial.

Even now, on Windows 11 and Mac OS 14, a network drive which does not respond will cause Explorer or Finder to hang because the network call blocks and Explorer/Finder does not service its event queue until it returns an answer or times out.

Nowadays there are asynchronous calls and threads to solve the problem, but both these billion dollar corporations must still be short of a bob or two because they never get round to fixing the problem.

[Jou (Mxyzptlk)]

Re: Not quite right...so never did any MacOS / Win16 programming then..

> Even now, on Windows 11 and Mac OS 14, a network drive which does not respond will cause Explorer or Finder to hang because the

> network call blocks and Explorer/Finder does not service its event queue until it returns an answer or times out.

Behold, 'cause a few month back on the "Carany Insider" channel of Microsoft they announced *dumrolll* threaded explorer.exe to *dddrrrrummmmmmrrrrooollll* fix that exact issue which annoys us since 1995. Before that they mitigated the blocking effect, i.e. it helped but didn't really solve the issue. MAYBE this will work.

While they are at it they should bring "refresh" to the "new" explorer menu. One of the reasons why I activate the "use old context menu" right from the start.

[HammerOn1024]

I have not!

I have not forgotten and this insipidly STUPID crayon inspired crap needs to go! MS, Apple, get your heads out of your backsides and go BACK to the IBM standards! They work! Have worked and will ALWAYS work because it wasn't implemented by the village idiot squad from the aforementioned trolls.

[Handy Plough]

Re: I have not!

Apple never followed CUA. Why would they? It came well after they had written and distributed their HIG (Human Interface Guidelines). The significant majority, if not all, of keyboard shortcuts that existed on the Mac 40 years ago still exist on the platform today.

[Anonymous Coward]

I just love Standards ... so many to ignore

As someone who sat in on some of those ISV meetings back in the 1980's no one outside IBM gave a toss about CUA. At least in the PC end of the shrink-wrap business. Pretty much the same for OS/2.

The closing up of the IBM PS/2 architecture effectively destroyed all goodwill that Don Estridges very open platform of the original IBM PC / XT / AT had generated. If it was an IBM idea no one wanted any part of it as the control freaks in Armonk were back in control of the PC platform now. Not the "free wheeling spirits" down in Boca Raton who at least did not try to screw you over at every turn. But those guys in Armonk.Total bastards.

And that's why CUA actually failed. If it was an IBM "initiative", not going to waste my time.

Plus there were a lot of proprietary roll your own MS/DOS application GUI like menus already out there which most people could work out how to use.So good enough. And by 1993, no one cared any more as Win 3.x sales quickly caught up with MS/DOS sales.

Other stuff.

Wordstar users only cared about the command keyboard combos. Less what was on the screen. Pretty sure there was a CUA version of WS2000 kicking around the building in San Rafael but like the OS/2 version got shelved for the Win 3.x version. Might have been shown at trade-shows. Projects like that often were.

Win 3.1 released in 1991 was the first usable version of Windows. Win 3.0 had "issues". Mostly involving memory. Which made non trivial applications unstable with large docs. You really did not want to do real work with Win3.0. Which included even running MSC6 on some machines. Would just blow up during compiles. In the same way Win 1.0 in 1985 ripped of VisiOn of 1983 Win 3.x in 1990 ripped off MacOS circa 1986. So nothing new there.

As for the Unix world. Linux continues the Unix workstation traditions of extreme fragmentation and almost standards. Which was less of a problem in the Unix workstation market when most of the sales were as VAR packages of some form or other. So each vendor effectively had their own GUI and OSF/Motif was the industry fig leaf standard to pretend that they cared about user interoperability. They did n't.

So there was no large horizontal end user software market / ISV eco-system for the Unix workstation market segment. Just like there isn't for Linux. And for exactly the same reasons. Ask the J2ME guys how platform GUI / API segmentation destroys platform viability as a deep and wide software ecosystem.

You need a rigorously enforced standard GUI (which is enforced by a vendor) if you want a mass market software environment. Which is something Google understands and why they stomped Samsung (and others) so hard when those hardware manufacturers tried to balkanize the Android GUI. Again J2ME is the poster child for getting it wrong.

Because its all about the end user application software. It you want a mass market platform with dominant market segment penetration. And a large ISV business segment needs a standard GUI and API.

Otherwise you stay at 2% market penetration. For ever.

[martinusher]

Re: I just love Standards ... so many to ignore

A lot of what you say makes a lot of sense --- but.......

>You need a rigorously enforced standard GUI (which is enforced by a vendor) if you want a mass market software environment.

True but Windows is not consistent within its product range. Linux has converged on a "older Windows" type of interface (sure, you can load all sorts of UIs but you're just trying to use the thing, right?). The Mac has been consistent for a long time (but I still can't get used to the way they work their mouse and menus!). Windows just can't seem to make up its mind what it wants and this uncertainty is common across the product range. The CUA just provided a standard for a basic GUI, nothing fancy but something close to how most GUIs work.

Incidentally, Wordstar is an example of why you don't need WYSIWYG, at least not the Word sort. WS worked in a very small environment but could edit and format large files. Its focus was on usability. WYSIWYG text editors have, for the most part, still not fulfilled that promise. Sure, they're close, but placing things on a page and getting them to stay put.....that's still a real challenge for the Words of this world.

[Anonymous Coward]

Re: I just love Standards...so many to ignore...follow the market share

MS GUI standards have always been (fairly) consistent at any given time. Not just sequentially over the decades.

And all major changes have been driven by the "There's Now A New Guy In Charge" syndrome. The biggest change was the ill fated catastrophic stumbling with a "mobile device friendly" tile centric interface model in Win 8. But that was just yet another typical Big Corp C-Suite company politics power play over in Redmond / Bellevue. By the latest "New Guy In Charge". Who gives a f*ck about the users - I'm the new head honcho in charge here and I've just shafted my biggest rival (and enemy) which means I'll walk off with another few $10"s millions of stock options by the time the dust settles and I move on.

Yes. It really is that simple and that stupid.

The big push by MS from CLI to GUI (from MS/DOS to Win 3.x ) was totally driven by MS using this platform change to lock down a dominant market position (and those juicy net margins) in all the application categories that MS had failed to make a dent in during the MS/DOS era. MS were minor players in the word-processor and spreadsheets segments but they used every dirty trick in the book to make sure they were top dog by the time the platform transition was over. The rise of Word and Excel from 1990 to 2000 being a perfect example. Both mediocre products. But in MS's defense almost all their major competitors back then had an almost unerring ability for repeated self harm.

You never actually thought MS pushed Windows and GUI's so hard because it made software easier to learn and the user more productive. If CLI's had given MS a strategy for application software dominance they would be shipping MS/DOS 21.1 by this stage. Dismissing GUI's as just a fad from the 1980's. Like skateboards and Big Hair heavy Metal Bands.

As for Wordstar. For those who had invested the large amount of time in learning the command sequences it could be very productive. But that was a very steep learning curve. And all internal MicroPro user research at the time showed that most WS users never progressed beyond a beginners / intermediate level of understanding of the commands available. A result repeated elsewhere most notably in John Carroll's research on IBM's Mulitmate word-processor which was the basis of his fantastic book The Nuremberg Funnel about user interfaces and how end users learn them. Or usually don't.

GUI's were better because they were easier to learn not easier to use. By greatly reducing the domain knowledge demands on users. And reducing the cognitive load while using the applications. There are real improvements to be made in how application GUI's are presented to end users. But CUA was never one of them. Even back in the late 1980's.

As for post Win 7 versions of Windows and their interfaces. I'm writing this on a test laptop running a Septimized Win 10. Which I spent about 30 mins during installation making sure it acted just like Win 7 and looked just like Win 7. And so far the packet-sniffers seem to show that its just as unchatty as Win 7 when it comes to phoning Redmond and acting as spyware. Which makes it almost, but not quite, as useful as MacOS 7.6.1. In 1997.

Gotta love the White Heat of Progress

[Liam Proven]

Re: I just love Standards ... so many to ignore

[Author here]

I think you are missing almost every point I made here.

> no one outside IBM gave a toss about CUA

Wrong. It was a huge hit, and the point I made at length, with examples, is that it took off and almost all big-name DOS apps changed to accommodate it, _even if nobody called it CUA_ or referred to IBM. There was a standard PC UI set more by Windows 3.0 than by OS/2 or anything IBM, and DOS vendors aped the Windows UI just so they could sell upgrades to people who couldn't afford new Windows-capable PCs.

Adopting standards means less user training, fewer manuals, more capable versatile staff who can do more because they already know their way around.

It is not about box-ticking and compliance to IBM docs. It's about selling product. It's about bums on seats.

> The closing up of the IBM PS/2 architecture effectively destroyed all goodwill that [...] the original IBM PC / XT / AT had generated.

Too extreme.

[1] It didn't. I installed Apricot servers at customers with MCA buses. Other vendors did adopt it.

[2] Forget the expansion slots: look at the rest of the box. VGA and the VGA connector, PS/2 ports, 3.5" HD floppies, even SIMM slots... all these quickly became industry standards and some of them remain so now, 35 years later. In its way the PS/2 changed the course of the PC industry even if it didn't sell all that well itself.

[3] Secondly, the negative impact. IBM crippled OS/2 1.x because it wanted to keep the promises it foolishly made to buyers of 286 PS/2s, none of whom (to a rounding error) wanted or bought OS/2.

The result of OS/2 1.x being crippled was that it opened up an opportunity for Microsoft and thus begat Windows 3.0 which is what shaped the modern PC industry.

In other words, the *failure* of IBM's planned PS/2+OS/2 combination is what set the stage for 32-bit PCs running 32-bit Windows and thus Windows 95, and the desktop UI that most Linux desktops still use today, including ChromeOS.

Nobody produced or has ever produced a Linux with a WorkPlace Shell style desktop. And I am willing to bet that they never will. There are Classic MacOS-like desktops, Mac OS X-like ones, Amiga-like ones, RISC OS like ones, but there's not enough residual nostalgia for a WPS clone.

Aside: the success of 32-bit Windows, notably Win9x and NT 4, pushed Apple to the edge of collapse... and as its in-house efforts to produce a successor OS (ironically, along with IBM in Taligent) are what caused it to buy NeXT... which resulted in the iMac and OS X, and that resulted in Apple being the first £1Tn computer company, and that resulted in Microsoft trying to imitate Apple in Vista and onwards. That Vista taskbar is visible an OS X Dock rip-off.

IBM's ports caught on. Its removable media caught on. Its core app UI caught on. What failed were its expansion slots and its OS.

But what worked is what defined the industry for decades.

Don't obsess over the details, and look at what really happened.

> Win 3.0 had "issues".

It certainly did but it sold in the millions.

[CRConrad]

Re: I just love Standards ... so many to ignore

And that's why CUA actually failed. If it was an IBM "initiative", not going to waste my time. [...] And by 1993, no one cared any more as Win 3.x sales quickly caught up with MS/DOS sales.

Tell me you either didn't read or didn't understand the article without telling me you either didn't read or didn't understand the article... Thanks to the humongous success of Windows 3.x, CUA was also a humongous success, since Windows 3.x and its interface guidelines for developers were an implementation of CUA.

...CUA version of WS2000 kicking around the building in San Rafael but like the OS/2 version got shelved for the Win 3.x version.

As per the above, the Win 3.x version was the CUA version.

You need a rigorously enforced standard GUI (which is enforced by a vendor) if you want a mass market software environment.

Which is exactly what Microsoft did for the CUA.

Holy fuck is it annoying when people who don't know WTF they're talking about confidently spew their drivel all over the Internet.

[My Coat]

Thanks for the fascinating article. I knew basically nothing about any of this - only read the article because disabling cua mode is one of the first things I do when configuring emacs. Now I actually know what cua mode means!

[rcxb]

Unix shell and web browsing

The Unix shell is a famously rich environment: hardcore shell users find little reason to leave it, except for web browsing.

Only because links/elinks hasn't been updated to handle CSS and modern JS nicely. Which is a real shame because I loved being able to browse the web (even huge complex pages) lightning fast with a tiny fraction of 1GB of RAM. Not to mention vision-impaired users who need to use a screen reader or other assistive device.

e.g.: https://i6.photobucket.com/albums/y223/mypapit/elinks.png

[David 132]

Re: Unix shell and web browsing

>Not to mention vision-impaired users

Speaking of... has anyone heard from our friend and vision-impaired fellow commentard ShadowSystems recently?

Your remark just made me realise that I don't recall seeing any posts from him for a while.

I do hope he's OK.

[Doctor Syntax]

Re: Unix shell and web browsing

You're right. Not since May '21. Like you, I hope he's still OK.

[jake]

Re: Unix shell and web browsing

I asked about him in May of last year in a thread that I'm fairly sure he would have been following. No answer.

Hopefully he's found better things to do than hang out here.

To missing comrades.

[Anonymous Coward]

Re: Unix shell and web browsing

I'm a blind user and we tend not to use CLI web browsers. The experience is much worse than using the GUI. The reason for this is the availability of navigation functions. In a GUI browser, for example the Firefox I'm using to read this, I can easily navigate by almost any HTML element. I can move to headings, landmarks, links, list items, form fields (generically or of a specific type) using keystrokes implemented by my screen reader. If I was using Lynx, it would all appear as one block of text. Navigation in that text is not something those programs were designed for since you can see the entire text there, so all you might need to do is scroll. I can and have used it when that was the only option, for example when I was on a text-only SSH session, but otherwise, I use normal ones.

There are a lot of blind people who have never used the CLI, and although I'm in it all the time, it is not automatically the case that a CLI program is superior to a GUI one from an accessibility perspective. The article's points about accessibility, however, are generally spot on.

[ldo]

Yawn

“they just never caught on, despite plaintive requests”

In other words, the solution was offered, but mysteriously, those who had been clamouring for it ... maybe were never that numerous, after all?

“Even so, these remain tiny, niche offerings.”

What makes it “niche” is not the software, but the degree of adoption. If not many people actually want to use it, doesn’t that point to a lack of need, rather than an unfulfilled need?

If you see an unfulfilled need with no obvious solution, why not get a like-minded group together to do something about it? Maybe use some existing partial solution as a starting point, if there is one?

“As the trajectories of KDE and GNOME show, even projects that started out compliant can drift in other directions.”

Like all open-source projects, KDE and GNOME are driven by the needs of the community that keeps them alive, not by random armchair commentary from passive users.

To paraphrase JFK: “Ask not what Open Source can do for you, ask what you can do for Open Source.”

[captain veg]

not just software

Much of what I remember about CUA concerned the meaning of function keys.

For some diabolical reason keyboard manufacturers have increasingly made the F-keys do something else. Like control the audio volume.

This is perverse.

Mind you, it's not as perverse as Microsoft reinterpreting Ctrl+F in Outlook. There are very good and obvious reasons why I might want to search for a certain string of text in a received email. Forwarding a copy is not among them.

Then there's Outlook's inscrutable treatment of modifier keys. Most programs that deal with multi-line text have some way of inserting a line break which is not an end of paragraph. Usually it's either Ctrl+Enter or Shift+Enter. In Outlook, one of those two will just send the message, ready or not. Are you feeling lucky, punk?

-A.

[David 132]

Re: not just software

>Usually it's either Ctrl+Enter or Shift+Enter. In Outlook, one of those two will just send the message, ready or not. Are you feeling lucky, punk?

Aaargh. This.

A trick that seasoned Outlook users adopt is... when composing a mail, fill in the recipient address(es) last. Any inadvertent attempt to send the mail - whether by misremembered keypress or accidental clicking on a Send button that has been redesigned in this version and is now a picture of a paperclip because some UI designer thought it looked cool - will thus fail, allowing you to curse Outlook in private without embarrassing yourself to others.

[Nugry Horace]

Re: not just software

Ctrl-F for 'Forward' was done at the insistence of Bill Gates, because that was the keymapping in the (presumably non-CUA) mail program he used previously: https://devblogs.microsoft.com/oldnewthing/20140715-00/?p=503

[Jou (Mxyzptlk)]

Re: not just software

> Usually it's either Ctrl+Enter

That Send-hotkey comes from Netscape Mail, since ever I used it. In my case: Before it was version 1.0. Outlook simply took that standard over from them. And maybe, before Netscape Mail, others used that as "Send" hotkey too, though that was before my time.

[Doctor Syntax]

Re: not just software

"Most programs that deal with multi-line text have some way of inserting a line break which is not an end of paragraph. Usually it's either Ctrl+Enter or Shift+Enter. In Outlook, one of those two will just send the message, ready or not."

My first reaction to that is that in plain text the notion of a non-breaking new line is nonsense so why would you need a key combination for it?

But wait, this is Outlook, the home of top-posting HTML, get as far as way from standards as possible without actually breaking email so - yes, you're right. It is odd.

[Fruit and Nutcase]

...

For the uninitiated, next time you are using a mainstream/well established application (I've just looked at the Edit menu in Firefox - also Office applications such as Libre Office, MS Office), look for the [...]

That's a visual cue described in CUA to indicate to the user the option will lead to a further dialog - whereas a menu option without indicates that the menu action will be applied immediately.

[ldo]

Re: Ellipsis

Apple had that before CUA.

[Doctor Syntax]

Re: Ellipsis

That was one of Liam's points.

[Doctor Syntax]

Re: ...

Also a > will indicate another level of menu.

Kate (KDE Atdvanced Text Editor) manages both an ellipsis and an arrow for Save with encoding - the arrow indicates a further menu for the choice of encodings and the ellipsis the save dialog which will follow once you've chosen the encoding. Not that I've ever had a need to use that option.

[CRConrad]

Re: ...

Kate (KDE Atdvanced Text Editor) manages both an ellipsis and an arrow for Save with encoding - the arrow indicates a further menu for the choice of encodings and the ellipsis the save dialog which will follow once you've chosen the encoding.

That seems pretty darn silly. The dialogue-ellipsis belongs only on those sub-menus; the higher-level one should have only the arrow.

[aerogems]

Torn

I'm torn on this one. On the one hand, I like the idea of certain conventions being adopted that make it quick and easy to sit down at any computer, regardless of OS, and be able to achieve at least some level of functionality without really thinking about it. On the other hand, this tends to stifle innovation and is why we're still using the same basic desktop metaphor almost a half-century later. Even when Apple created iOS they just recycled an idea from a product they were selling in the 90s. I forget the name*, but it was an alternate shell for MacOS that I think was marketed primarily to education customers who needed to keep their classroom computer reasonably secure from the curiosities of countless students trundling through every day.

My general thought is that open source is where you can give people a chance to try out new ideas and see what works. You don't have shareholders to keep happy, you don't have payroll you need to make, you're free to experiment. Sort of like the olden days of Linux when you had different window managers for X11. There was FVWM which was vaguely like Windows 3.x, FVWM95 which was a shameless Windows95 ripoff, IceWM which I think was OS/2 inspired, AfterStep and another -- the name of which is eluding me now -- that had varying levels of faithfulness to the NeXTStep UI, and Enlightenment for people who have giant monitors or hawk-like vision that can see tiny font sizes.

Let Microsoft and Apple be the ones concerned with the boring tried and true, Linux and *BSD land can be a laboratory for trying out new ideas. Maybe they work, maybe they don't, but there's a lot of talented people in *nix land who probably have more than a few ideas of things they'd like to see improved in UI design and have the necessary skills to actually implement it. If someone hits on the next great idea, it can be picked up by Microsoft and Apple.

* If anyone can help with the name, that'd be great. It's been bugging me for a while and my Google-fu is lacking.

[Dave559]

Re: Torn

If it wasn't actually AfterStep that you were thinking of, then it was probably Window Maker or GNUstep, all of which I had thought(?) were relatively similar (at least appearance-wise, if perhaps not under the surface?) homages or rewrites of NeXTStep, but I hadn't really used any of them enough myself to say for sure? I had tried each of them for a while out of curiosity, but then KDE<4 and GNOME<3 seemed to be starting to get the most traction by then, so I then moved on (rightly or wrongly)?

I did think they all looked rather cool, and the way that the menus were always zero-distance away (activated under the mouse pointer, when summoned) was definitely quite nice.

If you look up the page for any of these on Wikipedia, they all seem to also have links to their other counterparts, if it helps to jog your memory…

(On a related note, I'm pleased to see that the xwinman site still exists!)

[aerogems]

Re: Torn

Window Maker, that's the one! GNUStep is a whole other project to implement the OpenStep API as an open source lib, though I think it shares some relation to Window Maker. El Reg even did a piece on a recent update to it, and I still had the brain fart. Have an upvote for the help. Now if you can help me figure out the name of that old MacOS (like circa System 7.5) software, you'd really earn your upvote.

For some reason the xwinman site won't load with my VPN active, but it does seem vaguely reminiscent of the site I used to check every couple of months back in the day when I would get tired of one window manager and wanted to try something new.

[LionelB]

Re: Torn

Window Maker is still actively maintained and going strong. Always had a soft spot for it, and occasionally still fire it up. Rock-solid, fast, highly consistent, visually appealing and easily configurable UI with a distinctive personality, which springs no surprises. Other desktop shells could learn a lot from their approach (basically, stick to a central principle, and if it ain't broke don't fix it).

[Handy Plough]

Re: Torn

I always wondered this, then I remeber that GNU/FSF has always suffered from NIH syndrome...

[LionelB]

Re: Torn

Not sure what you're getting at. Window Maker is essentially an emulation of the NeXTSTEP GUI based on an OpenStep-compatible API. Those things were Not Invented by GNU/FSF, but WM was embraced by the Linux community (although it arguably never developed enough traction to become mainstream).

[Diogenes]

Re: Torn

I worked for company that used IBM 3090s, AS400s and PC-ATs all at the same time. We adopted CUA for new projects and started working at changing older systems on an "as time allows*" basis.

Usually the tertiary project we were working on, to help fill in the time while waiting for our 1 or 2 compiles/tests a day because the machine was (the 3090 - they ended up buying a second one for all the non "billing" stuff

[Doctor Syntax]

Re: Torn

"There was FVWM which was vaguely like Windows 3.x, FVWM95 which was a shameless Windows95 ripoff, IceWM"

A quick check on Devuan. "Was" is only the correct tense for FVWM95.

"On the other hand, this tends to stifle innovation and is why we're still using the same basic desktop metaphor almost a half-century later."

If it ain't broke... In fact there has been innovation that doesn't break and innovation which does. The former just fits in more or less unnoticed* because it seems natural, the latter gets complains.

* Unnoticed because people don't notice it wasn't there before.

[aerogems]

Re: Torn

You know what you call fixing something that isn't broken? Progress.

There was nothing wrong with walking, so why fix it by riding horses? There was nothing wrong with riding horses, so why fix it with trains? There was nothing wrong with trains, so why fix it with automobiles? There was nothing wrong with automobiles, so why fix it with planes? There was nothing wrong with swimming, so why fix it with boats? There was nothing wrong with using tape reels for storage on computers, so why fix it with random access memory? There was nothing wrong with hunting and gathering, so why fix it with agriculture? There was nothing wrong with 8-bit CPUs, so why fix it with 16-bit CPUs? You can go on with almost every single technical innovation throughout history. There was nothing wrong with what preceded it, but someone "fixed" it by making something better.

[CRConrad]

Re: There was nothing wrong with ..., so why fix it with ...?

All obvious BS: The fault with walking was that it was much slower than a horse; the fault with riding a horse was that it was much slower and carried much less cargo than the train; the fault with swimming was that it was much slower and carried much less cargo than a boat (and you got wet all over), and so on and on. There is not a single such obvious advantage to any newfangled GUI convention.

A few definitely non-newfangled ones had some, like IBM’s Workplace Shell and other such initiatives. (Perhaps Lotus Improv also belongs here?) But they didn't really change the basic WIMP premise at all, nor even the “desktop” metaphor all that much. And they were all from the 1980s and 90s; in more recent decades we've got shit like Ribbons, phone interfaces on 32-inch desktop monitors, and Flat Everything on screens large and small. Try to point out a single concrete advantage over the current desktop metaphor to any of those before next spouting off about “progress” in general.

[jake]

Re: There was nothing wrong with ..., so why fix it with ...?

"There is not a single such obvious advantage to any newfangled GUI convention."

The ability to run many x-terms on one screen, alongside whatever bespoke GUI software your job required.

Then Mosaic Communications Corporation released Mosaic Netscape (called "Mosaic Killer" internally; you know it better as Mozilla), and the world changed. Suddenly it was hip to release open source GUI code.

[CRConrad]

Standards vs “innovation”

On the other hand, this tends to stifle innovation and is why we're still using the same basic desktop metaphor almost a half-century later.

On the gripping hand,

1) Maybe the reason we're still using the same basic desktop metaphor almost a half-century later is simply that it works better than anything else. And

2) That much-ballyhooed “innovation” is usually form-over-function bullshit like Kai's Power Tools, so you might well want to be careful what you wish for.

[Downeaster]

Mac User Interface Guidelines, the Ribbon Interface, and Changes to Interfaces

The original Macintosh Interface Guidelines helped to "standardize" computer interface guidelines for a long time. The Mac was popular in the late 1980s and 1990s. Windows tried to keep up and eventually adopted most of it. Pull down menus, standardized keyboard shortcuts for printing, copying, and pasting. Many of these have stayed true over the past 40 years. I still like pull down menus on a Mac or on a PC. The ribbon interface drives me crazy. It is supposed to be more efficient but isn't. Then again I learned computers and programs in the 1980s and 1990s. Interface changes also are needed for software companies and operating systems companies are needed for a couple of important reasons. One is economics. Software and operating system companies need to continue selling software. Changes in a software interface sell software to early adopters and them it becomes the norm. The ribbon interface is now on most kinds of Windows software. Pull down menus are gone for the most part. Money needs to be made by companies! Many of us would still be happy using MS Office 2003 forever. Interface changes I think are trying to be more "international" also with more pictures. We have less plain English labels and now we must interpret pictographics to determine a software function. As ea new generation comes along, the interface changes become the norm. Also tablets and phones have changed the computer interface. Remember how Windows 8 wanted us to use Windows on a computer screen that looked like a phone? Also traditional Mac OS system preferences changing into a more iPhone look in the last two Mac OS releases? Smart phones are also driving changes to the computer interface. Also tablets and computers are a converging. MS Surfaces being an example. Things are changing with technology always. Money has to be made, innovations can be good for some and bad for others, and new technologies keep software changing. Nice article, Liam!

[aerogems]

Re: Mac User Interface Guidelines, the Ribbon Interface, and Changes to Interfaces

The ribbon UI is much more space efficient, allowing you to expose significantly more functionality without having half the screen taken up by toolbars. Ye gods, I remember the days of Word 6.0 and seeing like a good 1/3 of the screen (sometimes more) taken up by toolbars all because someone wanted to use one button and didn't know how to create a custom toolbar or couldn't be arsed to just track the feature down the like two times a year they'd actually use it. I still prefer the floating version in Office v.X and 2004 for the Mac, but I'll take the ribbon over toolbar mania every time. Sure, sometimes I have to check 3 different tabs before I find the function I'm looking for, but it's better than looking through 3 different separate dialog boxes like before. Sometimes you'd have to go 3 dialog boxes deep to find a particular function.

[Doctor Syntax]

Re: Mac User Interface Guidelines, the Ribbon Interface, and Changes to Interfaces

"Money needs to be made by companies!"

That's no excuse for abusing customers by breaking interfaces at the same time as you're exploiting them by selling them what they've already bought.

[Dan 55]

Re: Mac User Interface Guidelines, the Ribbon Interface, and Changes to Interfaces

The ribbon interface is now on most kinds of Windows software. Pull down menus are gone for the most part. Money needs to be made by companies!

Now you can disable the ribbon on Office 365 so MS can make more money by renting you the thing they originally took away from you in the first place.

This option is not on Office 2021 though. Again, to make more money pushing you towards the rented version.

[HMcG]

Common elements, such as standard menus and dialog boxes, were often reminscent of MacOS<CTRL-H><CTRL-H><CTRL-H><CTRL-H><CTRL-H> Xerox Alto.

Everyone copied the Alto, then accused the competition of copying them.

[Stuart Castle]

Interesting article.

I find it weird that both Apple and Microsoft spent mllions on research to develop their own common UI guidelines, then largely ignore them.

Case in point: The Windows 95 UI looks properly designed. It's simple, and generally efficient. The Windows UI carried on with pretty much the same layout until Windows 8. Then, as Windows 95 approached it's 17th Birthday, Microsoft decided to abandon years of tweaking, and threw up (literally, IMO) the Windows 8 UI. While I like Windows 11 a lot more than I did 8, the UI still isn't as good as it was. I know I can install Classic Start Menu, but I shouldn't have to install what is essentially a hack that might stop working because Microsoft have decided to change whatever CSM hooks into to change the start menu.

Also, while I am a Mac fan, Apple don't get off scott free here. They are better than Microsoft in that they don't make massive UI changes, but they have made some changes I don't like, particularly to System Preferences.. I used to be able to navigate System prferences relatively easy. On the main window, look down the window until you see the preference pane you want, click it, and change any options you want.. Most preferences are available in 2 or 3 clicks.

Now, you either have to use the search (which, admittedly, has always been an option), or you have to know which category they've put stuff in. Your account details. Are they under Users or iCloud (for instance), and the staggering array of privacy and security permissions you now have to grant, sometimes having to quit the app so they take effect.

In both cases, applications rarely follow the design guidlines written by the OS manufacturer, although Apple are better at this than Microsoft. I'm a long term PC user (since the 80s), an I don't think I've seen any MS application that follows the official Windows Interface guidelines since at least 2000.. Office certainly doesn't.

Linux doesn't really have any standard interface, and while the various Distro makers may have their own UIs and guidelines, the fact the user can just replace the UI negates the whole idea of UI guidliness..

As a user, I think it's important that applications on a given platform have a consistant OS.

[Doctor Syntax]

"Linux doesn't really have any standard interface, and while the various Distro makers may have their own UIs and guidelines, the fact the user can just replace the UI negates the whole idea of UI guidliness.."

It's true that Linux offers the opportunity to experiment and one way to do that with UIs is to produce a distro to feature it. It's also true that Ubuntu had a similar rush of blood to the head as did Microsoft, producing Ubity and W8 respectively. Ubuntu users were lucky - they just had to switch to whatever they'd used previously.

If you look more closely at the Linux GUI desktop world the most popular have followed CUA principles. Special mention goes to the founders of Mate and Cinnamon, both of whom took on the task of perpetuating the Gnome 2 look and feel when the Gnome developers started going rogue. The ability to switch desktops means that it becomes possible for users to main consistency through time where as Windows, with its imposed this-year's-standard approach, drags users along with it however unwilling.

[Bitbeisser]

Interesting that I mentioned (with a Wikipedia link) CUA in a (kind of) retro computing group of FB and two days later Liam is publishing an article about it at El Reg... ;-)

[AdB3]

Did Star Trek IV know...?

This article evokes a distant memory of the ancient the Star Trek IV movie, (which I re-watched, it's at just about the 1 hour mark) when Scotty sits down in front of a late '80s Macintosh Plus and enters a chemical formula for what we now call optical alumina with the keyboard alone. (Only after saying "computer" repeatedly to the mouse, RIP Cortana, you could have been the future user interface - please just ignore this little quip in any comments)

Point being: the universal user interface! or as el' Reg has now illuminated, CUA

I really wondered if the writers knew as much?

~ AdB3

[Lewis R]

CUA should be a design goal for all software

At Arca Noae, CUA compliance is something we strive to achieve in all of our work, and particularly in ArcaOS, from the installer GUI to the installed desktop (well, its OS/2 Workplace Shell is rather CUA compliant, after all <ahem>).

Aside from flat interfaces being ugly, the lack of easily identifiable elements (buttons, dropdowns, scrollbars, heck...window frames) makes it a chore to navigate. The desire for some platforms to adopt the "everything is a handheld" paradigm simply does not work for a desktop OS. I wish that these concepts were being taught to young developers today.

Despite what everyone believes of me (as the managing member of Arca Noae, an OS/2 consultancy and software development company for ArcaOS, based on OS/2), I do use other OSes and see "modern" Windows software (and not all free or even inexpensive stuff, either). The complete lack of regard for consistency in the user experience is appalling. Menus whose layout makes no logical sense, inconsistent button labels and positioning, weird child window placement in relation to the parent, missing mnemonics (because, yes, I know, nobody uses a keyboard anymore), accelerators which make no sense, so there is no possible way anyone can remember the twisted finger combinations (and those are supposed to be accelerators!?)...I could continue, but what's the point? To me, lack of CUA compliance is lack of fit and finish, like one set a bunch of school kids about piling up as much stuff as they could into a room without regard for fitment or balance[1].

The idea behind having user interface standards in the first place is to make things recognizable and immediately useful - for the user. Without them, we'd might as well go back to Lotus-style slash menus, MS Word for DOS <Esc> menus, and WordStar Ctrl-key combinations which change with every release.

I'll get off my soapbox, now. Great article, Liam. I wish I'd written it myself. ;-)

---

1. Contemporaries of mine who were also fans of the "Tennessee Tuxedo and His Tales" cartoon will appreciate the homage to Mr. Whoopee and his closet - and everything which would fall on him every time he opened it.