helloSystem: Pre-alpha FreeBSD project chases simplicity and elegance by taking cues from macOS A pre-alpha project to make a new FreeBSD-based desktop operating system has adopted a minimalist design intended to appeal to Mac defectors. FreeBSD is a Unix-like operating system first released in 1993 based on the Berkeley Software Distribution. The core of Apple's macOS, called Darwin, uses some code from FreeBSD. Despite …
... could be achieved by labelling GIMP as 'Photo editor' or some such.
There's nothing wrong with the name of an application giving a clue as to its function.
Okay, okay, everybody knows the GIMP by now, possibly. But how many times have you thought, 'Ah, yeah, I remember installing an Android app that does X... now, what the bloody hell was it called?'
Actually "wacky" names like GIMP are an advantage where you can at least search for related information.
On the opposite end of the spectrum is Dia, which is a diagramming application. Trying to search for anything related to that is a waste of time.
> Actually "wacky" names like GIMP are an advantage where you can at least search for related information.
That's a fair point; there are downsides to giving an application an overly generic name. However, I still believe that names that hint at the function of an application ( Inkscape, Word, Paint Shop, SolidWorks etc etc) lowers the mental overheads for newcomers.
If just one or two applications have wacky names, well, yeah, the user can get a handle on it quite quickly. But if the package is littered with wacky names for every common type of program, the poor novice user has to juggle say a dozen fuzzy names in their brain - head space they could be better use to complete whatever tasks they sat at the computer to do in the first place. Head space they could better use to learn sone handy nuances, concepts and efficiencies.
[Please pass me a hammer. Lump hammer or tack hammer? Oh, tack hammer. Here you go. Thank you. ]
[Please pass me a hammer. Stanley or Ewing? Just pass me a sodding hammer!]
> So I'm guessing it's some sort of ruby application
Yep, ruby is in there too, but I have half an idea what that one is.
The application is RedHat Satellite.
I find it strange that some of these off-the-wall names seem fine (perl, python, apache, yum...) whereas others really grate (celery, trousers, candlepin, pulp, ...).
Probably Grumpy Old Man syndrome.
It used to be a pet peeve of mine that every GNOME app used to start with a G and every KDE app would start with a K, e.g. gcalc, konqueror etc. and it took a long time for them to realise how stupid that was.
Fortunately some usability has seeped into desktop Linux since then.
But I wouldn't ever rename GIMP "Photo editor" because it's way, way, way too complex for that purpose. There should be an explicit photo editor that lets someone fix red eye, crop, frame, filter etc. from a simple user interface. Let people discover GIMP if they outgrow the simpler tool.
>There should be an explicit photo editor that lets someone fix red eye, crop, frame, filter etc. from a simple user interface.
Google's Picassa used to be very good for that (crop, level, rotate, basic levels, red eye removal etc) and it was quick to use because it didn't use the traditional Open > Edit > Save [As] method - changes were made virtually and not to the original image file.
> It used to be a pet peeve of mine that every GNOME app used to start with a G and every KDE app would start with a K,
There was a book of Gary Larson cartoon that had an index at the back... It looked like this
A
B
C..
..S
The one with the ants on the donut
The one with the beagle
The one with the cat by a piranha bowl...
...
X
Y
Z
But wasn't that a semantic rule, where T prefixes a type, I prefixes an interface etc?
The silliest naming, I think, is when people make a programming language with a single letter or common word for its name. Then they think about how it will be indexed for searching and add 'lang' to the search keywords. Then again, even .net can retrieve a lot of unrelated search results if you don't include enough specific search terms.
"The silliest naming, I think, is"
Not even close. Naming a railway company with a single digit number (one) is far more stupid. Especially when you think that now and again there might be announcements over the air of train times and platform numbers and other numeric info.
FWIW that's a take on a Borges piece about a catalogue of animals of which my favourite is: "that from afar look like flies". It's nice take on any taxonomy.
The fundamental problem with the GIMP is that the name is offensive to many people, and deliberately so in a let's-laugh-at-this-bacronym schoolboy sort of way, so the name is now any active hindrance to adoption, just as Computer User's Networking Tool or Naturally Intuitive Graphics Generator Expediting Research would both rightly be.
"GIMP is that the name is offensive to many people"
When I first started in IT we had a colleague who was named Gay. At that time it did not have any other meaning. Times and preconceptions change. Nowadays that name would be offensive, the not at all.
Gimp was no more offensive than grep.
it's way, way, way too complex for that purpose
ack - gimp does things no other graphics application seems to be able to do easily. However, they are things that you kinda have to be familiar with gimp to use properly. Example, paste a 2D image into a 3 dimensional perspective slot, such as "faking" a monitor screen for a meme...
worth pointing out, the screen shots for the desktop look Mac-like and have at least a 3D appearance, and not the 2D FLATTY FLATSO McFLATFACE every OTHER "modern" desktop is trying to clone...
"Minimalistic" can still look nice.
There is some merit, but I'd 100% be against this, purely for philosophical reasons.
The "woke" brigade on Reddit and elsewhere deem the usage of "GIMP", "gimp" and "gimped" to be extremely offensive, and should never be used under any circumstances. I refuse to give any ammo that those sanctimonious-vegan-hippies can proclaim as "success" to prove their righteousness.
I will obviously now crawl back into my cesspit, until the next opportunity arises!
I don't have an issue with GIMP. Not now, anyway, and always recommend it to people when they have a need.
But in the early days - when I first became aware of it (mid- to late-90s, I think), and in those times when you could be supplied with a virus or two on a magazine floppy disk - the name did make me a little suspicious, and even held me off installing it for a while.
I mean, I'm sure some people were also suspicious of the name, and actively sought to install it.
But the name of any package could have at least a small impact* on take up unless people are in the know.
Incidentally, El Reg covered this a while back where the code has forked to provide Glimpse in order to do away with the name connotations some people find offensive.
* Please note I am not criticising the name, or suggesting it be changed.
The "woke" brigade on Reddit and elsewhere deem the usage of "GIMP", "gimp" and "gimped" to be extremely offensive
Oh, I wondered why & so went searching and found a derrogatory term for someone that is disabled, & insult implying that someone is incompetent, stupid, & sex slave. All of these are different contexts from image editing - so what is the problem ? I assume that most of us understand that many homophones exist where words have multiple meanings ?
Of course: if I try hard I can switch off my brain's empathy circuits and view the world entirely from my own perspective and thus fail to understand that others are different and so become offended by their different use of words. But most of us are more intelligent that that ... are we not ?
I expect that some SJWs will down-vote this comment as they believe that everyone must avoid using words that their clique has decided are offensive and insist on pushing those memes much to the bemusement of others who were using them in a different context and had no intention of being rude about anyone.
I expect that some SJWs will down-vote this comment as they believe that everyone must avoid using words that their clique has decided are offensive and insist on pushing those memes much to the bemusement of others who were using them in a different context and had no intention of being rude about anyone.
It's not impossible, but most of the downvotes will be for your partisan virtue signalling.
"everyone must avoid using words that their clique has decided are offensive"
The latest is some people complaining that "breast feeding" is offensive to some trans people and think it should be renamed "chest feeding" so as to remove the "offence".
I'm not sure how that changes things since not everyone with a "chest" can feed a baby, but then again "breast" is pretty much a synonym of "chest" anyway in reality. Robin Red Breast anyone?
Additionally, I would imagine that a trans from female to male doesn't have a problem with the term breast-feeding while a trans going from male to female can't breast feed anyway. Maybe I'm missing the point?
Nothing actually is offensive. But people are sometimes determined to believe things are offensive. Let's take responsibility for own thoughts , values and self generations of emotion. You are self authored creature your responsible to yourselfand physics for the way you react to things. Offensiveness is not an objective property of matter. Its reflection of the way the person who is offended chooses to view the world. Being offended shuts down the side of your brain that is intelligent, to revel in a emotional distortion that is not pleasant. This can be fairly interpreted as a habitual mental act of self harm.
Just because other people react a certain way does not mean you have to. The more that offends a person the more limited the person is. One can oppose something out of pragmatism with out losing perspective and becoming emotionally compromised. Their is no need to practise the self deception of projected offense.
Social standards are no excuse for not personally Testing concepts for fuction and alternatives before accepting them as exclusive or authoritative. Are you a person or are you a fleshy unconscious robot?
> … labelling …
+1 to useful metadata.
Here with KRunner in KDE Plasma, seeking 'image' presents 'GNU Image Manipulation Program' and other relevant applications, with their icons.
-1 to misrepresenting the name or icon. I dislike that using Falkon in helloSystem 0.4.0 (0D26) does not present the name or icon of Falkon in the dock – this might have been mentioned in chat, but there's not yet an issue for it.
Also, the absence of the application name from the global menu.
NEXTStep, Mach microkernel and BSD userland. CUPS was never strictly Linux, nor was Webkit. Darwin, the underlying glue behind MacOS, is closer to FreeBSD than Linux, licensing being the motivation; it all depends on your definition of "free" as, for some of us - Cupertino included it would seem, having to release your mods as source isn't freedom at all.
Meh, whatevs, ancient history.
* Long-time *BSD user, jumped ship to De[bi|vu]an when the project became infested with divas.
Webkit would have nothing to do with Linux as it was a wholesale lift of KHTML (the K coming from the usual source). If memory serves Apple initially would not contribute back to the KHTML project (in fairness the LGPL did not require it) but continued "go on do the right thing" public pressure resulted in Apple doing a massive there-you-go-then undocumented code dump.
"having to release your mod as source isn't freedom at all" fair enough but the you are all free not to use the code you are modding too. It doesn't seem to be serfdom to acknowledge that you got something so you should give back. In general that seems to be called "society"
The dock is pretty awful but seeing as I have it set to autohide it rarely figures. Not that I'm cheerleading for MacOS design: it has good phases and bad phases (skewomorphic, yuck!) with the control panel currently a real collection of design ideas from the last twenty years – I haven't switched to MacOS 11 which is supposed to have cleaned this up a bit. The point being that most UIs borrow and inspire though I've yet to see a GTK-based one I didn't hate.
I assumed he was referring to internal parts rather than to the UI (and to the wider open source community than to the kernel, Linux). And misusing the verb 'stolen' unless he's alleging that Apple has used open source code without redistributing it?
Those details aside, I guess:
Apple adopted CUPS after its Linux-world debut, but eventually hired the developer and bought the code. Though it's still also open source.
WebKit is famously a fork of KHTML, but is also open source and has itself been forked onwards.
Apple used to maintain various GCC patches but upstreamed as much as it could; it hired the creator of LLVM and has been very active in advancing that, whilst also originating Clang and providing that as open source.
Then I guess there's everything in the terminal that Apple ships, from zsh to ls to python to everything else — though what proportion of that is from BSD rather than GNU I wouldn't like to hazard a guess.
So, yeah, Apple has benefitted hugely from engaging with the open source world. I don't agree with 'stolen' as no licenses have been broken and important things have been contributed even where licensing doesn't require it, e.g. Clang.
But you can't imagine macOS being anything like it actually is if it weren't for the large pieces of open source software it contains.
There have been other attempts to copy macOS design ideas on the Linux/BSD desktop: elementaryOS was one, PureOS another, Budgie/Solus yet another. But they are very niche; and unless someone like Canonical or Red Hat takes them up, they're likely to stay that way.
Makes me think of https://xkcd.com/927/
Personally I always thought Ubuntu's Unity desktop was one of the best rip-offs of OS X. Hoi polloi didn't notice because, shock, the dock-analogue was on the left, not along the bottom. (This being the sort of advanced functionality you probably can't do in Elementary OS, which doesn't even have menus in the menu bar.)
Unity was a better OS X than OS X: you could see how many app instances you had running, you could open new ones with a middle-click, launch them with Control+digit (like WinXP, not that most Windows users ever noticed, either.)
> Makes me think of https://xkcd.com/927/
… which makes me think of this, which referenced the same cartoon:
ActivityPub - one protocol to rule them all? - Dennis Schubert
https://schub.wtf/blog/2018/02/01/activitypub-one-protocol-to-rule-them-all.html
– and the 2019 follow-up, which I just discovered.
The idiocy is on everybody else's part. See Bruce Tognazzini and scroll down to Question 5.
As he points out though, part of the problem came from lawyers.
If I'm reading it right, global menus have been shown to be faster because the user accelerates the mouse pointer to the top of the screen, where it stops. They don't have to slow the cursor down, as they would if aiming at a window's menu bar
(It's like travelling to the moon - you can get there faster for the same fuel if you're happy to use lithobraking instead of retro rockets to scrub your velocity)
Having thought more about the article it's clear that he's only concerned with optimising one thing - speed to hit controls. There are more factors to be taken into account such as maximising use of screen space.
E.g. labels on menu bar buttons good because it's a bigger target to hit and therefore faster. In fact the space taken up by the larger buttons might be better used to display more on screen which be useful during all the time that the user isn't thrashing about the screen trying to hit buttons. For most users that's probably most of the time.
"global menus have been shown to be faster because the user accelerates the mouse pointer to the top of the screen, where it stops ..."
... unless you have 2 screens, and the second (larger, external) is above the laptop screen, so as to be in-line with the keyboard. Then, the global menubar is in the middle of your desktop, not linked to any window and hard to hit. Great
Tricky quiz as it assumes experience of both Mac and Windows and even as far as Windows goes I haven't really used it in an age and even then a W2K instance in a VM for one application is about it. However question 4 is spot on as far as KDE is concerned. The unhide location could be set to a corner up to and including V3, then it just became the entire edge and, exactly as he describes, it's too easy to hit unintentionally. It's a design choice that should be rolled back.
Piffle. Stuff and nonsense. It was just a way to avoid an Apple look-and-feel lawsuit. In the early days, everyone put it at the top of the screen, because of Fitt's Law. Amiga OS -- but you had to right-click to open the menus. GEM, but they opened themselves. All just ways to dodge litigation.
Only Acorn and NeXT put thought into it. NeXT put them into the corner and made them vertical, as the corner is nearly as easy to hit as an edge. Acorn went one better: the only menus were context menus, with a dedicated mouse button to summon them. Saved both screen space _and_ mouse movement.
NeXT's were really cool — you could tear submenus off and leave them sitting wherever you wanted permanently as regular buttons; I think the only disadvantage to Acorn's take was discoverability, but only in the first five minutes before you'd been taught what all three mouse buttons were for.
I love pie menus - though I only use see them in a CAD application. The pie menu is just so quick for for selecting a tool. Confused as to why they are not more common.
The same CAD application also let's you configure menus, tool palettes, 'ribbons', keyboard shortcuts... whatever is your poison, basically. It's almost as if they care about their users.
Amiga OS -- but you had to right-click to open the menus
But it wasn't particularly difficult to do as you didn't need to click before moving the mouse or when the pointer was inside the menu bar, you could just push the pointer up and hold right button down some time before the pointer hit the top of the screen, choose the menu option from the menu which would automatically open because the pointer was over the menu bar and the right button was pressed, then release the right button. Easier than the Mac's OS at the time I think.
> … link the menu to the currently live window? …
If you'd like to find/raise an issue for the global menu not showing the name of the application: I'll up-vote.
In parallel:
Make Chrome work with the global menu · Issue #29 · helloSystem/Menu
https://github.com/helloSystem/Menu/issues/29
Make Firefox and Thunderbird work with the global menu · Issue #37 · helloSystem/Menu
https://github.com/helloSystem/Menu/issues/37
An application's window should be in front following use of the global menu for the application · Issue #50 · helloSystem/Menu
https://github.com/helloSystem/Menu/issues/50
System tray (notification area) not working for some applications · Issue #96 · helloSystem/ISO — https://github.com/helloSystem/ISO/issues/96
From the article:
As you would expect from the creator of AppImage, application bundles play a part in the helloSystem design. The OS supports "simplified GNUstep-style .app bundles,"
There is also support for ROX-style bundles, ROX being another open-source packaging format, and for wrappers that allow traditionally installed applications.
Not much software is installed on first boot, but applications such as the Falkon web browser (as used by KDE, also Qt-based) download and install on demand.
This all indicates to me that either pkg is not used or is hidden under a "wrapper" and at least some apps are installed using specially created bundles or packages.
"wrappers that allow traditionally installed applications." is the closest indication that pkg may still be available and usable in this "distro" and that's fair enough, if true, but the article barely doesn't mention whether it's possible to do a pkg install without having to do it through a wrapper.
If it IS possible and DOES use the pkg system, then maybe the article needs updating.
In some ways it is. Especially on portable laptops. I also tend to use smaller monitors because I already have enough clutter on my desk.
But I suppose it isn't the screen space that is the issue, if the UI makes poor use of it, you also have to do a lot more travel to achieve things and it becomes slightly slower as a workflow.
Weirdly it also looks more consumer centric, like a TV or other multimedia appliance rather than a workstation class tool.
I wonder if anyone has ever done a serious usability study against Mac OS X vs Windows NT 4.0 UI systems? I know Sun Microsystems did one on Gnome 2 (a slight modification called Java Desktop System) for Solaris 10 as part of the CDE migration. The results of that did actually compare favourably against Aqua.
A bit. I know there are lots of people who have bragging rights with their trio of 54" monitors but even as someone who's not destitute, I'm not made of money and don't have enough space.
Still, not downvoting because I kinda sorta agree with you in some regards: virtual desktops provide "good enough" functionality, and efficiency doesn't matter if it's fugly, even speaking as someone who still uses fixed-pitch green-screen applications because I like the aesthetic (well, and functionality, e.g. Alpine for email). Er, I seem to have digressed; where was I...?
virtual desktops provide "good enough" functionality
It depends what you're doing. If you need a couple of documents open side-by-side, such as something you're working on and something you're referring to swapping workspaces doesn't cut it.
Yesterday, for instance, I was extracting dates from a PDF of medieval records to convert to conventional dates, tabulated in a spreadsheet. Workspace 1 had the PDF occupying the left half of the screen. The lower right had my own program for cleaning up rubbish OCR text from the PDF to paste into the spreadsheet. The dates there were along the lines of "Friday after the feast of St Barnabas[June 11]"*. The spreadsheet was in the top right, set to all workspaces.
Flip to WS 2 & I had not only the spreadsheet but a terminal open to run cal for the year concerned to find out what day June 11 was in 1309 or whenever. Also a jotting pad for working out some of the more complicated logic round the variable feasts (Easter etc).
Multiple desktops help but an essential part of the mix is being able to maximise use of the desktop to display the document being worked on alongside so much other stuff. Even the task bar is set to autohide to allow maximum use of the space.
* The original editor had provided most of the fixed feast days which cut out a lot of searching.
Ugh. I'm picturing *one* screen with *one* terminal window, a text file, a function, and a script.
$ cat feastdays.txt
year feastday date weekday
;The weekday here is for human readability. And I know about normalizing a database...
1309 Barnabas 1309-06-11 Wednesday
# define festivus(1309,Barnabas,Friday,after,1st) ...
# parse; if (found) {$answer<festivus(); echo $answer;} else {$date<prompt; echo $year $feastday $date $weekday>feastdays.txt; parse; $answer<festivus(); echo $answer;}
$ fest -- Friday after the feast of St Barnabas[June 11]
1309-06-13
It depends what you're doing. If you need a couple of documents open side-by-side, such as something you're working on and something you're referring to swapping workspaces doesn't cut it.
Well okay, "better than nothing" functionality, then. :D But y'know, thinking back to early in my career when development (er, such as it was for a junior programmer with bad habits) consisted of an 80x24 green screen using ed on a Unix with no job control. Then one day I managed to get a second terminal. Wow!
"Now a Windows NT 4.0 UI clone."
KDE, set the menu system to the original cascading menu style. Use the Windows 9x application style and some of the various downloadable styles for GTK apps. Download Reactionary decorations. Job done plus the options for multiple workspaces and the like. I have a desktop appearance that's changed scarcely at all in the last couple of decades end then only subtly and mostly for the better (except where KDE managed to screw up, especially on auto-hide).
> Honestly we really do not need another macOS clone. …
Honestly and more accurately, helloSystem is "Not a clone of anything, but something with which the long-time Mac user should feel instantly comfortable.".
https://github.com/helloSystem/hello#hello
I look forward to trying it out when it's past alpha status. Much depends on what applications I can get in binary packages though. It's cool that I have the option to compile from source, but I want my computers for other tasks than running GCC.
Shame about QT though, given their recent political issues.
... it wouldn't be the first time Aqua style desktops have appeared for Unix systems. The problem normally is that they're playing the notes without understanding the tune and/or they just rip off the Apple look and feel, sailing so close to the wind as to invite a cease and desist. To me this desktop seems to tick both boxes.
I also strongly disagree with the contention that a single pull down menu is the best. Personally I'd be fine with it on a small screen since it maximizes vertical space but as the screen size increases it just becomes annoying - my app is down here but I have to move the mouse all the way up there to click something and then go back down again? It just becomes an annoying brain fart, disconnecting the user from the task at hand. It has been tried on Unix too, most notably in the Unity desktop for Ubuntu and it just didn't work.
A better approach would be a more task centric UI toolbar and the menu is some kind of hamburger icon on the side. An app could even do what Firefox does - toolbar / hamburger by default but tap ALT and a traditional menu appears.
As I linked above, people with stopwatches have actual evidence that a single menu works best. Reason being, a window menu is quite a narrow target. Increasingly so as the screen resolution increases. OTOH the screen edges are effectively infinite in size. Unless you've configured your desktop environment like mine, that is. Then you can scroll past the edge into the next virtual screen. (I'm always flip-flopping on whether I prefer that or not. It's a pain when I do it by accident.)
In responding to this and to Greybeard's earlier reply to me (thanks!) I believe it was Unity which put me off top edge menus. The top limit to mouse movement (= less aiming required) is logical but DrXym's point about multiple small windows and large mouse movements was a stopper for me.
I wonder how much of this is simply 30 years of being exposed to menus-on-windows? Were the tests that Mr Tognazzini describes performed on naive users or with people who have used one or the other approach for years? I suspect the results might be different for different groups.
I can't help feeling that the biggest barrier to adoption of a new 'better' UI is simply familiarity with the old one - for example, I find the MS ribbon to be a largely unusable mess. But I am aware I am an old fogey; hell I even write in C so what do I know?
Nonetheless, I shall have a play with this project. You never know - I might like it!
"hell I even write in C so what do I know?"
Perhaps you know how to actually write a program that solves a problem rather than simply wasting time writing bindings to tie the underlying operating system to your "fancy cool modern" language of choice?
But I'm just a C++ developer so what do I know? (Other than wasting all my time trying to get RAII to play ball with the underlying native C libraries, dangerously close to a binding layer ;).
tldr; C is great. Its been a hard week wrangling C++ and Java!
If you wrap everything using std::unique_ptr with custom deleter, it's a lot less painful to wrap the underlying C libraries.
For MySQL (don't ask)
namespace std
{
template <>
class default_delete< MYSQL >
{
public:
void
operator()(MYSQL *ptr)
{
mysql_close(ptr);
}
};
template <>
class default_delete< MYSQL_RES >
{
public:
void
operator()(MYSQL_RES *ptr)
{
mysql_free_result(ptr);
}
};
} // namespace std
...
std::unique_ptr< MYSQL_RES > sql;
The problem is if the unique_ptr containing the statement object gets dropped *after* the unique_ptr containing the sql connection object you get a double delete. C is more than capable of stripping out the underlying data from under the smart pointers nose.
Holding onto your local data is only part of the challenge. I spend so much time crafting smart pointers in a way that they also hold onto their parent references up the hierarchy to stop them dropping in order to make things safe.
Chuck some callbacks in the mix and some already awkward deletion strategies and you have lots of fun. This issue is purely because our middleware doesn't provide proper C++ bindings of course ;)
I use macOS and Linux desktops, and prefer top-edge by quite a large amount. For the aiming speed yes, but also I only need space for one menubar on screen, which gives me more real-estate for multiple windows - useful on a laptop screen. Given most Window systems already have some form of status bar with wifi status, clocks etc up there, it's real-estate that's already in use.
Secondly, the UI doesn't become problematic when I have a wider menu bar than the width of my active window. Think small, widget-type windows: with local menus you often need a special type of UI for this sort of control, as the window isn't big enough. Not the case with top-edge windows, meaning things are more consistent.
I'm not a fan of screen edge menus for programs (although the "system functions" (Applications, Places, System) menus in Gnome 2 worked well, I felt).
Fitts' Law is important to remember for many things, but, for application menus, what might have worked well for menus on the smaller screens of 80s and early 90s computers doesn't work well for menus on large workstation monitors today.
If you have a large workstation monitor, it's usually because you have several less-than-full-screen windows open at once (for a good reason), and for any programs open on the right hand side, and especially bottom right, of my monitor, any time saving in being able to hit the screen edge is outweighed by the time (and physical motion required) to haul the mouse pointer all the way over to the top left. And for the most common use cases, I save even more time by using the keyboard shortcuts for the menu items instead anyway.
For those of us who prefer not to use a mouse, a hamburger menu is ... inconvenient. I get on fine with the old school pulldowns as they are keyboard enabled.
There is no one-size-fits-all best UI. The biggest UI problem I see is UI designers who interact with the system in only one way (theirs), and tough cookies for anyone who has a different method. There is an open source application I contributed to where the two maintainers used the system differently from the original author, each would constantly introduce UI bugs that only cropped up under an "other" UI paradigm, and 100% of bug reports were initially marked as "could not reproduce". Needless to say I am fed up with that application.
Aaaarrrrgh! Nooooo! Hamburger menu? No thanks.
I totally agree about the context menu behind a daft idea. I always found it very inefficient having to move to the very top of the screen for the menus when I used a Mac.
Personally I like each application to have a classic menu bar attached to its window and I'm dead against hamburgers on desktop OSes - it's an extra unnecessary click each time to get to the menu. With GTK3+ the GNOME team has deprecated classic menu bars (and might remove them in GTK4!) and they've converted most of their applications to use a fat title bar with a hamburger and other large buttons embedded into it (yuck!).
This is trickling, unavoidably, down into every other desktop that is based on GTK including MATE which I use, and XFCE which is really pissing of the desktop purists now.
I'm losing hope about it now - it's almost irreversible already :-(
> … fat title bar with a hamburger and other large buttons embedded into it (yuck!). …
True. Ugly AF, crow-barring of large icons without names into space that should be put to better use.
Are you married to those types of desktop environment, or would you consider KDE Plasma?
I'm willing to try KDE again as it's been a while but I am a bit fixated on using a no-frills UI which I think is almost perfect in MATE.
I was never keen on things like single clicking in KDE doing what a double click should do although I presume that's configurable?
The main problem for me if that there really isn't a mainstream web browser that's based on Qt, nor am I aware of a Qt based version of Thunderbird which I rate very highly as a mail client.
So if want to use Mozilla stuff, or Google stuff then it's all based on GTK and so will have all the new GNOME atrocities bundled in even if my main desktop environment were KDE.
One application that isn't part of GNOME but which has adopted the new toolbar layout is 'Remmina' which is a brilliant remote desktop client for RDP, VNC and others. When they first switched over they still have XFCE users a classic menu but then decided it was too much maintenance and removed it. So now it's got Client-Side Decorations, fat title bar with integrated hamburger and other buttons plus they've switched their icons from colourful bitmaps to monochrome SVGs which make it look terribly bland. GNOME/GTK could easily allow their menu building code to accept a user choice and render accordingly and Remmina could have allowed themes or a choice of classic Vs new icons. But they didn't. Some developer just decided they'd fundamentally change the user interface and all the uses just have to suck it up. That's the sort of thing we expect from Microsoft - I'm disappointed that some FOSS players are now taking the same approach.
Anyway... Next time I'm trying live USB images I'll give KDE a spin
PC-BSD was about the most complete attempt at a BSD-based desktop system. The last time I looked it suffered from the general problem of missing proprietary drivers for bits of hardware but overall it was pretty nice.
As for packaging apps: the ports system and the various port managers on BSD have always provided an excellent basis for this.
For drivers, you just have to manage expectations. For example FreeBSD supports vastly more hardware than macOS and yet this is a common complaint. A free operating system shouldn't be expected to provide better hardware support than a very rich commercial product (even though it does).
If a bit of hardware doesn't work on FreeBSD, rip it out and replace it with something that does. Older working hardware is almost free these days.
So, Thinkpad Laptop, HP Z Series desktop and pretty much any server. Can't go wrong ;)
Curious you should mention a Z series. I just bought an old one, which came with an old ATI FirePro graphics card.
I wasted several days of tinkering trying to get the thing to work with Linux (tried Manjaro and Ubuntu) but with no success at all - black screen as soon as I allow kernel modesetting. Scary messages about clock voltages being exceeded.
It was a pity because these old cards are supposedly very well supported by the old drivers. It opened a minefield for me because it seems that the most compatible graphics by far is Intel, which is what I've used on numerous laptops with no issues.
Nvidia seems to be proprietary drivers old, which means for some older cards you may have to downgrade X to an old version that still supports the binary blob driver (same seems to be true for the older proprietary ATI drivers as well) - this risks security as you then can't upgrade X.
So for my workstation I ended up with a cheap AMD card which worked first time out of the box using the open source 'amdgpu' driver but I've read plenty of accounts of folks having issues with some modern AMDs too so it was a bit of a gamble.
I'm sure BSD must be hitting all the same problems. I really don't understand why, in the 21st century, folks like NVIDIA still won't openly document their interfaces and allow for open source drivers... It's not like they're giving away the inner workings or blueprints. For want of a better analogy it's like a aeroplane manufacturer refusing to provide a manual of what all the switches do in have cockpit.
Yep, your experiences / views pretty much match my own.
With the Intel stuff, things are getting quite exciting with their recent interest in discrete cards (such as the Intel Xe). Yes, not the fastest but so long as I can get native resolution with an XTerm, I am happy haha.
I see NVIDIA a little bit like ImgTech's PowerVR. Fancy but useless for workstations. Don't waste another thought on them (that said, for my uses Nouveau is actually pretty darn usable, I am looking forward to FreeBSD and OpenBSD one day perhaps importing the source. I have a tonne of old inherited NVIDIA cards in cold storage).
For high performance acceleration, AMD is basically our saviour. For now.
When you say discrete cards do you mean they might put their graphics cores onto a PCI-E card?
That's what I'd ideally have bought this time round. The old HP Z620 with a Xeon doesn't have any means to support the CPU based Graphics as with their i-series processors so I have to use a card and it looked like there's basically only a choice between AMD and NVIDIA at the moment. This also seems to mean that the bulk of available cards are highly expensive 3d chips when all I want is a solid 2d desktop workhorse.
I noticed that there are some Matrox branded cards but I'm not sure if they're actually still AMD chips.
I suppose the problem now is that desktops are generally kept for either gaming or high end specialist stuff like CAD whilst your average user has moved to using a laptop for everything. That was/is me, but I wanted to offload some of my development stuff onto a beefier machine with more RAM.
Quote: "When you say discrete cards do you mean they might put their graphics cores onto a PCI-E card?"
Basically yup. Intel announced a while back they were getting back into discreate GFX cards. They are shipping an updated GFX engine with some devices now (Laptop OEM stuff). This is known as DG1 (Discrete Graphics 1), although it's a discreate chip, rather than a card, aimed at Laptops.
DG2 is apparently the discrete card, and these are in alpha currently, and are supposedly due for launch later this year.
Even if DG2 is only mid range, say equivalent to a 3060 or maybe a 3070, if they can get them in stock, they'll likely sell simply due to nVidia and AMD not have much stock anywhere atm.!
Sounds awesome! I'll keep my eyes peeled for this.
I'd heard that graphics cards are in short supply but the reality of it when I was trying to buy one came as quite a shock. Pages of "out of stock" and "coming soon" and very little available to buy now. I really wasn't expecting it to be quite so bad
If only Apple themselves still remembered that.
After the usual cautious wait for other people to find the initial bugs, I finally updated my own Mac to MacOS Big Sur recently.
Whereas before there were nice simple pull-down menus for choosing between WiFi networks, etc, Apple have now given us a rather iOS-like "Control Center" bulky blob which requires all kinds of extra clicketty-clicketty to open up its various nested contents (rather than just simply moving the mouse pointer down and up a menu). And I wonder how well it works for users who have special accessibility needs?
(Big Sur also has the rather strange feeling that it itself is a second-rate GNOME theme Aqua-knock-off, rather than the real thing: the slightly puffed up window titlebars just don't look/feel right (you now even have to click to open the search field in a Finder window, which is an extra unnecessary step), and neither do the frosted-glass and oddly vertically stretched out menus (I also feel that the translucency detracts from their usability, being not quite so contrasty/readable, and the fact that extra graphics rendering has to be going on for these unnecessary and backwards effects makes everything feel unnecessarily sluggish, responsiveness always being a key reqiurement for any interface).)
> … MacOS Big Sur …
Mac OS X Mavericks was the end of the line for me, after twenty-something years.
Apple span no small amount of bullshit around Yosemite. Too much of this BS was swallowed with glee by the unquestioning masses – this included the massed audience at WWDC.
A web browser that can no longer show the title in its title bar? No, thank you, Safari. Goodbye, Mac.
That's a good point. Because MacOS only supports "click to focus" (sigh), I just have to get used to that, sadly (unless there's any third-party software out there?).
But when I'm using Linux and go through one of my occasional "let's try a different window manager" moods, and am trying one with a screen edge menu, "sloppy focus" works as a reasonable compromise for this: you are given just enough time to skite from the window to the menu before the mouse focus decides to focus on any of the other windows en route.
> Already been done …
No, the two are quite different.
helloSystem emphasises the global menu. It seems that elementary OS will never have this feature.
And so on.
https://github.com/elementary/wingpanel/issues/195 duplicate of https://github.com/elementary/wingpanel/issues/81 …
It's easy enough to make KDE resemble Mac OS X, visually.
A quick search with Google found, for example:
Making Linux look like macOS with these easy tweaks! ▶ Making KDE Plasma looks like macOS
https://tipsmake.com/making-linux-look-like-macos-with-these-easy-tweaks#mcetoc_1dd28saph8
More exotically, although BackSlash is not recently updated:
Familiarizing with the Desktop | Docs
https://docs.backslashlinux.com/desktop.html
That is not a fail, that is a win.
The guy specifically states it's about returning to the simplicity of 90s computing, including GUIs which actually followed rules which were based on useability studies. You learnt the GUI, you could use every piece of software there was.
I realise this sounds like "old man yells at cloud" but I can't see how present-day GUIs are an improvement on that. Practically every piece of software is different today, huge areas of screen estate are used up for no good reason, the response time is abysmal, items in lists move around just before you click on them so you click on the wrong thing, badly contrasting colours and fonts hide information from you, not even the OS designers can follow their own rules. Bah humbug.
What a waste of time and effort that could be better spent on more useful things... there is no shortage of desktops out there by a long shot. BSD's problem isn't its desktop - it is its lack of application support, and a desktop will NOT make third party vendors support it any better than they do now, regardless of whether those are open source vendors or closed source.
There is no shortage of desktops, but most are based on KDE and Gnome, which manage to fail spectacularly, as he convincingly argues if you read through it all (link posted above by another commentard).
You can use them but have to constantly patch them to make them usable and you are held hostage to their next crazy idea, or you can roll your own. I guess developing his own desktop on a BSD is in some ways a challenge to Linux to sort its act out when it comes to GUI usability.
As much as I applaud anything that brings some attention to the *BSD world, this really isn't the thing stopping me from using it on the desktop. Having recently tried to do so again, here is a numbered list of what stopped me dead in my tracks.
1) Woeful Wi-Fi support. FreeBSD does not have any support for wireless AC, which means that no card newer than the N generation will work. This was an absolute dealbreaker.
2) No bluetooth support to speak of. I use this for audio quite frequently, and it not being there was a PITA.
3) Lack of support for containerized applications. This one isn't really FreeBSD's fault, since the issues come from that entire ecosystem relying on Linux-specific functionality, but it made it hard to use it for my day job.
4) Lack of support for USB HID devices in the udev emulation, which made USB peripherals annoying to get working in some cases.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
