Yes, crazy that something this good could turn so bad in just two years. In hindsight I am happy their move to Windows pushed me to switch to Linux though.
Was it really ever good though? I came to dos after us having an amiga and it mostly seemed like garbage except that there was more software. Having to jave different autoexec.bat and config.sys files for every game I wanted to play, and laboriously solving the knapsack problem forneach one to get all the drivers I needed into ram.
In MS-DOS 5.0 and later at least you could make a menu at start-up to pick what to activate. I looked at my old config.sys and autoexec.bat recently and it was amazingly simple. 100% of system configuration in maybe 1-200 lines of text, in two files. Not so bad compared to hunting for were to configure something in any modern OS?
DOS did what it had to do. For coding you could start up emacs (and ctrl-z somehow magically could drop you in a command.com shell that you could exit from to go back to emacs; I think there was a DOS interrupt to allow applications to start sub-shells like that?). And otherwise (i.e. most of the time for most of us) you just booted and started a game and then when you were done playing turned off the computer.
Anyway I must admit, many years later, that many games (and some applications) around 1990 were definitely better on an Amiga. DOS games for some time were rarely the best ports. But I never used an Amiga enough to learn if I liked the operating system much. I did buy Amiga Forever some year ago and that comes with a few pre-configured disk images with AmigaOS and various applications, but I am barely able to use that at all since I have no idea how to use it. It would be fun to play with some day.
For work maybe? As a gamer I never used 3.1 for anything. Even the gui apps I used were dos based. Seemingly a lot of them were ported from amiga. Trying to imagine a modern os with no window manager, just applications that all used gl or vulkan directly with some unmaskable keyboard combos to allow switching.
> Trying to imagine a modern os with no window manager, just applications that all used gl or vulkan directly with some unmaskable keyboard combos to allow switching.
Can I have that? Please!
Firefox won't open full-screen if started in X without a window manager.
The easiest approach would be configuring Linux to do this.
I have a Linux laptop without X or Wayland. It's mostly for CLI and TUI applications, but I do have some framebuffer applications like fbpdf.
There are a lot of little annoyances, many of which could be fixed with a proper Linux configuration, though others require improvement to the applications.
The framebuffer versions of graphical programs are unfortunately clunky compared against the conventional versions. And I'm disappointed that they don't work when called from tmux or Midnight Commander, or at least I haven't figured out how to make them work.
Some key combinations don't seem to work and I haven't been able to figure out why for all of them yet.
I miss having a unified clipboard.
I've also become more adept at Midnight Commander and its built in editor, diff tool, and viewer. Though this exposes some issues, for example: If a file is edited with F4 configured to be mcedit, then I can switch between the editor and file manager with the screen list (Alt-`). But if I open mcedit from the command line or user menu, then the screen list will only be for that mcedit run and not include any other screens open. I rarely use the Midnight Commander screen list outside of this specific computer.
(As for why I have this computer, it's to have a distraction-free setup. It works nicely for that purpose.)
You're right, it took a while for Windows gaming to really take off, even in the Windows 95 era. But Windows 3.1 was great for other kinds of productivity-related tasks.
Not only that, it was a couple of months after the first usability testing sessions of the original UI of Windows 95[1] (the one with a literal "tray" under the "desktop"[2,3]).
> In early builds of Windows 95, the taskbar originally wasn't a taskbar; it was a folder window docked at the bottom of the screen that you could drag/drop things into/out of, sort of like the organizer tray in the top drawer of you desk.
The image is really fascinating to me. The whole taskbar paradigm we're familiar with today started life as what sounds like a quick hack. Draw a bar that draws the contents of a folder in a narrow bar format. Later extend it to hold running applications instead.
A silly pet idea I have is that while time didn't literally pass slower, there was less overall entropy in the world back then. There were fewer people working on fewer things, so less new things happened overall. Today there are more people working on more things, so the rate of new things happening is higher, and thus the perception of time passing faster.
And we are all connected, so there are not only twice as many people on the planet, and more people with computers, but also people in China, India, Brazil, and Russia are able to collaborate instead of being completely isolated in their own locales.
Ah I wasn't talking about "Trusted Platform Module", I was talking about the technical product/project manager job, which tends to create meetings and various distractions for developers
If it was not for DRDOS 6.0, that would never have happened. DRDOS 6.0 was excellent, one thing that never got to MS-DOS was the ability to password protect files, preventing someone else to access it. Granted, that was very easy to get around, but back then the times were simple.
DR-DOS also had some kind of limited multitasking or at least switching between tasks. I do not remember using it much or know how stable it was. I do not think MS-DOS ever copied that either?
MS-DOS had TSRs (terminate and stay resident) programs, but those aren't exactly the same thing. I do recall some basic multitasking utilizing them in some way, though. My memory is too fuzzy.
MS-DOS 4.0 had a multitasking release, but it was not popular and only available through retail channels. IBM didn't want it so there were OEM MS-DOS releases of 4.0 that didn't have multitasking, and that's what you got when you bought an IBM PC. This meant it didn't get a lot of support and was discontinued after MS-DOS 4.1. If it sounds like a mess, it's because this was all when IBM and Microsoft were not very happy with each other, which is kind of why everyone mostly jumped from MS-DOS 3.3 (IBM and MS mostly cooperative) to MS-DOS 5.0 (MS independent).
DOSSHELL supported task switching, but only one process could execute at a time, and all tasks had to fit in memory.
I guess I could be wrong, but I thought that was only for DOSSHELL itself, not for the tasks you were launching. Perhaps it would work but only if the other tasks supported it?
My quick test of running multiple command prompts and running free showed that each one had a full set of memory. It's not multitasking, and I believe it assumes that the tasks are well behaved (not modifying DOS internals), so tasks don't really need to do anything special if they are well behaved.
You're thinking of what is sometimes called "European MS-DOS 4.0" which did indeed have limited multitasking features, including shared memory and pipes:
Correct, DR-DOS had a Task Switcher, but still single processing. It was much better than the DOSSHELL Microsoft had because you could switch at the command line as opposed to having to run a bloated shell.
Indeed, and I was a ViewMax fan, but eventually Microsoft managed to work around them as we all know, and like with other DOS clones compatibility never was 100% assured.
I grew up with PC-DOS and I always felt slightly like a second class citizen for missing goodies that others had, like qbasic, and also there would be subtle differences when it came time to set up stuff like CD-ROM drives and sound cards.
As a kid I used the MultiConfig (based on a guide in PC Magazine!) to have different menus. When all you were trying to run was games, maxing your conventional memory was key. I had different combos for sound, no sound, CD-ROM, boot to windows, moving stuff between EMS/XMS so older games could fit.
I would meticulously try different combos and seeing that `mem` command return something above 600KB was a reward in and of itself.
I did the same thing. I didn't have internet back then, so I assume I also got the knowledge to do this from magazines. Also, lots of time on my hand back then, so probably dove deep into various weird manuals.
I was stuck with MS-DOS 5.0 as a child, and was mildly envious of 6.x's MultiConfig that I read about in magazines; it seemed so much better than using boot floppies for nonstandard configurations. So I rigged autoexec.bat to present a menu of various configs and rename the related config.sys files so that the one I chose was used after autoexec.bat ran. Worked quite well; the only minor downside I encountered was having to reboot again when choosing a different config than the one config.sys currently held.
It seems his/her hack was to select a particular configuration, get autoexec.bat to copy the file with that configuration as the new config.sys, and then reboot the PC so that config.sys would get read at next boot.
A part of me still feels like I could still jump right in, like it would be the most natural thing for me to quickly juggle around some config.sys and autoexec.bat lines to get that balancing act between conventional and extended right for a given game. Iirc for wing commander the amount of one or the other was even key to adjusting simulation speed relative to CPU, more memory allowing more debris particles, slowing down in-game time?
The other part of me is fully aware that I have forgotten just about everything, even the filename of config.sys was only resurfaced by other posts here.
I'm guessing you were between the age of 10-20 at that time? Whatever tech minutiae you learned as a tween/teenager will be with you the rest of your life. I still can code in real-mode x86 assembly even though I haven't done so in years.
That said, the other reason you could still jump right in, is that these systems were accessible. Small enough to be understandable by a child's brain, a couple of text files you edited by hand, tinkering among a small number of parameters each of which you could test within a few minutes at most, so that worst case you spent a day dinking around with your setup. And without the distractions of the internet or thousands of free ad-supported dopamine pumps to displace your main purpose of getting this game to work.
Contrast this with today: you buy a game on Steam for $30, it doesn't work. What do you do? You have to search the internet or go to the forums, and try any number of random things that are in no way contained or enumerable and the knowledge isn't transferable. Our systems are 100x bigger than any single person can conceivably know.
I totally agree! I was 15 years old at that time. Just edit and reboot until the game run the way I like it. No internet, no distraction. As much a phone call (line phone) to a friend to discuss possible configurations.
Golden days!
I am pretty confident you and I could do it with no manual and just vibes. The satisfaction when you got it all running and star control II finally ran is perm burned into my base brain programming.
I never ran 6, but under 3.3 and 4 we (my pals and I) had all created a complicated array of batch files to create multiple booting options (autoexec/config combos).
It was originally borne of a need to allow us all to have our own preferred setup on shared PCs in the computer lab we worked in, and just grew from there. My pal Dave may have given us the idea with his scheme of multiple boot profiles on his Amiga, come to think of it.
>moving stuff between EMS/XMS so older games could fit.
Oh god this. My pal Mike and I -- who, I note fondly, I was visiting with last weekend, 30+ years later -- spent a LONG time basically becoming subject matter experts on DOS memory allocation and manipulation. It was my first experience with hard-won knowledge that was nevertheless useless relatively quickly; I remember NONE of it now. But holy cow we futzed around with this SO MUCH -- initially to accommodate various games or other heavyweight programs, and then out of momentum.
I learned Logo in elementary school, and some scraps of basic in junior high, but I have no doubt that my baseline computer systems engineering knowledge came about because if I wanted to play video games, I had to get my drivers and my EMS and XMS squared away.
When I hear people today ask, “how do I get into security?” My immediate answer is that you’re going to need a very underpowered 486 and a copy of Wing Commander 3 if you really want to do it right.
I remember being the only kid who knew how to hand-craft your config.sys/autoexec.bat to play some games, and random people would call me saying they got my name of some friend of a friend.
Microsoft tried to license it, failed, and wound up licensing DoubleSpace instead. As usual, the utility being built into MS-DOS 6 pretty much destroyed competition in that space. Stacker's ace in the hole was their coprocessor card, which I coveted pretty hard..
Windows 95, 98 and ME can read DoubleSpace drives. If you get USB drives set up, which you can do with 95 and 98, and format the drive FAT, not FAT32, put the drive file on it, those versions of Windows will mount it.
Some of those games were unplayable without the manuals or a map! Looking at you Star Control II! You needed a map for the coordinates to play that one.
At that early age, coming from the CoCo (which booted straight into a BASIC interpreter in less than a second) the thing that wierded me out the most about DOS was that it was not a REPL or interpreter for any language. It felt really weird to have to start up GWBASIC or QBASIC separately. I did eventually get used to it though.
Not long after this I called Microsoft on the phone somehow, deepened my voice as far as it would go to sound older, and got myself signed up as a beta tester for Windows 95.
Greetings fellow human born in 1981. I skipped the straight-to-BASIC machines (my first PC was an Amstrad 1512) and I didn’t sign up for the Windows 95 (‘Chicago’) beta but I remember queuing for it when it was released on 24th August 1995.
(I later signed up to be a beta tester for Windows 2000. That was truly impressive when compared to Windows NT 4.0 that preceded it. By then I had a multiprocessor dual PIII-450 system, so gaining USB support was a great deal.)
Fun MS-DOS fact: The MSDOS.SYS, IO.SYS & COMMAND.COM files had a timestamp that matched the version number (as you can see in the screenshot in this page). The file time is 6:00a [am]. In MS-DOS 6.22, the last version before it was integrated into Windows, the timestamp of those files was 6:22a.
I happened to have a folder containing DOS 6.22 on my desktop and indeed, all the files in the DOS directory are dated 1994-05-31 06:22. Thanks for the interesting knowledge, I never noticed that! Maybe it was touch-like command run as a part of the release process.
Other random MS-DOS fact, if your filesystem is badly corrupted and it attempts to read sector 0 as a directory, you will see what looks like a file named "δ<ÉMSDOS.5.0"
I'd forgotten MS-DOS 6.0 included antivirus. I have fond memories of VSAFE (I think) buzzing the speaker and warning me more or less at random about something or other (it never was an actual virus).
Friends who used other antivirus solutions of the time (Dr Solomon and F-Prot were two of the other popular ones amongst friends from what I remember) ridiculed it for having so many false positives but somehow for me it was reassuring to see it was "working".
Oh, wow. I remember when it came out. My multi-configs were epic. At that time I was actually an OS/2 programmer but we also had a DOS product so some time reluctantly spent in DOS.
Interesting that a lot of the updates were actually licensed software.
Deciding to (re)build a system level disk compression system the month before a major release, in an era when distribution of patches was done through floppies, was a balsy move.
It also somewhat explanatory of the reputation regarding bugs of MS back then.
And how many multiples over a machine capable of running MS-DOS would I have had to spend to get a NeXT? I'll let you do the math on that one, because I already did it 30 years ago.
It wasn't a mainstream gaming platform, true. But the productivity and development tools available would absolutely run rings around what was available for DOS. Easily worth the price if you needed it.
I can spend several (2-3x) as much for a computer that doesn't run the software most people need? From a company with no track record? Well, at least it's pretty.
You can always tell the replies from someone who only read about it on the interwebs.
You could indeed... But a budget PC with DOS was good enough, had hundreds of thousands of industry-standard apps, could be networked with any one of tens of corporate network servers, from Novell NetWare to 3Com 3+Share.
And most of all, you could equip an entire department of a dozen people with a PC each for the cost of a single NeXTstation.
I think I was on DR-DOS at that point from MS 5.0 until MS released 6.20, and ISTR feeling that dblspace was a huge step down from the DR equivalent SuperStor.
I used MS-DOS since 1.25, but I kinda wish I'd instead been using Unix all that time.
MS-DOS had much better video games, but I didn't much play those. Overall, the things I did do would've been better on Unix, and I would've gotten plugged into that scene earlier. (When I finally got access to Unix workstations, I sold my PC, and never again used another MS box at home.)
Unix was very expensive at the time of course, meaning only used by big business. I remember asking a computer/IT teacher what it was, in the early 90s. Sounded interesting and somehow we found out about Minix and downloaded it from somewhere to tinker, maybe a BBS?
Luckily this happened, because when I later got an IT job in the mid-nineties at a aerospace company I was good to go. At which point I discovered Linux and then BSD.
Also, engineering-ish small businesses, and universities, and the occasional non-profit saving money with dumb terminals where they'd otherwise need a PC on each desk.
I once paid ~$100 for Coherent (Unix-like) for my '286, then sent it back due to the 64KB process limit, while I used Unix workstations at an internship. My next home computer was to buy a used Sun-3/75 from the local Sun office, for a fraction of a what a PC would cost.
Worked in IT in the mid-90s, was a great time and I became a sought after "expert" to free conventional memory for lab PCs connected to equipment. Did it all by hand at first with mem.exe and edit.com?, but I think in 6.0 I started also using a TUI to visualize the upper area—the helpful msd.exe:
More recently in the XP era, I used the DOS multi-config menu in a classroom environment. Combined with ghost the partition imager. There was a menu that defaulted quickly to booting the current OS partition. But hit the right key at boot and you could restore the computer from a pristine pre-configured XP image.
Since the images were on a primary partition it would be hidden from the end-user and not mounted for them to prevent issues.
You could also update the current image and save it back to the menu partition for next time. This happened once in a while to apply windows/app updates.
Worked incredibly well, until the whole classroom was dismantled around 2015ish.
If you had included MultiConfig options that added Soundblaster8/Soundblaster16, with CDROM or without it... AND their complementary AUTOEXEC.BAT options that would trigger based on the CONFIG.SYS, then that would have been my system config! Without being able to squeeze every single kb of memory possible, there was no way to play Sam and Max or Indiana Jones and the Fate of Atlantis, while still having a system that can boot Windows 3.1 :-)
Yeah, it was great having a menu instead of a collection of different boot floppies. Though at this point I forget how much I did with MultiConfig versus once I had enough RAM what I was doing with OS/2's boot manager. OS/2 Warp 3 was wild, but mostly because it was a better way to configure DOS and/or run some DOS and Windows 3.1 apps.
I think there is a lot of disingenuity in this post.
Microsoft was an extremely hostile, aggressive company in those times. (I have seen little to no evidence that it is very different now, but then it was famed for it, and was constantly in court for what I will diplomatically call intellectual property violations.)
The real story is:
* MS went to all these companies and said, in essence, "we're going to bundle your code so either give it to us free or we'll go to your rivals. But you'll make money selling updates and upgrades to all the people who get free versions!"
So, as the company says, it got a free version of Central Point Backup from the PC Tools vendor, and other tools from Symantec.
This was an obvious lie but many companies felt that they had no choice so they signed up.
* STAC Inc was the author of Stacker, the most feature-rich and highest-compressing disk compression package. It was the only one that ran on OSes other than MS-DOS and PC-DOS -- it also ran natively on OS/2 2, where I used it myself. It also offered a hardware compression-offload expansion card, for faster data compression/decompression. (Cheap, much cheaper than a bigger hard disk in the very early 1990s.) Stac entered negotiations with MS, and very foolishly, it gave MS access to the Stacker source code.
* Then, Stac management changed its mind and said no. MS copied some of the data compression code into the inferior disk-compression tools it had got from Vertisoft and shipped it.
* MS denied the theft. STAC sued. STAC proved it, with the help of Geoff Chappell (among others) who just died very recently.
1. Remove the functionality (thus MS-DOS 6.21, with no compression, followed by MS-DOS 6.22, with DriveSpace instead, with the offending code rewritten as the blog post describes.)
2. Pay Stac damages. (IIRC, ITRO $200 million.) Stac bought ReachOut with the proceeds and tried to get into remote control tech support, knowing that now disk compression was bundled, it was doomed. The effort only succeeded briefly: Stac was correct, but the future was remote control over the Internet, not direct point-to-point dialup connections which was the focus of ReachOut.
* Nobody paid for antivirus updates for the DOS antivirus; if they cared, they bought a whole new antivirus. Nobody paid for an enhanced defrag, or defrag at all any more; or for backup tools, or anything else that was bundled. With Scandisk there was little need for disk-repair tools, either, so the PC utilities market cratered. The MS-DOS memory management was good enough that the market for memory management tools such as 386Max, Quarterdeck QEMM386, and the like collapsed as well.
As did the market for DR-DOS: MS-DOS 6.x was perceived as a safer bet, and it was good enough, while a rival _might_ not be perfectly compatible. It was -- better in places -- but there was doubt.
Regarding the DoubleSpace context, Wikipedia helps (1):
Sept 1991: Digital Research releases DR DOS 6.0 with AddStor's SuperStor disk compression.
March 1993: Microsoft introduces MS-DOS 6.0, the first with DoubleSpace disk compression
June 1993: IBM announces PC DOS 6.1
November 1993: Microsoft MS-DOS 6.2, leapfrogging IBM's PC DOS 6.1 and with improved the stability of DoubleSpace.
February 1994: Microsoft found guilty of patent infringement, and Stac Electronics guilty of trade secret theft. Microsoft MS-DOS 6.21, removing DoubleSpace.
June 1994: After a judge ordered Microsoft to recall all unsold infringing products worldwide, Microsoft settled its dispute with Stac, and released MS-DOS 6.22, bringing back disk compression with internally developed DriveSpace, which is about 5% slower than DoubleSpace.
An interesting detail:
In the MS-Stac trial, "Stac’s lawyers showed a videotape of Microsoft Chairman Bill Gates" "at the unveiling of MS-DOS 6.0 in which he wore a T-shirt saying, “We came, we saw, we doubled,” to underscore the importance of the compression feature." (2)
A second detail, the "trade secret theft" was:
Stac was found "to have committed a trade secret violation by reverse engineering features of a beta version of MS DOS that they had gotten in confidence and then using the information they gained in making their own product." (3)
110 comments
[ 1.7 ms ] story [ 165 ms ] threadDOS did what it had to do. For coding you could start up emacs (and ctrl-z somehow magically could drop you in a command.com shell that you could exit from to go back to emacs; I think there was a DOS interrupt to allow applications to start sub-shells like that?). And otherwise (i.e. most of the time for most of us) you just booted and started a game and then when you were done playing turned off the computer.
Anyway I must admit, many years later, that many games (and some applications) around 1990 were definitely better on an Amiga. DOS games for some time were rarely the best ports. But I never used an Amiga enough to learn if I liked the operating system much. I did buy Amiga Forever some year ago and that comes with a few pre-configured disk images with AmigaOS and various applications, but I am barely able to use that at all since I have no idea how to use it. It would be fun to play with some day.
Can I have that? Please!
Firefox won't open full-screen if started in X without a window manager.
I have a Linux laptop without X or Wayland. It's mostly for CLI and TUI applications, but I do have some framebuffer applications like fbpdf.
There are a lot of little annoyances, many of which could be fixed with a proper Linux configuration, though others require improvement to the applications.
The framebuffer versions of graphical programs are unfortunately clunky compared against the conventional versions. And I'm disappointed that they don't work when called from tmux or Midnight Commander, or at least I haven't figured out how to make them work.
Some key combinations don't seem to work and I haven't been able to figure out why for all of them yet.
I miss having a unified clipboard.
I've also become more adept at Midnight Commander and its built in editor, diff tool, and viewer. Though this exposes some issues, for example: If a file is edited with F4 configured to be mcedit, then I can switch between the editor and file manager with the screen list (Alt-`). But if I open mcedit from the command line or user menu, then the screen list will only be for that mcedit run and not include any other screens open. I rarely use the Midnight Commander screen list outside of this specific computer.
(As for why I have this computer, it's to have a distraction-free setup. It works nicely for that purpose.)
firefox -geometry 1024x768+0+0
does not work ?
[1] https://betawiki.net/wiki/Windows_95_Usability_Testing_Build...
[2] https://devblogs.microsoft.com/oldnewthing/20030910-00/?p=42...
[3] http://bytepointer.com/resources/old_new_thing/20030910_053_...
The image is really fascinating to me. The whole taskbar paradigm we're familiar with today started life as what sounds like a quick hack. Draw a bar that draws the contents of a folder in a narrow bar format. Later extend it to hold running applications instead.
MS-DOS 4.0 had a multitasking release, but it was not popular and only available through retail channels. IBM didn't want it so there were OEM MS-DOS releases of 4.0 that didn't have multitasking, and that's what you got when you bought an IBM PC. This meant it didn't get a lot of support and was discontinued after MS-DOS 4.1. If it sounds like a mess, it's because this was all when IBM and Microsoft were not very happy with each other, which is kind of why everyone mostly jumped from MS-DOS 3.3 (IBM and MS mostly cooperative) to MS-DOS 5.0 (MS independent).
DOSSHELL supported task switching, but only one process could execute at a time, and all tasks had to fit in memory.
https://en.wikipedia.org/wiki/MS-DOS_4.0_(multitasking) https://virtuallyfun.com/2013/03/17/european-ms-dos-4-00-aka...
It should not be confused with the mostly unrelated DOS 4.0 later released.
A Microsoft executive later said the multitasking DOS should probably not have been named "4.0" because it created a lot of confusion.
I have demo disks for download on my blog with the service configured. It's readily Googled.
https://liam-on-linux.livejournal.com/58013.html
What were the good parts of PC-DOS?
I wonder how well it would work to copy the MSDOS 'DOS' folder into the PC-DOS one, just keeping the programs with different names.
Note, 7.1, the last version and capable of FAT32, not to be confused with 7.01.
https://liam-on-linux.livejournal.com/59703.html
As a kid I used the MultiConfig (based on a guide in PC Magazine!) to have different menus. When all you were trying to run was games, maxing your conventional memory was key. I had different combos for sound, no sound, CD-ROM, boot to windows, moving stuff between EMS/XMS so older games could fit.
I would meticulously try different combos and seeing that `mem` command return something above 600KB was a reward in and of itself.
I remember Tornado needing 612kb. That took me some serious optimizing.
The other part of me is fully aware that I have forgotten just about everything, even the filename of config.sys was only resurfaced by other posts here.
That said, the other reason you could still jump right in, is that these systems were accessible. Small enough to be understandable by a child's brain, a couple of text files you edited by hand, tinkering among a small number of parameters each of which you could test within a few minutes at most, so that worst case you spent a day dinking around with your setup. And without the distractions of the internet or thousands of free ad-supported dopamine pumps to displace your main purpose of getting this game to work.
Contrast this with today: you buy a game on Steam for $30, it doesn't work. What do you do? You have to search the internet or go to the forums, and try any number of random things that are in no way contained or enumerable and the knowledge isn't transferable. Our systems are 100x bigger than any single person can conceivably know.
It was originally borne of a need to allow us all to have our own preferred setup on shared PCs in the computer lab we worked in, and just grew from there. My pal Dave may have given us the idea with his scheme of multiple boot profiles on his Amiga, come to think of it.
>moving stuff between EMS/XMS so older games could fit.
Oh god this. My pal Mike and I -- who, I note fondly, I was visiting with last weekend, 30+ years later -- spent a LONG time basically becoming subject matter experts on DOS memory allocation and manipulation. It was my first experience with hard-won knowledge that was nevertheless useless relatively quickly; I remember NONE of it now. But holy cow we futzed around with this SO MUCH -- initially to accommodate various games or other heavyweight programs, and then out of momentum.
A great retro YouTuber, Phils Computer Lab, has a similar menu available for download. It's a pretty slick menu kind of like the one you've described.
https://www.philscomputerlab.com/ms-dos-starter-pack.html
When I hear people today ask, “how do I get into security?” My immediate answer is that you’re going to need a very underpowered 486 and a copy of Wing Commander 3 if you really want to do it right.
https://en.wikipedia.org/wiki/List_of_DOS_commands#MEMMAKER
It would help you change your configuration so that you maximized the amount of conventional memory. This was useful in getting games to run.
https://www.pcjs.org/software/pcx86/
I wonder what email system it was being used at the time (I mean, was it even over TCP? SMTP/IMAP?)
The Rainbow magazine (https://en.wikipedia.org/wiki/The_Rainbow_(magazine)) ceased publication that year, which I was very bummed about.
At that early age, coming from the CoCo (which booted straight into a BASIC interpreter in less than a second) the thing that wierded me out the most about DOS was that it was not a REPL or interpreter for any language. It felt really weird to have to start up GWBASIC or QBASIC separately. I did eventually get used to it though.
Not long after this I called Microsoft on the phone somehow, deepened my voice as far as it would go to sound older, and got myself signed up as a beta tester for Windows 95.
(I later signed up to be a beta tester for Windows 2000. That was truly impressive when compared to Windows NT 4.0 that preceded it. By then I had a multiprocessor dual PIII-450 system, so gaining USB support was a great deal.)
Friends who used other antivirus solutions of the time (Dr Solomon and F-Prot were two of the other popular ones amongst friends from what I remember) ridiculed it for having so many false positives but somehow for me it was reassuring to see it was "working".
Deciding to (re)build a system level disk compression system the month before a major release, in an era when distribution of patches was done through floppies, was a balsy move.
It also somewhat explanatory of the reputation regarding bugs of MS back then.
Having done it so many times, those things are burned into my brain forever.
(Note: I have never owned a NeXT machine.)
Also, https://doomwiki.org/wiki/NEXTSTEP
Anyway my main point was that DOS was nearly as primitive then compared to state of the art as it would be considered today.
You can always tell the replies from someone who only read about it on the interwebs.
And most of all, you could equip an entire department of a dozen people with a PC each for the cost of a single NeXTstation.
The people the leaps in tech the simplicity the naivety
MS-DOS had much better video games, but I didn't much play those. Overall, the things I did do would've been better on Unix, and I would've gotten plugged into that scene earlier. (When I finally got access to Unix workstations, I sold my PC, and never again used another MS box at home.)
Luckily this happened, because when I later got an IT job in the mid-nineties at a aerospace company I was good to go. At which point I discovered Linux and then BSD.
I once paid ~$100 for Coherent (Unix-like) for my '286, then sent it back due to the 64KB process limit, while I used Unix workstations at an internship. My next home computer was to buy a used Sun-3/75 from the local Sun office, for a fraction of a what a PC would cost.
https://i.postimg.cc/2jX8LpF0/memorymap.png
More recently in the XP era, I used the DOS multi-config menu in a classroom environment. Combined with ghost the partition imager. There was a menu that defaulted quickly to booting the current OS partition. But hit the right key at boot and you could restore the computer from a pristine pre-configured XP image. Since the images were on a primary partition it would be hidden from the end-user and not mounted for them to prevent issues. You could also update the current image and save it back to the menu partition for next time. This happened once in a while to apply windows/app updates.
Worked incredibly well, until the whole classroom was dismantled around 2015ish.
If you had included MultiConfig options that added Soundblaster8/Soundblaster16, with CDROM or without it... AND their complementary AUTOEXEC.BAT options that would trigger based on the CONFIG.SYS, then that would have been my system config! Without being able to squeeze every single kb of memory possible, there was no way to play Sam and Max or Indiana Jones and the Fate of Atlantis, while still having a system that can boot Windows 3.1 :-)
lolwat
Microsoft was an extremely hostile, aggressive company in those times. (I have seen little to no evidence that it is very different now, but then it was famed for it, and was constantly in court for what I will diplomatically call intellectual property violations.)
The real story is:
* MS went to all these companies and said, in essence, "we're going to bundle your code so either give it to us free or we'll go to your rivals. But you'll make money selling updates and upgrades to all the people who get free versions!"
So, as the company says, it got a free version of Central Point Backup from the PC Tools vendor, and other tools from Symantec.
This was an obvious lie but many companies felt that they had no choice so they signed up.
* STAC Inc was the author of Stacker, the most feature-rich and highest-compressing disk compression package. It was the only one that ran on OSes other than MS-DOS and PC-DOS -- it also ran natively on OS/2 2, where I used it myself. It also offered a hardware compression-offload expansion card, for faster data compression/decompression. (Cheap, much cheaper than a bigger hard disk in the very early 1990s.) Stac entered negotiations with MS, and very foolishly, it gave MS access to the Stacker source code.
* Then, Stac management changed its mind and said no. MS copied some of the data compression code into the inferior disk-compression tools it had got from Vertisoft and shipped it.
* MS denied the theft. STAC sued. STAC proved it, with the help of Geoff Chappell (among others) who just died very recently.
https://news.ycombinator.com/item?id=37384911
* MS had to:
1. Remove the functionality (thus MS-DOS 6.21, with no compression, followed by MS-DOS 6.22, with DriveSpace instead, with the offending code rewritten as the blog post describes.)
2. Pay Stac damages. (IIRC, ITRO $200 million.) Stac bought ReachOut with the proceeds and tried to get into remote control tech support, knowing that now disk compression was bundled, it was doomed. The effort only succeeded briefly: Stac was correct, but the future was remote control over the Internet, not direct point-to-point dialup connections which was the focus of ReachOut.
* Nobody paid for antivirus updates for the DOS antivirus; if they cared, they bought a whole new antivirus. Nobody paid for an enhanced defrag, or defrag at all any more; or for backup tools, or anything else that was bundled. With Scandisk there was little need for disk-repair tools, either, so the PC utilities market cratered. The MS-DOS memory management was good enough that the market for memory management tools such as 386Max, Quarterdeck QEMM386, and the like collapsed as well.
As did the market for DR-DOS: MS-DOS 6.x was perceived as a safer bet, and it was good enough, while a rival _might_ not be perfectly compatible. It was -- better in places -- but there was doubt.
* CPS and the others went bankrupt, killed by MS.
Sept 1991: Digital Research releases DR DOS 6.0 with AddStor's SuperStor disk compression.
March 1993: Microsoft introduces MS-DOS 6.0, the first with DoubleSpace disk compression
June 1993: IBM announces PC DOS 6.1
November 1993: Microsoft MS-DOS 6.2, leapfrogging IBM's PC DOS 6.1 and with improved the stability of DoubleSpace.
February 1994: Microsoft found guilty of patent infringement, and Stac Electronics guilty of trade secret theft. Microsoft MS-DOS 6.21, removing DoubleSpace.
June 1994: After a judge ordered Microsoft to recall all unsold infringing products worldwide, Microsoft settled its dispute with Stac, and released MS-DOS 6.22, bringing back disk compression with internally developed DriveSpace, which is about 5% slower than DoubleSpace.
An interesting detail:
In the MS-Stac trial, "Stac’s lawyers showed a videotape of Microsoft Chairman Bill Gates" "at the unveiling of MS-DOS 6.0 in which he wore a T-shirt saying, “We came, we saw, we doubled,” to underscore the importance of the compression feature." (2)
A second detail, the "trade secret theft" was:
Stac was found "to have committed a trade secret violation by reverse engineering features of a beta version of MS DOS that they had gotten in confidence and then using the information they gained in making their own product." (3)
1) https://en.wikipedia.org/wiki/Timeline_of_DOS_operating_syst...
2) https://www.latimes.com/archives/la-xpm-1994-02-24-fi-26671-...
3) https://lwn.net/Articles/134642/
Cheers to my fellow old-farts in keeping these memories alive. :-)
It's crazy to think that we will be nostalgic about today in another 30 years!
How far and fast the tech scene is moving!