back to article How Windows got to version 3 – an illustrated history

Windows 1 and 2 flopped almost as badly as OS/2 did. How did Microsoft stage one of the greatest comebacks ever with Windows 3? Earlier this month, we took a look at how Microsoft learned important lessons from the failure of OS/2, even though less than a decade later, it had already started forgetting what went wrong and why …

  1. Dan 55 Silver badge
    Devil

    Conspiracy-led nonsense

    "Microsoft intentionally sabotaged OS/2 because it wanted to sell Windows!"

    That's the motive. Why wouldn't they want to sell their own OS instead of dancing to IBM's tune?

    Now for it not to be a conspiracy theory (although it was a conspiracy) we just need some evidence...

    1. diodesign (Written by Reg staff) Silver badge

      Re: Conspiracy-led nonsense

      I think you need to look at the timeline. The article's argument - which we can make clearer - is that Microsoft didn't set out to sabotage OS/2 to benefit Windows. Why would it bother in the first place?

      As you entered the 1990s and Windows 3.x was looking like a goer, then yes, there is court-submitted evidence that appears to show Microsoft wanted to push its Win OS. By that point, Microsoft and IBM had pretty much broken up, which we all know about.

      C.

  2. Anonymous Coward
    Anonymous Coward

    An enjoyable read, once I managed to get past "Kalinina anatomizes the first three Windows versions". A perfectly cromulent word which, according to the OED, was quite common in the 1700s. Not so much since then. I'm off to dissectivize some more articles.

    1. Bebu sa Ware
      Mushroom

      "Kalinina anatomizes the first three Windows versions"

      At first glance I had Kalinina atomizes the first three Windows versions. If only ... and all the rest too, please.

      Anatomy, as one demonstrator put it, was just Greek for cut (τομία, tomia) up (ἀνα, ana-) ... classicism in a scalpel wielder was remarkable. Atom likewise is just Greek for cannot (a-) be cut. :)

      Postmortem might be an equally appropriate description of Kalinina's opus but for the inconvenient truth that subject still haunts us.

    2. Doctor Syntax Silver badge

      It may have been unfamiliar to you. It's certainly a word I recognise, understand iand have met before n that context even if I wouldn't use it myself - I'd probably just say "analyse".

    3. Liam Proven (Written by Reg staff) Silver badge

      Thanks! Glad you liked it.

      > once I managed to get past "Kalinina anatomizes

      TIL that the word for "to dissect and display the anatomy of something" is an obscure word. Huh.

    4. MyffyW Silver badge

      Dare I suggest "anatomise" for us dwellers in Airstrip One?

  3. AndyMTB

    Thanks Liam!

    Thanks for pointing out this article, Liam. Thoroughly enjoyed it, and brought back a lot of memories! Am now off to to find some of the documents mentioned in Sources.

    1. John Brown (no body) Silver badge

      Re: Thanks Liam!

      Yes, it was fun! I lived through it :-)

      I was teaching IT at the time, started with BBC Micros, TRS-80 model IV's and various CP/M based kit. My first experience with Windows was the run-time only version of Windows 1.0 that came with Aldous Pagemaker. Likewise, run-time Gem that cam with Ventura Publisher. We never got full fat Gem, but did get the full fat Windows 2.0 and the SDK to play with. Windows 3.0 on a proper 386 was a revelation at the time in terms of PCs. Then I moved into a whole different area of IT related shenanigans.

      1. ICL1900-G3 Silver badge

        Re: Thanks Liam!

        I had full fat Gem on my Atari ST, but soon installed MINIX, much more fun ... and useful.

      2. Andrew Scott Bronze badge

        Re: Thanks Liam!

        interesting article and the pointer "inside windows 3" also. As i remember, we were primarily dos and ran everything from a menu system on a network. At some point we were making win 3 available from the menu to run some apps, but mostly stayed with dos for word processing, spreadsheet, and calendaring apps, all dos based. No '386 machines at the time. i argued for the sx but was shot down on the premise that it wasn't a true 32 bit cpu and was limited. It wasn't until a gift of 15 '386sx machines from dec that windows became more important. That and the packet driver enabling the first use of mosaic. After that no stopping windows 3. still ran pc's for some things for a number of years and had people running xywrite in dos boxes into the 2000's.

    2. bombastic bob Silver badge
      Megaphone

      Re: Thanks Liam!

      Thia was pretty comprehensive, but it's missing some important points in the evolution from OS/2 PM 1.2 to Windows 3.0. Solitaire was *mentioned* but its marketing importance OVERLOOKED.

      Aside from the under the hood features there were TWO major factors... no, THREE, THREE major factors, that made Windows 3.0 the huge success it became.

      1. Unlike OS/2 it ran on just about EVERY PC. It did NOT need a boatload of "OEM Support". I had taken a class in OS/2 PM Programming at a local city college and understood the SUPERIORITY of OS/2's API. Unfortunately, ONLY PS/2 machines seemed to work with OS/2 PM 1.2 (and the 2D FLATNESS of the earlier 2.0-like UI made 1.1 UNDESIRABLE!)

      2. With Windows 3.0's "New shiny 3D SKEUOMORPHIC LOOK" (something HORRIBLY LACKING since Windows 8) , as well as DOS compatibility, it just LOOKED COOL running in the store, AND for your old DOS applications, it was a bit like a DOS upgrade. I actually did BOTH at once. It was like getting a Model A Ford that came in colors OTHER than black.

      3. Windows 3.0 Solitaire. I do not ,now HOW many copies of windows weer sold JUST FROM THIS, but seeing it on the monitor and hearing people comment sticks in my mind.

      and maybe a 4th... Micros~1 actually LISTENED to USERS *AND* DEVELOPERS!!! The OS itself showed this on every level, and CONTINUED to do so, minus ME, Vista, and ANYTHING AFTER 8.0 !!!

      MOSTLY it was "pretty". But SOMEWHERE along the way, some "MENSA CANDIDATE" (read: Brain Donor) **FELT** that EVERYONE! ELSE! MUST! CHANGE!!! So a "New, shiny?" (read: a horrible plastic fake-look cheaply stamped-out pile of CRAP) interface was FOISTED UPON US ALL, by a man who was LATER FIRED [though his legacy lived on... but why?]. That became the TIFKAM interface designed for a low-CPU-horsepower limited resolution PHONE DISPLAY that really NEVER HAPPENED, thus *RUINING* the customization and PLEASANT appearance we all wanted, and ACTUALLY HAD in 7.

      In short, it's the rise to 3, the evolution to 7, and the FALL of a once decent desktop+OS into a CRAPPY look that is a LOT more like the Windows 1.x and 2.x that were SO UNSUCCESSFUL!!! And even FORCING us to use it gains NO FANS. We mostly have NO OTHER CHOICE.

      Those of us grey hairs who have WATCHED THIS must acknowledge this truth. Unfortunately Micros~1 NO LONGER LISTENS.

      1. Terry 6 Silver badge

        Re: Thanks Liam!

        Never thought I'd upvote BB*. But there you are.

        *And those bloody CAPITAL LETTERS do make it upleasant to read.

        1. ICL1900-G3 Silver badge

          Re: Thanks Liam!

          @Terry... My thoughts exactly!

      2. Liam Proven (Written by Reg staff) Silver badge

        Re: Thanks Liam!

        No big disagreements here. The couple of things I'd note are:

        > Windows 3.0's "New shiny 3D SKEUOMORPHIC LOOK"

        The 3D buttons in Windows 3.0 were taken directly from OS/2 1.2.

        See for yourself:

        https://winworldpc.com/screenshot/c3b23b57-c3a7-1b19-11c3-a4e284a2c3a5

        Released 1989, the year _before_ Windows 3.0.

        Secondly, neither Win 3.x nor OS/2 1.x, or >= 2.x needed activation, license keys, serial numbers or anything.

        Windows 95 and NT 4 needed a key, but "11111-1111111" worked fine.

        XP was the first with online activation.

        OS/2 was just such a pig to install, and had so few drivers, that there was no _need_ to copy-protect the thing as well.

      3. Efer Brick

        Re: Thanks Liam!

        Our chief weapon is surpiiiise, surprise and fear...

  4. Kev99 Silver badge

    Way back when, I used a an 80286 box running DR-DOS, later switched to MS-DOS 5. My main apps were Quattro Pro, pfs:Professional Write, and pfs:First Publisher. No fancy windows, no unnecessary bloat, just basic, it works, software. I ran a multi-million dollar county with those three bits. QP had tabs which I could easily link. PW and FP could easily work together. With FP, I created annual budgets for our county, with each year on it's on page. I used QP to prepare tax settlements which would easily run into several files because of how the tax laws were written, and then run a macro to pull everything together into a single file. Another macro and the 30 or so settlement sheets were created for printing out on our Panasonic wide carriage dot matrix. (We tried using our IBM Selectric as the printer but it was slower than Fred Flintstone's printer. But it was great for when letters needed to done.)

    Basically, for 90% of I needed to then and since these packages worked fine. Most of what's in Word, LibreOffice, et cetera I never use.

    1. NATTtrash
      Pint

      These packages worked fine...

      Basically, for 90% of I needed to then and since these packages worked fine.

      Indeed, very true. And that is something that never stops to amaze me: listening to "others" nowadays, it all seems to hinge on "The Only One"™ and no alternatives "exist". Or should even be considered!

      Reading through Nina's piece I had vague (Age? Booze?) memories of GEOS on the C64. Which did the job too, making me happy since I could not afford a PC yet. It all worked fine, despite the PITA of printing the final version of your project report on a 8 pins Oki printer... the night before you had to hand it in... ALL through the night..... Niiiiii, clank, niiiiiiiiiii, clack, niiiiiiiiiii, clank. Luckily my bed was located above the cupboard with the Oki... in a very careful configuration so the paper didn't tangle/ jam...

      But yes, I agree. Later managed to get an IBM 30 out of a bust office auction. And WordPerfect 4.2 on it was... well, perfect. The WYSIWYG graphics of 5.0 were just "unnecessary indulgence". Quattro Pro did everything I do today too. CorelDraw let me make those images that looked very nifty in Harvard Graphics. And yes, ReferenceManager gave you that feeling automation really, REALLY saved you SO MUCH time and work. (Who of you have done, like me, their in text literature refs by hand,,, and then had to reorder all subsequent references and their list at the end 10.000 times, because you added reference #24? ;)))

      Microsoft got lucky

      @Liam: THX for the piece Liam! Let me add a thought here: I also think it was (just) very clever marketing. What I do remember from those days too that Microsoft "allowed" people to copy and use Windows everywhere. Everybody had a copy of Windows on their home computer. The running argument was that if people used it at home, they would nag their boss to also get it at work (where the MS money is made). It was all over the place, everybody gave it to everybody. And it was all good. And possible, since all could use it on their PCs. Contrary to the (way more expensive) Apple (software). And thus MicroSoft applied the "dealer-dependancy" model which works up to this day: give it away for free, get them hooked on it, and then charge them for it.

      But yeah... "No fancy windows, no unnecessary bloat, just basic, it works, software". Happy days...

  5. Doctor Syntax Silver badge

    The mention of HP NewWave is is worth following up. The really big advantage from my PoV was that it added a spill chucker to basic text processing without having to fork out for Word. The real advance, however, was what they called object orientation which was essentially being able to click a data file and have it automagically opened with appropriate application. AFAICR prior to that you used program manager to open the application and then the application to find and open the file. It seemed to depend on a l-o-t of files which strange looking names.

    AIUI it became the basis of the functionality in W95 with the strange looking files being corralled into the registry with equally strangely named keys. Certainly the copyrights of W95 included HP along with Berkeley for the network stack.

    1. Liam Proven (Written by Reg staff) Silver badge

      > being able to click a data file and have it automagically opened with appropriate application. AFAICR

      Eh? Windows had file associations in 3.0. I think Windows *2* had them.

      It just links the 3-letter file extensions to a binary. This is the original use of the registry; it _registered_ which program opened which file.

      .DOC -> c:\msoffice\winword.exe

      .XLS -> c:\msoffice\excel.exe

      ... and so on.

      You did not need NewWave for that.

      What NewWave did was things like long filenames, descriptive labels on icons, and an equivalent to symbolic links -- later called "shortcuts" in Win95 and "aliases" in Mac System 7.

      NewWave looked very impressive but none of my clients back then were willing to pay for it. HP later ended up with an enterprise email system, HP OpenMail, and it could have made a killing early on if it had bundled, say, a bunch of enterprise tools together:

      * email

      * workflow (forms you could send, attachments, etc. -- that stuff was hard in the early Internet era)

      * a network directory that deployed your apps to workstations

      * on the back of that, an enterprise app launcher like this.

      But only Novell had even some of that much vision and it was too busy trying to compete with DOS and Windows to seriously invest in _improving_ Windows. Noorda got distracted with "let's out-Microsoft MS" before he lost his marbles.

      There _were_ rivals and there _was_ room to build something amazing on 16-bit Windows clients to a grown-up 32-bit server, but nobody put all the pieces together until it was too late.

      If IBM grasped the nettle as soon as Win3 shipped, bought HP and Novell, it could have sold OS/2 servers running Netware for OS/2 and NDS, and grown-up email and so on. But it didn't. Novell and IBM both tried to compete with MS instead of building on it.

      HP at least had the wit to suck up to Redmond and piggy-back. That's why it's still alive.

  6. IvyKing Bronze badge

    Paterson vs Kildahl

    Tim Paterson's libel suit was tossed because the judge ruled that Paterson was a public figure and thus his suit needed a higher burden of proof, i.e. malicious intent, for a judgement in favor of the plaintiff. Tim would likely have prevailed if he was not considered to be a public figure.

    86-DOS copied CP/M's API but not the code in the same way that Linux copied the UNIX API but not the code. CP/M was written in PL/M, 86-DOS was written in SCP's assembler and thus would involved a lot more work than simple copying.

    1. Falmari Silver badge

      Re: Paterson vs Kildahl

      @IvyKing "Tim Paterson's libel suit was tossed because the judge ruled that Paterson was a public figure and thus his suit needed a higher burden of proof, i.e. malicious intent, for a judgement in favor of the plaintiff. Tim would likely have prevailed if he was not considered to be a public figure."

      Not a chance, public figure or not the the Judge would have found in the defendant's Evans favor.

      "Plaintiffs fail to provide any evidence regarding 'serious doubts' about the accuracy of the Kildall chapter. Instead, a careful review of the Lefer notes... provides a research picture tellingly close to the substance of the final chapter." Judge Zilly*.

      *https://www.theregister.com/2007/07/30/msdos_paternity_suit_resolved/

      BTW great article Liam.

  7. Roopee Silver badge
    Windows

    Brilliant!

    Thank you Liam, another great article.

    I once, just the once, saw Windows 2 I use in an office. I’m young enough that my first PC had Windows 3.0 and came with a mouse. I learnt about mice doing Computer Science A level, but we never saw one. We learnt more about punched cards and paper tape, and the teacher brought in examples. I’ve never seen either in use since.

    My first PC, bought in 1990 (an Austin iirc) was going to be a 286 - I agonised about the choice; they were 25% faster than a 386SX for the same (huge) price, >£1000, - but at the last minute the 386SXs were upgraded to 20MHz for the same money so it became a no-brainer.

    The thing I remember most is the many unhappy hours I spent trying to optimise memory for multiple configurations. Running DOS programs such as SuperCalc and Lotus 1-2-3 in a window was sooo slow it was worth the effort to dual-boot... oh, and how frequently and frustratingly Windows crashed/locked up without warning.

    But I still thought it was wonderful! How times have changed...

    1. CorwinX

      Re: Brilliant!

      Back in the mists if time, "computing" was taught as an "Advanced Mathematics" course.

      And punch cards and tape did indeed feature.

      I was literally taught how to "patch" punch tape by cutting out a bit of the tape and inserting a new bit.

      You do all know where the phrases "Cut & Paste" and "Patching" come from?

      1. John Brown (no body) Silver badge

        Re: Brilliant!

        "You do all know where the phrases "Cut & Paste" and "Patching" come from?"

        Yes, and we even had hand punches, similar two but smaller than the splicer, for those minor correction where you just needed to add an extra hole or two to change a value, or stuck some tape over it when the holes were in the wrong place and re-punch those few characters you'd managed to fat-finger, similar two but smaller than the splicer :-)

        Yes, kids, we learned to read the punched tape. Most fun was with 5-hole tape where you had to know if the characters were before or after a shift code, ISTR there were two different shift codes, but certainly no lower case and limited non-alphanumerics anyway, not sure if it was even ASCII we used then :-))

        And.....OH WOW! I just Googled to find out what the 5-hole code was called and the very first link pointed me at the first tape punch I ever used at school! We also had the teletype version, two or three desk sized units, ie they didn't sit on a desk, they WERE the desk :-) Note especially the odd shaped box at the far left with the black rubber hose going into it. That's the "bit bucket" where the "waste" of punching all those holes ends up.

      2. MJB7

        Re: Brilliant!

        "Cut and paste" don't come from paper tape (though patching may do). They come from graphical layout where pictures and typed text were "cut" (with scissors) and "pasted" (with glue) into a layout before reproduction. ("copy" was a quite a _lot_ more difficult in those days).

        1. Doctor Syntax Silver badge

          Re: Brilliant!

          When drafting text it was "cut and staple". The word processor has saves a lot of punctured fingers.

          1. VicMortimer Silver badge

            Re: Brilliant!

            It was literally paste. During the phototypesetting days, text and image blocks were glued to markup boards with a waxy glue. It was removable and repositionable. You then take a photo of the completed page and turn that into a printing plate. Much easier than working with blocks of linotype castings.

            1. Anonymous Coward
              Anonymous Coward

              Re: Brilliant!

              That explains cut and paste.

              Patching, of course, is what your Mum did to the knees of your trousers (pants) after you'd been out playing in them. :-)

            2. gordias

              Re: Brilliant!

              Ah, you mean Cow Gum.

            3. Doctor Syntax Silver badge

              Re: Brilliant!

              I know that layout was done by cutting and pasting. I can also assure you that when drafting the original text there was a lot of cutting out bits and stapling then onto fresh sheets of paper with scribbled annotations. Staples were less messy than paste for that purpose. When it got too much the whole lot was send to the departmental typist for a clean copy so we could start again.

              I remember also getting proofs as galleys which, I think was hot metal. And the subsequent page proofs with the added fun of ensuring that any corrections made at that stage didn't alter the pagination.

              One of my first jobs as a research assistant was checking page proofs for a Proc.Roy.Ir.Ac. (Irish equivalent of ProcRoySoc) and finding a whole load of numerical errors converting from imperial* to metric which, on checking further, went right back to the guy's PhD thesis and had gone through every other stage unnoticed.

              * Yes, imperial measures. The work had been done with a surveyor's staff marked in decimal feet. My first essay into the world of punched cards and Fortran was combing the levels from that staff with metric from the Hiller borer to tables that I could use to plot on old-fashioned graph paper with old-fashioned Indian ink. Happy days and a strong hint that this computing stuff deserved a closer look.

    2. Dan 55 Silver badge

      Re: Brilliant!

      The thing I remember most is the many unhappy hours I spent trying to optimise memory for multiple configurations. Running DOS programs such as SuperCalc and Lotus 1-2-3 in a window was sooo slow it was worth the effort to dual-boot... oh, and how frequently and frustratingly Windows crashed/locked up without warning.

      Those of us who migrated from other systems (read: Amiga) wondered if this was supposed to be the future of computing or the past...

      1. Anonymous Coward
        Anonymous Coward

        Re: Brilliant!

        Those of us who didn't (Macintosh) know it was a bastard offshoot of computing.

  8. Baudwalk

    Love this....

    ...old shtuff.

    My favourite books in this area are:

    Barbarians Led by Bill Gates

    On the Edge: the Spectacular Rise and Fall of Commodore - (My copy is from 2006. Later versions have been re-edited and split into two volumes. I don't know if they're better or worse.)

    UNIX: A History and a Memoir

    Any other recommendations?

    1. captain veg Silver badge

      Re: Love this....

      Hard Drive: Bill Gates and the Making of the Microsoft Empire

      -A.

    2. Toastan Buttar

      Re: Love this....

      "In the Beginning was the Command Line" by Neal Stephenson is still worth a read, despite its age.

      https://web.stanford.edu/class/cs81n/command.txt

    3. Daniel M

      Re: Love this....

      An unreformed fan of WordPerfect, I would like to suggest Almost Perfect

      http://www.wordplace.com/ap/

  9. CorwinX

    A masterful summary sir.

    Being an old git I remember highmem.sys, tweaking Config.sys with LOADHIGH, figuring out which order to load the drivers...

    I think my record was 560-ish Low Memory, with a working TCP/IP connection. In DOS.

    Then W4WG came out and changed the playing field.

    1. Tim99 Silver badge
      Windows

      My personal favourite was that if someone had "upgraded" a program written in dBase III to dBase IV there was not enough available memory to run Novell's IPX/SPX network. In the end we produced a floppy boot disk that the punter had to use without networking. They got their work done, and either removed the floppy and booted to the network, or used sneakerware to transfer the file.

      I wrote much of my (similar) stuff in R:Base - Wikipedia (Caveat: I wrote some of this) - Which was, in my not so humble opinion, superior. I had a couple of R:Base applications that ran well with thousands of rows for 5-20 concurrent network users. The secret was to load the program files onto each client, and just share the data files. In the mid-1980s, Microsoft did not have a database, it licensed R:BASE to sell in Europe and the UK.

  10. Gene Cash Silver badge

    To make sense of the industry today, you need to know where things came from and how they evolved

    Heh. Had someone at work ask "why does Linux have /bin and /usr/bin?"

    "Oh because the PDP-11 disks were too small to contain the entire OS needed to boot"

    "Wot. Stop bullshitting me..."

    "LOL... oh you sweet summer child..."

    1. bombastic bob Silver badge
      Linux

      To make sense of the industry today, you need to know where things came from and how they evolved

      In FreeBSD anything NOT actually part of the OS goes into /usr/local/xxx and things in /bin and /sbin (vs /usr/bin and /usr/sbin ) are the core features needed to make the OS work at all... generally speaking. "man hier" explains

      Unfortunately many Linux distros use a symlink for /usr/bin (or is it the other way round...?) to /bin (similar /sbin) and dump EVERYTHING in there, even some obscure package you install from outside of the distro.

      I guess that's sorta consistent with the PDP11 reference you made...

      1. Liam Proven (Written by Reg staff) Silver badge

        Re: … you need to know where things came from and how they evolved

        No, Bob. You need to know your history a bit better. You're seizing on later emergent stuff, adaptations to lack of disk space, and trying to post-facto justify them as part of the design., They weren't.

        You _need_ to read this.

        https://lists.busybox.net/pipermail/busybox/2010-December/074114.html

        The plan was simple. All the bogus stuff about "admin stuff goes in /sbin" and so on is cargo culting -- trying to justify bodges and lash ups as if they were planned.

  11. Gene Cash Silver badge
    Pint

    "built-in automated plagiarism generator"

    Smack, there it is. That's going on a plaque.

  12. Gene Cash Silver badge

    Dead links

    It's amazing how many failing links there are in both articles. And all of them return WordPress error messages.

    I do wish she hadn't hidden so much important stuff under the little triangle things.

    1. Liam Proven (Written by Reg staff) Silver badge

      Re: Dead links

      > It's amazing how many failing links there are in both articles. And all of them return WordPress error messages.

      "Both"? I linked to dozens. What two are you thinking of?

      I checked all the links. All my links work; I checked before publication. Where needed I use the Wayback Machine or something.

  13. 45RPM Silver badge

    The first GUI I ever used was Gem 2 on my first PC - an Amstrad PC1512. It sucked so hard that my desk was permanently dust free. I ignored Windows until I got a 386 with Windows 3 - I remember being oddly impressed by the 3D aspects of the user interface (and unsurprisingly unimpressed by the poor usability). It was fine as long as programs were installed by an installer program (otherwise you needed to fart around with PIFEdit) and provided you didn’t expect to launch a program by double clicking a document or use drag and drop etc. Otherwise it was a bit of a poor effort - and it didn’t really get usable until Windows 95.

    But by that time I’d given up on the mess of Wintel and moved to the fruity side.

  14. navarac Silver badge

    Happier Days

    It is just a great pity that Windows has degenerated into the mess that is Windows 11. To my mind, the crap started with the disaster that was Windows 8. Better to have gone back to 7, and then forward from there, but what do I know? Now I am happily using Linux; thank goodness I don't have to put up with Microsoft's shenanigans any more.

    1. Doctor Syntax Silver badge

      Re: Happier Days

      "To my mind, the crap started with the disaster that was Windows 8."

      I'd put it earlier - XP that wanted to phone home.

      1. VicMortimer Silver badge
        FAIL

        Re: Happier Days

        Much earlier still - the crap that was MeSsy-DOS.

        Micro$loth had no idea how to write a disk operating system, let alone a GUI. Everything they released after BASIC was a disaster.

    2. Liam Proven (Written by Reg staff) Silver badge

      Re: Happier Days

      > the crap started with the disaster that was Windows 8

      Nah. Windows 2000 was the last good one, and it was already polluted with IE junk.

      1. ibmalone

        Re: Happier Days

        Basically yes. Although whether you draw that line at XP or 2000 is slightly open. XP has the distinction of being the putting out to pasture of 9x series windows in favour of the NT based stuff, making NT the consumer option too. That was actually a big change, the stability of NT was significantly better, prior to that you'd not be surprised if a home or small office machine completely locked up. We do still make fun of the Blue Screen of Death, but, unless you've got some particularly bad drivers, you can now go a whole day without the machine needing a hard restart.

        But of course the Active Desktop stuff had really kicked off in Windows 98 and got carried through, so there's never really been a version of windows without flaws. However awful that was (and it really was), MS really were somewhat ahead of their time there, given the amount of Javascript and other dynamic features in modern desktops.

      2. Doctor Syntax Silver badge

        Re: Happier Days

        "Windows 2000 was the last good one"

        Definitely. That's the one that lives in a seldom used VM on my daily driver for the rare occasions when I need anything Windows related. Anything Windows related would also date from that era. The only exception to that is the W7 which shares a little net-top with Linux and has an OBD package on it - except that I haven't seen the OBD to USB lead for a long time.

      3. ReggieRegReg

        Re: Happier Days

        Yep, loved Windows 2000 - rock solid too. And you could rerun the hardware discovery retrospectively so changing CPU and motherboard was possible without disturbing your files and apps.

    3. Roland6 Silver badge

      Re: Happier Days

      > "To my mind, the crap started with the disaster that was Windows 8."

      No it started much earlier.

      We forget Win 3 was originally only 16 colours and ran on a CGA 640x200.

      Looking at the screenshots, compared to current versions, the screen is very clear and uncluttered, focusing more on function.

      It would be nice to be able to set W10, say, to implement the W3 windows furniture on a HD display, whilst taking advantage of TrueType/Wysiwyg fonts, effectively pushing the OS into the background, rather than the attention seeking toddler it has become.

  15. AlanSh

    I remember the early multi tasking

    The app had total control. If it didn't give it up, it could hog the CPU and no one else could get a look in. Writing Windows apps was fun in those days.

    Alan

    1. 45RPM Silver badge

      Re: I remember the early multi tasking

      Which, at the time, wasn’t as bad as it seems. In fact, the major nuisance was formatting or copying disks - anything with a tight loop that didn’t expect user input.

      Let’s face it, in most GUI applications, a significant amount of time is in the event loop - waiting for the user to do something. Move the mouse. Click a button. Press a key. And that event loop was when applications could be polled to see if they had something to do.

      Remember, at that time, you probably weren’t listening to MP3s while you worked - and, if you were, some of the processing would have been done on a sound card - so as long as the CPU had decoded enough and the card buffered enough then perhaps stuttering would have been kept to a minimum? But it was the early 90s - we were used to stuttering from our DiscMans. I don’t know how the magic happened but I was using a Tracker on my SE/30 (booted from MacOS 7.5.5 in this case), listening to MOD music. Copying files didn’t affect the beats. And then, most surprisingly, MacOS 7.5 crashed (in itself not a surprise, 7.5 was rubbish) - but the music kept playing until I rebooted. No. I have no idea what happened - but the point is that Co operative multitasking a) wasn’t at all bad and b) I’ll bet most people didn’t notice the different when upgrading to pre-emptive multi-tasking.

      Lack of memory protection - that was the real embuggerance. And, in fairness, kind of goes hand in hand with pre-emption. After all, ifs no good having protected memory if the crashed program won’t hand back to the OS event loop.

    2. Liam Proven (Written by Reg staff) Silver badge

      Re: I remember the early multi tasking

      > I remember the early multi tasking

      Yes but no.

      The key thing is in here but buried a bit. It's this:

      16-bit Windows had 2 different multitasking mechanisms.

      DOS didn't multitask. So, to run >1 DOS app, Windows needed to pre-empt them. _Shigata ga nai_ -- there is no other way.

      But Win16 apps didn't and had to yield control.

      So, 2 forms of multitasker: pre-emptive for DOS, cooperative for Win16.

      All the way to WinME.

      NT could pre-empt Windows apps and it's normal for Win32 but in NT 3.x it was _optional_ for Win16 apps.

    3. ChrisC Silver badge

      Re: I remember the early multi tasking

      Yup, which is why, as a postgrad, I dug into my own pocket to buy a copy of OS/2 Warp, *specifically* so that I could pre-emptively multitask the Windows versions of Matlab and Office, both of which I was using quite heavily for my research work. Having Matlab turn Win3.1 into a completely unresponsive brick for an hour at a time whilst it went and churned through the latest set of calculations, with me then having to twiddle my thumbs instead of being able to continue working on something else in the meantime, was enough of an embuggerance that the one-off hit to my finances in buying Warp (combined with the ongoing hit to my sanity of maintaining the damn thing) felt like an entirely acceptable thing to do...

      Even today, W11 will sometimes get itself into a complete tizzy over *something* and decide that pre-emptive multitasking is simply too much bother for it - just this morning my work PC decided to ignore all user input for a couple of minutes whilst Outlook had a think about something during its startup process, and I sat there wondering how much longer I might need to give it before the three finger salute was called into action - wouldn't mind *as much* if such periods of UI inattentiveness were at least accompanied by *some* degree of feedback to let me know, even if it's just to the same level of accuracy as the Explorer "time to complete" estimates when copying a load of files around, but when the entire UI without warning suddenly adopts a persona of near total lifelessness, with only the mouse pointer still being able to move around to show that the OS hasn't *completely* frozen solid, I really wonder at just how much crap there still lurks beneath the surface in these later versions of Windows, and why it is that despite the decades of supposed improvements in the OS combined with the clear improvements in hardware specs, we're still faced with systems that can still end up being less responsive to their users than a mid-80's spec Amiga.

  16. Blackjack Silver badge
    Trollface

    Read this for OS/2

    [https://www.karlstechnology.com/blog/history-of-ntfs/

    https://www.abortretry.fail/p/the-history-of-os2]

    And this for Windows NT:

    [https://www.abortretry.fail/p/the-history-of-windows-nt-31]

    "In NT, APIs are implemented as subsystems on top of a native API that isn’t publicly documented. Because of this, Microsoft can have compatibility with many different systems (as the modern Windows Subsystem for Linux shows), and Microsoft can make major changes to the NT microkernel and NT executive without disturbing userland. A program written for NT in 1993 can run unmodified on Windows 11 in 2023 as the entire API and userland haven’t changed while everything underneath most certainly has. "

    This is how Microsoft ended ruling the Computer market because it keeps compatibility going on way longer that anyone else.

    And that's also why Windows is a buggy mess that causes problems with each "update".

    I love Linux but I hate when I have to upgrade a distro or install a different one and an App no longer works so I have to resign myself to use a different App.

    And if you need to use some kind of emulator or VM to run something? Most of the time you can have that emulator or VM on Microsoft Windows and for the other times? That's why they implemented that Linux Subsystem.

    What? Is Mac only you say? There is probably a clone of that Aop for Windows!

    There isn't? Oh well, I guess you are a Mac user then.

    1. joeldillon

      WSL1 was an NT kernel personality. It worked, but only sortakinda, which is why we now have WSL2.

      This is /not/ an NT kernel personality because it turns out that wasn't actually fit for purpose, it is a tightly integrated virtual machine and does not rely on the NT native API; it runs a full, actual Linux kernel.

      1. This post has been deleted by its author

    2. Anonymous Coward
      Anonymous Coward

      "This is how Microsoft ended ruling the Computer market because it keeps compatibility going on way longer that anyone else."

      Ehheh, no.

      Either you sell a PC with MS OS (DOS, later Windows) or you don't sell *any* Microsoft products. *That* is the reason for 'ruling the market'. Still is.

      After MS managed to kill DR-DOS, either you paid MS-tax or you didn't sell PCs.

      "Compability" is there because MS isn't actually changing anything, all of it pure cosmetic changes in the UI.

    3. ReggieRegReg

      "This is how Microsoft ended ruling the Computer market because it keeps compatibility going on way longer that anyone else."

      z/OS says "hold my Program Status Word"...

      1. jake Silver badge

        I see your MVS (51 years old), and raise you MCP (64 years young).

        Doesn't time fly when you're having fun?

  17. Mike VandeVelde
    Windows

    Hrumph

    No mention of GEOS so not worth reading.

    https://en.m.wikipedia.org/wiki/GEOS_(8-bit_operating_system)

    1. Liam Proven (Written by Reg staff) Silver badge

      Re: Hrumph

      GEOS was great but this is not about 16-bit GUIs. It's about Windows. It's significant but it's not relevant.

  18. Bebu sa Ware
    Windows

    "Most people don't remember this period"

    "Most people don't remember this period – if they're old enough to have been there, then they're old enough for memories to fade."

    Bit like saying "if you can remember the sixties you really weren't there."†

    Prior to WfW, Windows and I parted rags. It was just too painful to write non trivial code for, at least without access to undocumented interfaces* that MS happily used for its own applications.

    At the time the Win2 and Win3 SDKs were seriously expensive in this part of the world as were the assembler (masm) and C compiler (msc) - no C++ compilers except the cfront based Glockenspiel compiler and Walter Bright's Zortech C/C++ v3 native compiler which I purchased but neither of which were Windows applications (development was basically under DOS.)

    "Old enough for memories to fade." - That I live that long. I don't imagine I am likely to forget the enormous amount of wasted time and money invested in MS's pile of steaming horse apples‡.

    * "Managing Multiple Data Segments Under Microsoft Windows: Part I & Part II" Tim Paterson & Steve Flennikan DDJ v15 (1990) Feb. (#161) p113 and Mar.(#162) p227 Notably: "A little-known alternative ... [whose] description does not appear in the Windows Software Development Kit."

    Attributed to comedian Charlie Fleischer. ‡ Pferdeäpfel.

    1. Ken G Silver badge
      Windows

      Re: "Most people don't remember this period"

      I remember booting Windows 1.x just for shits and giggles (it was shit and made me giggle) then restarting with plain DOS.

  19. Doctor Syntax Silver badge

    "So, yes, these tiny apps [Calendar and Cardfile] were more useful than they looked ... although more so to Microsoft than to you, especially today."

    I found Cardfile quite useful, so much so that I knocked up something similar with Lazarus/Free Pascal.

    Does Outlook really have that facility built in other than for contacts? I don't remember it.

    1. Liam Proven (Written by Reg staff) Silver badge

      > Does Outlook really have that facility built in other than for contacts? I don't remember it.

      Nope, not AFAIK.

  20. Doctor Syntax Silver badge

    It's really an example of the big organisation being a big organisation pitted against a couple of guys being empirical.

  21. captain veg Silver badge

    Nina Kalinina?

    I don't know this author, or her bona fides, and neither, it seems, does Wikipedia. I've skimmed the article referred to and, to be honest, it don't impress me much.

    While seeking confirmation for my recollections of the book "Hard Drive" I stumbled upon this: https://www.filfre.net/2018/08/doing-windows-part-7-third-times-the-charm/

    Again I don't know this author and Wikipedia remains silent, but the story told there at least chimes with my direct memories of the period, and not any kind of conspiracy theory or refutation thereof..

    -A.

    1. Roland6 Silver badge

      Re: Nina Kalinina?

      Nina, has effectively provided a set of notes and screenshot examples, whereas Jimmy’s article you refer to has made it into more of a story.

    2. Liam Proven (Written by Reg staff) Silver badge

      Re: Nina Kalinina?

      > I don't know this author

      Nor do I. I don't know her from Adam.

      > neither, it seems, does Wikipedia

      It's not so useful for biographical info. Even if someone adds it, it tends to get deleted. There's a camp among the editors called "deletionists" who remove anything they consider "insufficiently notable". A bunch of my Wikipedia articles have been removed, or edits and additions removed. It's why I no longer bother: they are a curse to the project, but they are annoying think-they-are-do-gooders and there's no central management to slap them down.

      However, I did go back through Kalinina's Mastodon posts and looked at a bunch of her other OS posts. She really knows her stuff, and on any area where I personally know the history, she is rock solid. I was there for much of the stuff in this article: I used Windows 2 in production myself. There is not a single fact wrong I can trace in this article.

      Unfortunately, there are a lot of people who take this stuff personally and who are deeply invested in projects that have nothing to do with them. The result is a lot of conspiracy-theory nonsense about Windows, Microsoft, OS/2, etc.

      I've had emails about this article calling me an evil Microsoft shill who's trying to rewrite history.

      They wave articles from 1991 and 1992 with MS bashing OS/2 and showing OS/2 crashing as "proof" that Windows 3 was "sabotage".

      The idiots apparently can't track dates.

      The new look of Windows 3.0 first appeared in 1989 in OS/2 1.2. The proof is that at that time MS was still committed to OS/2 and was releasing new product versions for OS/2.

      I do not know how long it took to get Windows 3.0 ready to ship. At least a year. This article says that the skunkworks effort that moved the Windows "kernel" -- not terribly meaningful for 16-bit Windows but trolls don't know that -- into protected mode was 1988.

      We can assume 18 months to 2 years to get it ready to ship, and to combine Windows 2.01, Windows/286 and Windows/386 into a single product. It was a big job.

      3.0 shipped in 1990 and you bet your patootie that after that, _yes_ MS was gunning for OS/2 and wanted to show it failing, crashing etc. By that point it had its own rival.

      I wrote last year about _MS_ OS/2 2.0 from 1990. It was nearly ready. IBM didn't ship it for 2 more years.

      In those 2 years, _yes_ MS was trying to make OS/2 look bad. It knew the 32-bit one was nearly ready; it worked on it. It had a version.

      But IBM dallied and tarried for 2 years.

      By then it was too late. It was all over.

      MS wasn't especially smart, or visionary, or bold, or pioneering. It just acted fast when it got an edge.

      That's how it beat DR and that's how it beat IBM.

      1. Dan 55 Silver badge

        Re: Nina Kalinina?

        I've had emails about this article calling me an evil Microsoft shill who's trying to rewrite history.

        They wave articles from 1991 and 1992 with MS bashing OS/2 and showing OS/2 crashing as "proof" that Windows 3 was "sabotage".

        The idiots apparently can't track dates.

        [...]

        3.0 shipped in 1990 and you bet your patootie that after that, _yes_ MS was gunning for OS/2 and wanted to show it failing, crashing etc. By that point it had its own rival.

        So the articles coincide with when MS wanted to play dirty and it all came out in court later. So where's the conspiracy theory?

        They also beat DR by making DR-DOS look bad and again it came out in court later.

  22. SuperG
    Terminator

    I remember with alacrity exactly where I was and what I was doing in 1990. Heck, I was the only one (the Highlander?), in an office full of engineers with nothing but Dec terminals linked to a VAX, that even had a PC. It would be another year or two before they'd all get PC's, when they'd realized that their compiler vendor now made an edition that newly ran under DOS, and by Jove, you could get a cheap EEPROM burner thingy on a ribbon cable attached to your PC. No need to trot down the hall to the VAX to burn a prom - woohoo! Pre Windows 3.1, you had a choice of running some El-cheapo IBM LANBIOS compatible DOS driver ethernet stack provided gratis by 3COM with the purchase of an adapter, I think it was, or you went the high dollar route with IPX/SPX & NetWare. 1992 rolled around with Windows 3.1 and overturned the whole table, mooting the expense of a separate driver you had to pay for. One would love to have been a fly-on-the-wall at Novell HQ to witness the very beginning moment of it's demise.

    1. Anonymous Coward
      Anonymous Coward

      "Pre Windows 3.1, you had a choice of running some El-cheapo IBM LANBIOS compatible DOS driver ethernet stack provided gratis by 3COM with the purchase of an adapter,"

      Nothing 'el cheapo' in Trumpet Winsock. No idea where it appeared, but I'm quite sure I didn't buy it. Better quality than anything 3com offered.

      I'd say it was better TCP/IP stack than what MS provided in 3.1 for workgroups. And it cost more than ordinary 3.1, while Winsock was essentially free.

  23. beardman
    FAIL

    Use of 386 protected mode was a lie until WindowsNT

    Windows has not used 386 protected mode until WindowsNT. All 3.0, 3.11 and even 95 did not use protected mode. How do I know? In between 3.11 and 95 as a student I read about 386 protected mode and was fascinated. Even agreed with uni to do an assignment outside of my CompSci program. And I had 486 made not by Intel, but AMD. And mine 486 had a bug. It entered protected mode and worked as expected, but could not leave it. Upon instructed to exit protected mode it would just lock up. At first, I thought my code was wrong. But when OS/2, which did use protected mode, could not reboot my PC it dawned on me - this CPU had a bug and all the Windows I've tried on it did not use full potential of 386 architecture and instruction set.

    1. Dan 55 Silver badge

      Re: Use of 386 protected mode was a lie until WindowsNT

      There was more than one way to exit protected mode, and the IBM way (safe to say it was used by OS/2) required AT-compatibility with BIOS support.

    2. Liam Proven (Written by Reg staff) Silver badge

      Re: Use of 386 protected mode was a lie until WindowsNT

      > All 3.0, 3.11 and even 95 did not use protected mode.

      Yeah, no, that's not true.

      I don't think you understand things like a 386 mode kernel controlling v86 mode instances and so on as well as you think.

      Full potential? Nothing uses the whole architecture. Nothing ever did.

      I explained the protection rings and how they worked 14 years ago. OS/2 used more than almost anything.

      https://www.theregister.com/Print/2011/07/11/a_brief_history_of_virtualisation_part_one/

  24. Tanaka

    I remember when ipx/spx was bundled with Doom so that you could get networking up in DOS with a BNC terminated ring circuit. Mis-spent youth!

  25. Willy Ekerslike

    Runtime Win2

    I never used Windows 1 (my first GUI was GEM on an Amstrad 1640) but I recall a runtime version of Windows 2 came with one program I bought, sometime in the early 1990's (though I can't remember what that program was - probably a project management or flowchart one). It was back when Harvard Graphics was becoming popular (its output usually being printed to B/W overhead transparencies).

    Regarding the comment about the RN forgetting the cure for scurvy - that was over a longer timeline than it takes many companies to forget how they fixed their earlier mistakes. In my 40+ years in industry, I reckon it only took 10-15 years for previous problems to return. I study I ran put it down to individuals being relied upon to retain the knowledge on the fixes: whilst one would often pass that knowledge onto their replacement when they moved on, it was unlikely to be highlighted at the next handover (and people tended to move jobs every 3-7 years). The result would be a relaxation of whatever control had ben put in place (in the erroneous belief of efficiency) and the original problem then recurs. I carried that concept into shift patterns: for critical work, where there is a reliance on the individuals, rather than systems, 12 hour shifts are safer than 8 hour ones.

  26. big_D Silver badge

    Terminal network...

    A couple of minor points, MS-DOS could network, just not natively. When I started work in 1987, the company I worked for had networked PCs, plus one acting as a bridge to the Mac (AppleTalk) network, with an AppleTalk card stuck in one slot and an Ethernet card in the other. It could copy files back and forth between the two networks.

    We used HP Vectras with HP Ethernet cards and drivers and we also used DECNET for MS-DOS to use the VAX as a file server. In fact the classroom PCs were configured to remote boot into Windows. I then moved onto a client that used Novell Netware and Token Ring!

    As to the terminal, it was pretty basic, but I used it to talk to my Amiga and it could talk to a VAX, but you really needed something like Reflection 4 for proper emulation, including ANSI and DEC escape code interpretation.

    As part of an acquisition, we also got a VT100 terminal emulator, although it was never really marketed... Ran on an ACT Apricot Xi - ACT's GUI, MS-DOS, WordStar, Multiplan, dBase II, C compiler, C interpreter(!!), MS BASIC compiler and interpreter, a few databases and documents, plus the source code for the terminal emulator (C code) all fit on a 10MB hard drive with plenty of room to spare!

    1. Roland6 Silver badge

      Re: Terminal network...

      > MS-DOS could network, just not natively.

      That’s how I took it, back in 1985 we were using Sun Pc-NFS, FTP Software and KA9Q on MSDOS and Win3 before Trumpet WinSocket, which given we were also running Unix hosts (Sun and NCR) was more suitable than Netbios et al.

      We forget just how many proprietary networking stacks abounded at the time. I seem to remember one of the intentions for LanManager was to try and create a unified (proprietary) application network interface to the differing PC network stacks that existed.

      As for terminal emulators, Reflection I seem to remember was one of the better ones, but was on the “heavier” side.

    2. big_D Silver badge

      Re: Terminal network...

      I just started reading Nina's article. The first thing that caught me was her pricing. $600 for an XT? $1000 for an AT? We were buying HP Vectra 286 at the time, because they were about 1K less than an equivalent IBM AT, but they were still expensive at 3,500-4000UKP.

      1. Liam Proven (Written by Reg staff) Silver badge

        Re: Terminal network...

        > $600 for an XT?

        Yeah. Americans had it much, much better. Europeans did badly.

        That's why the Amiga and ST did well over here after they started flopping in sales in the USA. That's why when Commodore went under, German vendor Escom bought them.

        The first affordable PC clones over here were from Amstrad. (Schneider in DACH.) Americans have never even heard of Amstrad.

        It was a different world.

        1. Terry 6 Silver badge

          Re: Terminal network...

          I mentioned in some thread or other a week or so back. There had been an article in PCW magazine to the effect that the US manufacturers had given large discount to UK resellers to get the prices down and kick start sales here. The resellers pocketed the cash and kept the prices high. I think I explained it by saying that UK companies then(?) often had the mentality that a price reduction was taking money out of their own pockets. They'd rather sell 1000 units with £100 profit on each than 10000 units with £90 margin.

    3. Liam Proven (Written by Reg staff) Silver badge

      Re: Terminal network...

      > MS-DOS could network, just not natively

      I would reword that. I think it is misleading in this phrasing.

      MS-DOS *could be networked*.

      On its own, no version of MS-DOS had _any_ networking support at all. No NIC drivers, no protocol stacks, no modem dialler, no filesystem redirector (except the repurposed one that became the CD driver: `MSCDEX.EXE`.)

      Novell DOS did. MS-DOS did not, nor did IBM's PC DOS.

      You could put clients on it, sure. You could take the product and network it with additional tools. That's not the point here.

      The point is that DOS did not do it out of the box. Nor did Windows 3.0 or 3.1. Only the different product Windows *for Workgroups* did that, and it cost money. It was not a free upgrade. WfWg was a different SKU.

      Windows 3.1 is not the same product as Windows for Workgroups 3.1.

      Windows for Workgroups 3.11 is a whole different OS version with a new filesystem, VFAT, running as a 32-bit VXD, and a whole new 32-bit network stack, also as VXDs.

      (Which did not talk TCP/IP over dialup -- Internet Explorer had to bundle that separately and they did not interoperate.)

      The 32-bit network stack didn't include 32-bit TCP/IP because IP was niche then. Out of the box it included NetBEUI and IPX/SPX. 32-bit TCP/IP was an addon. It's still out there:

      https://winworldpc.com/product/microsoft-tcp-ip-32/tcpip-32-3-11b

      There was even a Windows 3.11, which is extremely rare and which I never saw in the day. It was Windows 3.1 plus updates and did not include 32-bit file access.

      https://winworldpc.com/product/windows-3/311

      There was even Windows 3.2 for the Chinese market.

      https://microsoft.fandom.com/wiki/Windows_3.2

      Again, Win 3.1 with updates and none of the enhancements in WfWg 3.11.

      OS/2 2.0, 2.1 and 3.0 did not include network support either. I think Warp 4 did but I had left the OS/2 camp by then. I only played with it as a curiosity.

      1. MarkMLl

        Re: Terminal network...

        There's a considerable degree of internal support- "hooks"- in DOS from at least v3 onwards, as you can see from the data structures etc. in Ralph Brown's Interrupt List.

        I can't speak for MS-DOS, but you could certainly get a networking layer from IBM easily and cheaply, and that gave you streams (?) and mailslots which were compatible with those later implemented by Windows for WorkGroups and Workgroups for DOS: I've programmed them for interprocess communications, e.g. to write an SMS server.

        What DOS lacked was server capability, for which you originally had to go to some poorly-understood (to the average sales/support people) product from 3Com+MS. However the significance here is that it was this combination that became LAN Manager, and to at least some extent it was the LAN Manager API which gave both OS/2 and Win-32 their filehandling and interprocess communications APIs... I'm a little unclear whether than includes the "godothisoncompletion()" callback but that was certainly in place by OS/2 v1.

        Another significant strand is whether the DLL structure came from OS/2, LAN Manager, or something even older: i.e. "European" multitasking DOS-4. I've actually come across a development tool from that era which backported DLLs onto straight DOS in lieu of using overlays.

        So irrespective of the extent to which IBM and MS were able (through wisdom or luck) to plan things to each others disadvantage, some of the ancient history might have roots even deeper than Nina Kalinina realises.

        1. jake Silver badge

          Re: Terminal network...

          The networking hooks first appeared in DOS 3.0, but they were just placeholders at first, and not usable until DOS 3.1

    4. Anonymous Coward
      Anonymous Coward

      Re: Terminal network...

      Back in 1985, I was using a DEC Rainbow, running DR-DOS and part of the company network (and it was a BIG company). It was quite crude, looking back, but effective for the time. The physical network was via the RS232 serial ports and gave access to both local and remote printers, telex and a simple messaging service. It was probably just VT100 emulation, but it worked for us.

      This was a time when telex use was widespread, at least in our industry, and it was the first choice for our team to communicate with suppliers - and being able to send them direct was a boon (receiving needed an extra step for our telex operators to redirect them, but it was a lot quicker than prints being sent through the office mail system). And speed was sometimes essential as delays of even a day could cost >$1million.

      I remember using my Psion II and its serial adapter to test network ports in our office (because IT periodically repatched the system for office relocations and our ports were occasionally swapped around). I also used the Psion II to check printer configurations - plug it into the printer's port and run a simple program that ran through the various permutations - the ones that printed out sensibly were the ones to then use.

      How things have changed in the succeeding 40 years - more speed allows things to get out of hand even quicker. The human part of the chain can often be an essential sense-check - I suspect replacing that with AI will ensure more proverbial hits the spinning thingy, more often and faster.

      1. jake Silver badge

        Re: Terminal network...

        In 1986 the HDD-less Rainbow on my desk was connected to the VAX cluster via thicknet ... I don't think that one ever booted DOS of any kind, it was only used to interact with the corporate ManMan system.

      2. big_D Silver badge

        Re: Terminal network...

        In 1988, I had a DEC Rainbow, CPM/86 HDD, Ethernet, serial to Gandalf for connection to the VAX cluster, a Burgoughs BTOS terminal connected to the BTOS server, an IBM PC (dual floppy), an HP Vectra MS-DOS PC (5.25", 3.5" and 20MB HDD), an HP 125 CP/M PC, an HP 150 MS-DOS PC and a Mac Plus. I was supporting various systems on each of those and none were compatible with each other.

        The Vectra was supposed to be the first HP MS-DOS PC that was IBM PC compatible, but wasn't really, it ran about 80% of IBM PC compatible software.

  27. dharmOS

    Thanks for the scurvy advice

    Hi Liam

    Thanks for the article and the reference to scurvy. I worked at Hasler Hospital where James Lind had been a doctor some 200+ years before me. The amnesia of the military to advances made in war and lost in peace time was coined as the "Walker Dip" by the late Surgeon General, Surgeon Vice Admiral Alasdair Walker. This regression in military medicine that occurs between conflicts is repaid in the lives of service personnel at the start of the next campaign.

    Should anyone wish to read the article, it is linked below.

    https://jrnms.bmj.com/content/104/3/173.full

    1. dharmOS

      Re: Thanks for the scurvy advice

      I meant Royal Hospital Haslar (not Hasler)

  28. Acrimonius

    Mouse was a godsend for spreadsheets

    I remember using SuperCalc and feeling jealous of Excel on the Mac

  29. Step'n'Fixit

    I've been involved in the industry since the early 1970s and Nina Kalinina's fine work released a flood of memories, but more than that filled in quite a few blank spots. Thank you for bringing it to your readers.

  30. Anonymous Coward
    Anonymous Coward

    Ah memories, gave me the nudge I needed to go and do a little googling

    My secondary school experience stradled the time when BBC micros (and Spectrums) were still to be found but the school had just procured a room full of shiny new RM Nimbus machines for computing purposes. I have the vaguest of memories of using said beasties with a version of Windows and looking through all of the info it must have been PC-186's probably running Windows 1 originally (I do seem to recall seeing fancier boxes running Windows 386 in later years but those were for teachers, not for the plebs). By the time I left the entire building had been cabled with 10 base 2 and we'd gained a second dedicated computer room as well as RMs being available across all subjects but I think we still used a beeb with a pH sensor attached in Chemistry :)

  31. The answer is 42

    Grumpy

    Can Windows 4, 5 and 6 be translated into words?

    1. Doctor Syntax Silver badge

      Re: Grumpy

      If they could they wouldn't be words to be repeated in polite company.

      The reality is, of course, that the simple (even with point releases) numbering had been running for some time. Basic marketroid behaviour requires that by then an entirely new branding has to be introduced because (a) the existing stuff was done by the previous lot and (b) there's an awful lot of crud to be distanced from.

  32. midcapwarrior

    Butterfly affect

    "Windows was supposed to die, and it would have if not for one fateful encounter.

    "Well, at a late-June Friday-night party celebrating the opening of Microsoft's big new Canyon-Park manufacturing facility, Murray spied his good friend David Weise, a Windows developer, fellow physicist, and all-round computer whiz. Looking for some fun, Murray teased David that David's new Windows 286 (Windows 2.x with access to the 64-KB HMA) was basically a joke. What one really should do was to get Windows into protected mode and blow away the 640-KB RAM barrier altogether. Much to Murray's surprise, David said, "Yes, let's go do it!" So Murray said, "OK, how about tomorrow?" David said, "No, let's go right now!" -- "The Personal Computer from the Inside Out", by Richard Shoemaker and Murray Sargent.

  33. Sceptic Tank Silver badge
    Windows

    If we could all just cooperate here.

    The article(s) talk of preemptive multitasking in Windows/386. I never used the product, but I seem to remember that all Windows versions up to 3.11 used cooperative multitasking. I can't think that they would have regressed from preemptive to cooperative multitasking from W/386 to Win 3.0. Cooperative multitasking was what made Win 3 so suckey to work with: one app crashes, so it does not hand control back to the scheduler and the whole house of cards comes crashing down. IIRC there was even a Yield() API function (PDF, p 119) to interrupt a long-running tasks and give some CPU time to other waiting applications.

    1. Dan 55 Silver badge

      Re: If we could all just cooperate here.

      Only DOS executables were pre-emptively multitasked, Windows executables were co-operatively multitasked.

  34. Zippy´s Sausage Factory

    That screenshot of Windows 2 took me back. I really kind of miss the Executive even now... probably one reason why I still like the Commander style of file managers so much.

  35. Frank Leonhardt

    Killer Application

    I was there. In the PCW office, mostly. We could run whatever we wanted - vendors were falling over themselves to get their product on our desks.

    It had nothing to do with OS technology. It was a platform to run an application.

    There was simply no software anyone wanted to run on Windows (or OS/2), until there was. The majority were running WordPerfect (or WordStar) and Lotus 123 on MS-DOS, and any OS that didn't support these was a dead duck. If it supported it, but not as well as native DOS, it was a dead duck.

    It was only worth switching to an OS if it offered something significantly better.

    Microsoft Word for Windows 1.0 was somewhat flaky, but did WYSIWYG - which was a main selling point for the expensive Apple Mac (MacWrite).

    As soon as Word for Windows became viable with Windows 3.0, a significant number of people had a reason to run Windows.

    I say it was that simple.

    1. jake Silver badge

      Re: Killer Application

      I'd go as far as pinpointing Windows dominance becoming real with Office 3.0 (Word 2.0, Excel 4.0 and PowerPoint 3.0 ::spit::) in one package. In late summer of 1992, one could purchase a Windows 3.1 PC[0] running the above as a package deal from Fry's Electronics in Sunnyvale for under $1500 out the door. Or so says the receipt I'm holding, dated Friday, September 4th.. I installed several dozen similarly configured PCs in small offices across SillyConValley over the following year or so. They were relatively cheep and cheerful, they worked, weren't really crash-prone (they key there was video drivers), ugly little beige boxen.

      [0] 386sx16, no math-co, 4 megs RAM, 80 meg Maxtor, 1 meg video card, soundblaster compatible, 4x CDROM, internal modem (NOT a "win-modem"; they came later), with speakers, mouse & keyboard, monitor extra. Obviously an obsolete parts-bin build.

      1. Frank Leonhardt

        Re: Killer Application

        Excel was okay, but spreadsheet geeks were hooked on Lotus. But word processing was the one thing every user wanted, and WYSIWYG was a feature that the incumbent WordPerfect couldn't do. Paired with a LaserJet II, a Windows 3.1 PC with Word for Windows 2.0 was way cheaper than a Macintosh and LaserWriter - and it could still run your database and other DOS applications. The Macintosh couldn't.

        Windows 3.1 was important for WYSIWYG word processing because it added TrueType fonts, IIRC.

  36. Randy Hudson

    MacOS today isn't really a descendant of the 80s

    The macOS we have today is from NeXTSTEP. It was a competitor to the mac OS of the 80s, not an evolution of it.

    1. jake Silver badge

      Re: MacOS today isn't really a descendant of the 80s

      Yes. At it's core, the MacOS of today is far more Berkeley than it is Cupertino.

  37. CrazyOldCatMan Silver badge

    Scurvy..

    For instance, in the middle of the 1700s, the Royal Navy worked out how to cure scurvy, but by the end of the 1800s, it had forgotten again.

    There's a reason for this - in the 1700/early 1800s ships were at sea a *long* time between prots and had very little fresh food storage. 100 years on, neither of those things were so true any more - so sailors were not away from port long enough to get scurvy.

    Until conditions changed and the grog ration (which had had the citrus mixed into it) was non-citrus again (those lemons/limes don't pay for themselves y'know!). Ships started to get more long-range and the average time at sea went up again - and scurvy became a thing again.

  38. CrazyOldCatMan Silver badge

    As omeone who lived through that period..

    OS/2 failed for a number of reasons:

    IBM (initially) only allowed it on PS/2 machines (using the MCA bus) and had fairly stringent BIOS requirements that, at that time, only PS/2 machines could pass. I worked for a company that used PS/2s (my work PC was a PS/2 50z which IBM developed for *us* and the majority of the 400k machines they sold went to travel agents that we supported).

    It was on that machine that I first tried Windows 1 and 2 (they crashed a lot - my job involved using the 3270 terminal emulator and,after about 30 minutes, it would crash Windows). I well remember the days of carefully crafting config.sys to actually have some usable RAM in DOS with the network/netbeui stuff running.

    Second big mistake, once they had opened up OS/2 to non-IBM machines, was the price. I bought OS/2 Warp 3 (with a freebie SB16 card and CD-Rom drive) and it was several hundred pounds (got a free t-shirt though which I sometimes still wear today) Microsoft were still (basically) giving away Windows 3 for free (especially Window 3.11 with built-in networking) even though, nominally, it was an expensive product - in practice there were many, many easy ways of getting it for free.

    Third - it was complex - written for geeks, for geeks. Whereas Windows was just a form of what people were used to at home with their Amiga/Atari et. al. machines and didn't have any complex ideas like object orientation. So people like me loved it, some of my less technical colleagues (yes, mainframe systems programmers were not necessarily technical - or more accurately, were 'narrowly technical') loved it - even if it crashed more often.

    1. Frank Leonhardt

      Re: As omeone who lived through that period..

      I was running Microsoft OS/2 version 1.x on whatever I wanted - it wasn't tied to PS/2. In fact it arrived early, before the PS/2 was read, and targeted at the IBM-AT (and thus all its clones). IBM and Microsoft both released 1.0, and Microsoft's didn't even run on a PS/2 (lacked the drivers).

      Microsoft kept OS/2 going a lot longer than people think because it was the platform for their LAN Manager file server.

  39. Anonymous Coward
    Anonymous Coward

    " Windows 3 was perhaps the greatest comeback story in the history of software at that time."

    Under the hood it was almost as bad as the previous versions and programming for it was a nightmare.

    But UI was designed to please a toddler, so of course MBAs and PHBs bought gazillion units. Pure marketing.

  40. Atlantic Roller

    Brilliant article

    But underplays the significance of CP/M during the evolution of MS-DOS and Windows.

  41. Anonymous Coward
    Anonymous Coward

    (about networking) "the first Microsoft client OS with it was Windows for Workgroups 3.1, released in 1992."

    Yes, and cost extra compared to normal 3.1. Networking wasn't included for free, so a note that it cost extra for OS/2 isn't whole truth.

  42. Anonymous Coward
    Anonymous Coward

    "conspiracy-led nonsense like "Microsoft intentionally sabotaged OS/2 because it wanted to sell Windows!""

    Bill-boy has and had 2 goals: To be a monopoly and make maximum amount of money.

    Of course they did that: Their modus operandi has *always* been sabotaging competitors. *Every time*.

    You can argue about effectiveness, but the will to do so definitely is there and has always been there: You can' argue about that.

  43. Anonymous Coward
    Anonymous Coward

    "And it still couldn't multitask DOS apps ... although a version of Windows that could do that shipped the year before."

    Kind of "multitask": Any program could hog all of the cpu and then all the others were just stalled.

    It kind of worked with applications behaving nicely. But OS/2 couldn't do even that, so even worse.

  44. Anonymous Coward
    Anonymous Coward

    "Microsoft didn't sucker those industry partners into doing expensive R&D as a feint. It was doing the same itself. "

    Semi-truth: MS could easily afford it because they had a huge cash cow directly from PC makers, selling DOS to them. The others couldn't.

    A strategy MS has used to butcher competition since then.

  45. Anonymous Coward
    Anonymous Coward

    "There was no long-term corporate strategy here"

    That is bullshit: Bill himself had very clear strategy: 1: Be a monopoly and 2: Maximum profit.

    Both directly copied from IBM, per se not a bad idea. Except for the customers.

    MS at one point had more lawyers than people actually writing code, just for goal #1.

    It's not *all* wrong though: MS didn't have *technology* strategy, just because technology was irrelevant. They still don't have: They do whatever marketing says.

  46. Anonymous Coward
    Anonymous Coward

    " just as it did to land the IBM deal for PC DOS a decade earlier"

    Revisionist history: IBM asked for it, zero acts from Bill or Microsoft needed.

    Now the actual question: Why would IBM ask for an outsider to make OS for their little toy machine, while they could do it themselves effortlessy?

    And even bigger impossibility: The terms: IBM had *never* given rights for anyone for anything. Even less paying royalties. Not a single time, for any reason. Never.

    But of course, Bill-boy had his mom pulling the strings behind the curtains. Directly from the top: That's how you do it.

    https://www.cnbc.com/2020/08/05/how-bill-gates-mother-influenced-the-success-of-microsoft.html

  47. ReggieRegReg

    Post breakup IBM then rewrote OS/2 increasing efficiency massively, OS/2 Warp and Merlin would run nicely in 4Mb of RAM. Job done - memory guzzling NT (would hardly boot with under 24Meg) Microsoft would have its work cut out to complete - er, 'cept memory became cheap and plentiful overnight! 32Mb becoming the norm - Doh! Sometimes events aren't on your side.

  48. ReggieRegReg

    Windoze

    I remember using Word for work under Windows 3 - what a backward step it was from using Quill+Spellbound on my Sinclair QL. Real-time spellchecking (multitasking) in 1986. Took years for other word processors to catch up with that feature. As for Windows in general, only casual users are faster with a mouse, tab key and keyboard was (and still is) king for efficiency if the software is developed to allow the flow.

  49. steviebuk Silver badge

    an illustrated history?

    One image.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like