Former Microsoft Developer Would Like To See MS-DOS Open Sourced (youtube.com) 113
For over an hour on Saturday, retired Microsoft OS developer David Plummer answered questions from his viewers on YouTube.
Long-time Slashdot reader destinyland writes: He began with an update on a project to test the performance of the same algorithm using 30 different programming languages, and soon tells the story of how he was inspired to apply for his first job at Microsoft after reading Hard Drive: Bill Gates and the Making of the Microsoft Empire.
I decided that this is where I wanted to work, because these guys sound like me, they act like me, they are what I want to be when I grow up. And holy cow, they pay them well, apparently. So I wrote to everybody that I could find that had a Microsoft email address, which was about four people, because I had a software product people had been regisering on the Amiga. And one guy, Alistair Banks... responded and he hooked me up with a hiring manager directly in Windows that had an open slot that was hiring... And a couple of interview slots later, I wound up as an intern at MS-DOS working for Ben Slivka.
So you would think, "Oh, an intern on MS-DOS. What'd you do? Format disks?" No — it's amazing to me, actually. They give you as much work as they believe that you are capable of, and — they get you for all that you're worth, basically. They had me write a bunch of major features, like the Smart Drive cache for CD-ROMs was the first thing I wrote. Then I wrote DISKCOPY, making it work, single pass, bunch of features in MS-DOS. I re-wrote Setup to work on a single floppy disk by using deltas and patching in place, DOS 5 to turn it into DOS 6, something like, or maybe it was DOS 6 into 6.2... A whole bunch of features, within the span of, like, three months, which to me was fairly impressive at the time, I thought. And that only got me an interview...
Later he says that he'd like to see most of 16-bit Windows and all of MS-DOS open sourced, along with some select application code from that era.
I don't think there's any reason to hold back any of MS-DOS at this point. They have absolutely no reason to open source any of it, really — other than PR, because all it brings them is potential liability, complaints and angst, and probably nothing positive for putting the code out there and exposing it to ridicule. Because it's ancient code at this point. It's like, "Ha! Look what Microsoft did!" Well, yeah, I know Linux is cool now, but go look at Linux code from 1991 — and I worked on some of that code. Well, '93 I did. It's not the same as what you see today.
So yeah, MS-DOS probably looks archaic — although it's super tight, it doesn't have many bugs. It's just written differently than you would write code today, because you're targetting something that is a very different CPU and memory system and PC as a whole, and it's so much more limited that everybody's sacred, every cycle matters. That kind of thing that you don't worry about now. But I'd still like to see all the code from back then that's not embarrassing released.
And when asked what he misses most about being a Microsoft developer, he answers:
I miss going for lunch with the people that I went for lunch with, and talking to the people that I worked with. Because they were a lot like me, they had similar interests, they had similar abilities, they were people like me. We went for lunch, we ate food, it was awesome, and then we talked about cool things. And we did that every day. And now I don't get to do that any more. I get to do it rarely, because I take guys out for lunch and stuff, but it's not the same. So that's really what I miss.
And I miss somebody always feeding me something interesting to do. Because now I have to go out and find something that's interesting to do on my own. And I can't make everything be monetarily remunerative...
Long-time Slashdot reader destinyland writes: He began with an update on a project to test the performance of the same algorithm using 30 different programming languages, and soon tells the story of how he was inspired to apply for his first job at Microsoft after reading Hard Drive: Bill Gates and the Making of the Microsoft Empire.
I decided that this is where I wanted to work, because these guys sound like me, they act like me, they are what I want to be when I grow up. And holy cow, they pay them well, apparently. So I wrote to everybody that I could find that had a Microsoft email address, which was about four people, because I had a software product people had been regisering on the Amiga. And one guy, Alistair Banks... responded and he hooked me up with a hiring manager directly in Windows that had an open slot that was hiring... And a couple of interview slots later, I wound up as an intern at MS-DOS working for Ben Slivka.
So you would think, "Oh, an intern on MS-DOS. What'd you do? Format disks?" No — it's amazing to me, actually. They give you as much work as they believe that you are capable of, and — they get you for all that you're worth, basically. They had me write a bunch of major features, like the Smart Drive cache for CD-ROMs was the first thing I wrote. Then I wrote DISKCOPY, making it work, single pass, bunch of features in MS-DOS. I re-wrote Setup to work on a single floppy disk by using deltas and patching in place, DOS 5 to turn it into DOS 6, something like, or maybe it was DOS 6 into 6.2... A whole bunch of features, within the span of, like, three months, which to me was fairly impressive at the time, I thought. And that only got me an interview...
Later he says that he'd like to see most of 16-bit Windows and all of MS-DOS open sourced, along with some select application code from that era.
I don't think there's any reason to hold back any of MS-DOS at this point. They have absolutely no reason to open source any of it, really — other than PR, because all it brings them is potential liability, complaints and angst, and probably nothing positive for putting the code out there and exposing it to ridicule. Because it's ancient code at this point. It's like, "Ha! Look what Microsoft did!" Well, yeah, I know Linux is cool now, but go look at Linux code from 1991 — and I worked on some of that code. Well, '93 I did. It's not the same as what you see today.
So yeah, MS-DOS probably looks archaic — although it's super tight, it doesn't have many bugs. It's just written differently than you would write code today, because you're targetting something that is a very different CPU and memory system and PC as a whole, and it's so much more limited that everybody's sacred, every cycle matters. That kind of thing that you don't worry about now. But I'd still like to see all the code from back then that's not embarrassing released.
And when asked what he misses most about being a Microsoft developer, he answers:
I miss going for lunch with the people that I went for lunch with, and talking to the people that I worked with. Because they were a lot like me, they had similar interests, they had similar abilities, they were people like me. We went for lunch, we ate food, it was awesome, and then we talked about cool things. And we did that every day. And now I don't get to do that any more. I get to do it rarely, because I take guys out for lunch and stuff, but it's not the same. So that's really what I miss.
And I miss somebody always feeding me something interesting to do. Because now I have to go out and find something that's interesting to do on my own. And I can't make everything be monetarily remunerative...
A fantastic channel (Score:5, Informative)
Dave Plumber wrote the original version of Task Manager single handed so he knows his onions wrt Windows. But he's also clued up on linux, mac and embedded systems too. A true geeks geek. Well worth subscribing too.
Re:nope (Score:4, Informative)
vxworks/unix/plan9 and you think DOS is tight code?
Older versions of DOS would run on 64kB RAM machines. This was important because the original IBM PC only had enough sockets to get to 64kB. Maybe you could do the bank switching trick with stacked DIP chips to get to 128kB, I dunno. Mine had an AST ISA card with another 384kB on it, designed to get a PC XT (which had sockets for 640kB) up to 1MB, so I had 448kB RAM which was enough to run almost any DOS software that would run on a PC 5150.
Show me vxworks, any version of Unix, or plan9 running in 64kB...
Re: (Score:3)
System V Unix ran on PDP-11 systems with only 64 KiB RAM -- the processor natively used 16-bit addressing.
Now show me any of the systems you named that had such limited functionality as those "older versions of DOS".
Re:nope (Score:5, Informative)
BZZT. PDP-11s supported substantially more than 64 KiB of RAM after 1975. System V first came out in 1983. The kernel alone for its predecessor, Unix version 7 was 40 KiB.
Re: (Score:2)
The original IBM PC, model 5150, was available in several configura
Re: (Score:2)
Maybe you should watch his channel. He has no need to monetize anything, the dude is doing it for fun, and any programmer worth his or her salt should watch what he has to tell, which is a hellot.
Re: (Score:2)
One of his videos shows a tour around his house. You can be pretty sure he retired from Microsoft pretty loaded.
Re: (Score:2)
He's more than loaded. He retired from Microsoft around 2003 or so. He started another company selling software and sold that for several million dollars.
He's made not just Microsoft money but successful software company money as well. He only sold the company because he wanted to code for fun, not code to a requirement or schedule, or even worse, manage people, something owners of businesses tend
Re: (Score:2)
Re: (Score:2)
To be fair, DOS didn't have to do much other than load some drivers and start programs, its not an OS in the true sense of the word.
they did. MS-DOS 1.0 & 2.0/2.1 were open sourc (Score:1)
did everyone forget?
Re: (Score:2)
Later versions are where all the interesting stuff is.
Microsoft has released some stuff as open source lately, but I have a feeling that MS DOS needs more work than they are willing to put in. For a start, like most operating systems, it's really a collection of different applications and tool, all of which need to be checked out for legal issues.
There may be incriminating stuff in there too, comments about deliberate incompatibilities that were designed to sabotage DR DOS and the like.
Re: (Score:1)
Re: (Score:3)
Re:they did. MS-DOS 1.0 & 2.0/2.1 were open so (Score:5, Interesting)
The other thing people don't realize about DOS is there is more to it than just having the source and the tool chain to build it. Things like EXEPACK were uses on the binaries and based on overvaluations like somethings being packed, some not, different compressors used it appears this was all probably done manually; when preparing releases.
So even if you had all the source you can't probably do the equivalent of 'make world' and get the entire OS as a result. Not that its needed for most 'uses' of the source at this point but the retro-computing crowd looking for a truly authentic experience is still going to be hunting down old copies of binaries.
Re: they did. MS-DOS 1.0 & 2.0/2.1 were open s (Score:3, Informative)
The 'tool chain' to build MS-DOS was a PDP-10 minicomputer.
DOS did not compile/assemble natively like a modern BSD or Linux.q
Re: (Score:2)
Re: (Score:2)
There may be incriminating stuff in there too, comments about deliberate incompatibilities that were designed to sabotage DR DOS and the like.
Doubtful that any of those once-smoking guns would still be smoking at this point; statute of limitations for civil cases is only a couple of years in most US jurisdictions (not sure about worldwide though).
Re: (Score:3)
Still bad PR to have the rumours proved.
Re: (Score:2)
Doubtful that any of those once-smoking guns would still be smoking at this point; statute of limitations for civil cases is only a couple of years in most US jurisdictions (not sure about worldwide though).
It doesn't matter that the "guns" are no longer smoking. What mattered was that they were fired at the opposition at the time the opposition had a chance.
DR DOBBS Journal had an article about Win3.1 installing on top of DR DOS producing an "error" msg and then aborting the install. DR DOBBS tech staff replaced the part of the code with checked for DR DOS with NOP's and then did the install of Win3 on top of DR DOS without problems. Win3.1 ran fine on DR DOS, but too many people saw the "error" msg and
Re: (Score:2)
Re: (Score:2)
Look up the "discovery rule". The statute of limitations doesn't even begin to run in many places until the facts of the wrong come to light.
Re: (Score:2)
Re: (Score:2)
If anyone is curious, it's not just really old versions of MS-DOS [github.com] but you can also get sources for GW-BASIC [github.com] as well as the old File Manager [github.com] (plus updates to run and build on win32).
I wouldn't mind seeing newer release of DOS, like a version with support for subdirectories and multilingual support. Or some of the more obscure GUI things like Comic Chat [wikipedia.org].
Re: (Score:2)
Yes. And not a surprise. MS-DOS was never very good. Basically all competitors wiped the floor with it, but not many people noticed.
These days, you just go FreeDOS if you really want DOS.
It should be out already (Score:2)
Re:It should be out already (Score:5, Informative)
Even if the copyright no longer applied, MS still would be under no obligation to provide the source code. Since they never published the source code in the first place.
You would be able to freely copy and install MS-DOS in binary form freely though, but not source code unless it had been actually made available.
Re: (Score:3)
A retired engineer could probably release it off old floppies or tapes if there were a reasonable expiration date to software copyright.
Re: (Score:2)
That might still be another issue. If the IP holder *never* intended it to be released, it would probably be another non-copyright claim for leaking code thtat would be out of copyright had it been released, but it was never released.
Re: It should be out already (Score:1)
Too late (Score:1)
MS-DOS will be mostly abandoned because it can only works on old or virtualized hardware. BIOS are disappearing in favor of UEFI which turn DOS unusable.
And for virtualization open source is not a requisite. DOS developers are decreasing very quickly because of that.
MSDOS code, although as a "leak", has been open for more than a decade. FreeDos was a reasonable alternative from scratch. But all of this has becoming abandoned with the removing of the BIOS even UEFI compatibility mode.
Re:Too late (Score:5, Informative)
MS-DOS will be mostly abandoned because it can only works on old or virtualized hardware. BIOS are disappearing in favor of UEFI which turn DOS unusable.
Companies are still making old school PCs. Most of them are very tiny; they are ISA SBCs at the largest, or more likely PC/104 bus SBCs which are the same size as a PC/104 card (quite dinky.) Most of these seem to be 386s. There are tons of these machines still in use in industrial control; if it ain't broke, don't fix it.
Re:Too late (Score:4, Interesting)
Agreed. One of the new "old" systems I'm looking forward to is the Commander X-16. That looks like a lot of fun to tinker with.
Re:Too late (Score:4, Interesting)
Don't laugh, but DOS is a pretty darn good realtime OS. Yes, all the onus of programming winds up in the application programmer, both having to create a .sys file, a TSR, as well as an active application, but you are guaranteed that at a certain time, you can fetch a packet, or move a stepper motor. For something that doesn't require the complexity of an embedded SBC, or even an Arduino, a MS-DOS based MCU can do the job, and there is not just a lot of x86 assembly documentation, but a ton of tools and compilers.
Of course, it isn't my idea of a platform, but it does do the job, and can keep doing the job for years in the future.
Re: (Score:2)
Re: (Score:2)
MS-DOS may fade away, but I booted FreeDOS last week to run Spinrite 6 and fix a current hard drive. Real-mode DOS has very limited, but still valid, use cases.
I would not bother fixing up an MS-DOS 6.22 bootable anything. Not worth the effort.
ps - MS-DOS 2.x ran on the IBM DisplayWriter. Most common use was to run a legal firm conflict manager app. Time and billing systems also, like 'TABS'.
Re: (Score:3)
Subscribe to the User mailing list: https://sourceforge.net/projec... [sourceforge.net]
Re:it already was (Score:4, Interesting)
Re: (Score:1)
I always viewed MS-DOS as more of a disk-loaded firmware. This was especially evident when programming in 16-bit DOS. Many of the calls were to the BIOS.
Most of the power of MS-DOS (especially the post v3.0) comes from the utilities packed with the OS, rather than the Kernel and command interpreter. In some ways, MS-DOS can be considered a bootloader, too. Some "Booter games" took advantage of this in the early 90s.
Just as old OS-es like this it is also worth studying old boot loaders. Recently, there was s
I'd like to see Windows 7 opensource much more! (Score:1)
Death to WIndows 10!
Re: (Score:3)
Not going to happen for exactly that reason. ReactOS [reactos.org] is a better chance at that goal.
Windows 10: Destructive to Microsoft's reputation. (Score:3)
The poor management of Windows 10 development has been extremely destructive to Microsoft's reputation. My opinion.
The ReactOS News is interesting. [reactos.org] I like the ReactOS Gallery. [reactos.org]
65 specific reasons for wanting ReactOS [reactos.org]
Sometimes the writing on the ReactOS is excellent. Sometimes it is not perfect. Good communication has a very positive long-term effect.
Writing error fixed: (Score:2)
Re: (Score:2)
So He Dined With Those (Score:1)
Re: (Score:2)
Yes. Scumbags the whole bunch and they never paid for their crimes.
Longing for youth... (Score:5, Interesting)
Sounds like a guy that misses being care-free and enjoying his 20s and/or 30s. I miss it too. The hanging out with friends and talking about all the crazy things that came through our heads.
As you get older, responsibilities of family come to (almost) all of us, and you cherish the time you had with your friends more than you did back in the day.
This last year has made these feelings of nostalgia more potent, since most of us in our 40s, 50s, and older haven't been able do all the fun things in life and more stuck with the grind of work.
Re: (Score:3, Insightful)
As you get older, responsibilities of family come to (almost) all of us,
To be fair, it doesn't just come to people, they choose it. Many seemingly without having given much thought to how significant of a commitment it is, with many crappy childhoods resulting from it.
I thought about it and decided I had no desire to spend the next few decades of my life doing little else than catering to my children, so I didn't have any. Closing in on 50 now and have no regrets. I get to sleep in every day off and all my disposable income gets spent on things and activities I enjoy. Admittedl
Re: (Score:2)
Re: (Score:2)
I do not think you are wrong, many crappy childhoods do result from people who neglect their children, but I think that has more to do with modern society than parenthood. Done right the relationship between parent and child can be reciprocal and beautiful. I am childless myself in my mid-forties, but I teach and the kids I teach easily bring me the greatest joy in my life. Seeing the world through their young eyes gives me back a little bit of my childhood wonder.
Modern society emphasizes compartmentalized
Re: Longing for youth... (Score:2)
Re: (Score:3)
Re: (Score:2)
Part of it is that the field of IT evolved very rapidly in the 1990s and 2000s. Part of the nostalgia is that in the past decade, little other than ad slinging, data slurping, and more intrusive DRM has been invented, especially with the move to cloud services. At best existing protocols have received some security updates. Anything new tends to be of a proprietary nature.
Platforms have also changed. The locked down environments on modern day computers have made it difficult to make anything new and exc
Re: (Score:2)
Well that's the nice thing about FOSS. If one wants to be a cowboy programmer they can and no one can tell them different.
Lunch in Silicon Valley (Score:3)
I miss going for lunch with the people that I went for lunch with, and talking to the people that I worked with. Because they were a lot like me, they had similar interests, they had similar abilities, they were people like me. We went for lunch, we ate food, it was awesome, and then we talked about cool things. And we did that every day.
I was visiting a friend in Silicon Valley once, some years ago, and we went out for lunch with a bunch of developers we knew. The talk was about how this or that web browser needed to handle this or that condition better. It briefly seemed confusing because these were browsers I'd never heard of. I leaned over and asked my friend, and he explained that they had decided to see if each of them could write a fully-functioning web browser in a minimum amount of code from scratch. One was writing in Scheme, another C++, another in assembly, etc., each using their favorite tools and comparing notes. Just for fun. It was more like a book club than a contest, since they were making suggestions to help each other. Oh, I should mention that they were all MIT LCS / AI alums.
So, yeah, hanging out with like-minded people who grok the same things you do is important.
It's dangerous to open-source old proprietary code (Score:5, Interesting)
Developers get complacent and put comments in it that they assume won't be seen by anyone outside the company, such as "This added here to break compatibility with competitor XYZ" or "That's broken but fixing it is too costly and nobody's gonna know so we won't be sued".
One of the companies I worked for in the past open-sourced one of their flagship products back in 1999, after careful sanitizing. But some things slipped through the review - namely one former developer who was named in one of the comments as having "fucked it all up but now we can't remove this shit otherwise it would break compatibility" - none of which was really true, but there: the poor guy's name was out there. Not cool.
I'm assuming MS isn't all that keen to spill its former guts because it would require a lot pointless of work to clean up the comments and remove former dirty company secrets and developers' opinions that shouldn't have been there in the first place, and also because the code speaks for itself - if you know anything about DOS internals, it ain't pretty.
Re: (Score:2)
While it would make the code much harder to consume for the OSS world, It would not be much work at all to simply strip all the comments; if that was a concern.
Without them "proving" anything even if it seem unnecessary out or place was there only for some anti-competitive reason is going to be nearly impossible. Maybe it looks like it was there to break compatibility with Dr.DOS or something, or maybe it was there because the 286-LODALL fakery in some 386+ bios they tested on some obscure Japanese clone wa
Those are the best parts (Score:1)
"That's broken but fixing it is too costly and nobody's gonna know so we won't be sued".
What's really a shame is, this is exactly the same thing that would bring a ton of value to open sourcing old by key software like MS-DOS...
I wish there was a way companies could dump source into an archive that would be open sourced in 100 years, or whatever period of time was well beyond suing range... then you would still get the full historical value of releasing the source without any of it being scrubbed. But eve
Re: (Score:2)
I wish companies could be required to dump source into an archive that would be open sourced in 100 years in order to get IP protection for their software.
Refining your wish with my own fantasy, there.
Re: (Score:1)
Yes, fully agree!
Re: (Score:2)
And one of our customers ran strings on it. Instead of seeing as an attempt by the customer probing for weaknesses in licensing the management were upset by the developer. But he goes back a long ways to the founding days of the
The Prime in 30 languages Github (Score:2)
There is no link to his github where he posted a Prime algorithm in 3 languages and people contributed dozens more. It's here [github.com].
Should be done (Score:2, Interesting)
Of course it would make sense to open up MS-DOS.
However even though MS-DOS looks like it's highly efficient with very small code, back when it was important, it wasn't seen that way back then. For example there was PTS-DOS which achieved essentially the same in far less code, freeing up lots of conventional RAM.
OpenDOS (Score:5, Informative)
Re:OpenDOS (Score:4, Insightful)
He isn't arguing for open sourcing old versions of DOS because he thinks they're going to run on anything being released today. He's arguing for it because he thinks the code is worth putting out there for people to read and learn from.
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
EA Play is also using DosBOX for some games in it's PC subscription.
Yeah, um, paragraph 2? Helloooo? (Score:1)
Sounds to me like Microsquish milked an intern for valuable features and didn't compensate him for them. Am I missing something here?
Re: (Score:2)
No he definitely got paid for his many years of work at MS...
Well, it's NOT that great (Score:3, Interesting)
Re: (Score:1)
Well, that pretty much supports the TLD -- people who think they could've done better can't resist mentioning how "shitty" it was - even though the article mentions how petty and anachronistic that is.
Re: (Score:2)
In a fit of politicizing programming all the "S" will be replaced with "$".
Re: (Score:2)
Does not surprise me. Basically all competitors were significantly better and years ahead.
Who care as we have freedoms? (Score:2)
Anymore?
Considering what Microsoft paid for it (Score:2)
Re:Better to bury it (Score:5, Interesting)
DOS was a good OS for its days of the mid-late 1980's
The standard PC of the time,
used a 16bit CPU around 1mhz clock speed.
256-512k of RAM (640k were for the high end systems)
Hard drives were expensive add ons. Most PC had 1 or 2 double density 5 1/4 floppy drives (360kb)
Standards such as Serial and Parallel ports, CGA or Hercules Display. Modems were Hayes Compatible, and Printers were Epson compatible.
DOS was designed to to be small and get out of the way to allow the program to run. The programs themselves would interact with the hardware. As hardware was expensive back then, the likes of Printers and Modems, had built in their own command language. Modems ad the AT commands (to dial a number with a phone contract that you using a dial phone, would be ATDPnumber or if you had the new touch tone, ATDTnumber. Dumb terminals uses ANSI or VT100 codes (They were very close), Printers used Epson esc key language, Advanced printers used Postscript. Because they all more or less talked the same languages, you didn't need drivers. Also because the computer was so under-powered, you are only going to be running 1 app at a time.
However the real sin was when the 386 CPU came out, we were still using DOS, and Windows up to around 2001 (Windows ME) was still based on the MS DOS Core. Where the system had a lot more power and could handle multi-tasking and other features so much smoother.
I think today we are in an other gap where the OS isn't well suited for our computers. My Laptop which is already a few years old has 12 CPU Cores. Yet most applications only use 1 core, and most programming languages today make multi-threading something to do if you can't do it in one thread, and having to use strange calls to get things done.
Re:Better to bury it (Score:5, Interesting)
However the real sin was when the 386 CPU came out, we were still using DOS, and Windows up to around 2001 (Windows ME) was still based on the MS DOS Core.
DOS was nothing but a handful of disk access and file system routines laying around in memory while Windows 3.1x or later was running. DOS was more a boot loader and set of libraries for BIOS functions and file system operations at that point - it was harming or limiting the experience at all. With the exception of cases where Windows was falling back to the DOS routines for disk I/O, all DOS represented was a little memory overhead. It was not limiting the system in anyway at all - that was entirely about Windows 3.1x on
Re: Better to bury it (Score:1)
The earliest IBM pc used a 4.77 Mhz clock. The common clones ran at 8-10.
Re: (Score:2)
Yeah but the early CPU's were all microcoded, so most instructions took multiple cycles, so effectively about 1 MIPS.
Re: (Score:3)
The first Mac was running a gui with 128kb of ram.
The first Amiga was running a preemptively multitasking gui with 256kb of ram.
The computers were not so much under powered, dos was just a very poor attempt at an os.
Re: (Score:2)
The Mac was mostly limited towards running smaller programs than the PC did, with the exception of some desktop publishing apps that used the OS Features.
The Amiga, while ahead of its time in many aspects, really failed to catch on, out side of gaming, and Video Production.
It was a case that the OS hindered the performance of a lot of the apps for the Computer, even gaming for windows never became a thing until Direct-X came out, which allowed to bypass a lot of windows for direct screen access, were games
Re: (Score:2)
Lack of success is not really the issue, as that has nothing to do with the source code being open or not. The fact is dos was significantly inferior to versions of macos or amigaos that were available at the time.
The OS still hinders performance of apps, and it actually makes a bigger difference today because systems are more complex with far greater abstraction. This difference is offset by the greater diversity in hardware which makes it impractical to program the hardware directly and take full advantag
Re: (Score:2)
The main problem with Mac (and Amiga and Atari ST) is that it wasn't made of off-the-shelf chips and semi-open specs that launched a cottage clone industry. No technical advantage can overcome the sheer volume and competitive pricing of PCs. Even IBM's own PC was beat by the (clone) PC itself. Mediocre graphics, terrible sound, difficult to use. But everyone had it and from the mid-1980's and on it got cheaper every year. Things would have been different if clone Apple II's were allowed to flourish later in
Re: (Score:2)
Re: (Score:2)
The ST's MMU, DMA, and shifter were not off the shelf. Probably difficult but not impossible to second source. GLU could have been made with discrete components but such a clone would have spanned multiple PCBs.
Regardless, the question was about operating systems, and MS DOS was awful compared to the available options. A standard PC generally had no less than 256K of RAM (yes, the originals "came" with 64K, but nobody actually used 64K PCs, PC DOS 2 barely ran in that.
The price was right for DOS. It did what people needed, stored files on floppy disks and let you load your program.
I agree on the RAM part. I think my XT came with 256K loaded on it, and most people quickly upgraded it to 512K. The XT of course was a little later than the PC it hits kind of a sweet s
Re: (Score:2)
The amiga actually did have clones.
Commodore would happily sell you the chips if you wanted to make your own custom motherboard, and several companies did this.
They would also happily sell you motherboards that you could house in your own custom cases.
The OS was also modular enough that you could produce your own hardware without using any of the commodore custom chips. Some companies did this (eg the macrosystems draco). It wouldn't run most games and other software that programmed the chipset directly, bu
Re: (Score:2)
Macs and Amigas needed system ROMs to hold part of the OS.
If you wanted them to perform you had to patch those routines into RAM, which is faster than ROM. Especially on Amigas, some of which would support SRAM.
DOS was a poorly featured OS, but it also got fully out of the way of programs. That was a big benefit at the time.
Re: (Score:2)
AmigaOS could also get fully out of the way of programs, most games operated in this way. This only works when the program running knows how to directly drive the hardware, which was possible on the Amiga because there were a limited number of well known models and chipsets.
It's not done anymore for two reasons:
1) security is nonexistent if software can take over the entire system like this.
2) hardware is too diverse, you would need to drive it in a very generic lowest common denominator way (ie generic vga
Re: (Score:2)
security is nonexistent if software can take over the entire system like this.
Surely there must be some way to encrypt the filesystem yet let the game operate. Maybe all its files live in some kind of archive. Which I guess is only a good idea if your filesystem has some kind of redundancy...
Re: (Score:3)
Re:See how that flies. (Score:5, Informative)
Sure. Expose the fact that Microsoft Windows is still just a fancy wrapper around DOS. See how that flies at Corporate.
That literally hasn't been true at all since Windows ME, which used DOS only as a loader. The last version of Windows which was even capable of running on top of DOS was Windows 98, and it did that only if you were loading protected mode DOS drivers because you needed them for some kind of hardware, almost always a CDROM drive.
Nice troll tho
Re: (Score:3)
I remember working as a PC tech back then; If I saw 98 in DOS mode it was a strong hint that the system had a boot sector virus. It would come in the door with "My CD doesn't work." and the failure chain was that the boot sector virus caused it to load into DOS mode and since the default config.sys didn't load ATAPI.sys by default that killed the CD drive.
I know it's nostalgia knocking off the rough edges, but I do miss those days. The work was easy (compared to changing Truck Tires), the money was amazin
Re: (Score:3)
Windows ME was really just a third respin on Win98. I seem to recall people had patches to restore 'MS-DOS Mode'
I don't think there were any changes that prevented 'exiting windows' and returning to DOS other than the UI. If there were it was simply because they were no longer bothering to protect/restore DOS structures in memory while Windows was running. If memory serves though the win9x disk subsystem Dragon/LADDER still could use the DOS routines as a fallback if 32-bit disk access was not usable just l
Re: (Score:3)
Re: (Score:2)
Windows did use DOS's file system APIs
Very inaccurate recollection of the Windows evolution. Early versions of Windows were explicitly described as a "graphical operating environment" that ran on top of DOS, which was the actual machine operating system; Windows was little more than a big collection of libraries that abstracted some of the BIOS (not DOS) functions; e.g. int 14h BIOS serial calls.
Starting with Windows 3.1 there was a 32-bit sector-level disk driver that overlaid/replaced the int 13h BIOS services, however the filesystem access
Re: (Score:2)
Windows 2 could already do things like swap out code segments while there were references to them in the call stack. It relied on DOS for filesystem access, but it was well on its way to providing operating system functionality.
Re: (Score:2)
As I said, the transition f
Re: (Score:2)
https://packages.microsoft.com... [microsoft.com]
You can already install microsoft packages with apt-get or yum, they make a bunch of stuff for linux now including powershell, edge and mssql, and they actually publish a proper repository instead of creating their own stupid installer.
Re: (Score:1)
Why, because he reminisces about the good old days? Maybe he just didn't allow all the joy to be squashed out of him by his work and adulthood.