SCO v. IBM reopened

Posted Jun 17, 2013 20:03 UTC (Mon) by viro (subscriber, #7872)
In reply to: SCO v. IBM reopened by JoeF
Parent article: SCO v. IBM reopened

Give me a break... "Didn't allow" != "printed a warning in debugging version along the lines of 'you are running it on unsupported kernel, don't bother us with your bug reports'". I'm not fond of MS (and I have even less fondness for their products), but extortionists are quite willing to go after very unpleasant companies, as long as they hope to get a settlement. Consider e.g. patent trolls - would you suddenly consider their brand of racket right and proper just because the current victim happens to be somebody nasty? If you do, you are in for disappointment when (and that's when, not if) the buggers attack somebody else in the same way...

As for "way better" part... I can only admire your ability to make distinction between one pile of garbage and another such pile. I've used both and there's too much CP/M in all of them. Junk is junk...

SCO v. IBM reopened

Posted Jun 17, 2013 20:05 UTC (Mon) by dlang (guest, #313) [Link] (13 responses)

It didn't just print a debugging warning, it refused to run at all. I remember patching my DR-Dos binaries to change the string that it searched for in order to be able to run the Windows software that I had purchased.

SCO v. IBM reopened

Posted Jun 17, 2013 21:05 UTC (Mon) by viro (subscriber, #7872) [Link] (12 responses)

Frankly, as far as I'm concerned, they had every right to calculate a strong hash of the kernel text segment and refuse to run the damn beta of their code with anything other than what they'd been interested in. The same reason why anybody bringing an oops from a system with binary modules loaded will be told to take a hike until they can reproduce the damn thing without said modules - developers' bandwidth being limited. Especially with the crap as brittle as *DOS and Windows 3.1. I've seen claims that some variant of DR-DOS managed to run that garbage "flawlessly"; I'm sorry, but I don't believe that it ran flawlessly on _anything_ (including MS-DOS).

It implemented things that would be a part of the kernel on any sane OS, ran with the priveleges equal to those of what passed for kernel there and dealt with very crappy userland running with the same priveleges and not shy of messing with the hardware directly. Trying to debug such a thing is a job for masochists and they apparently had failed. I had very little contact with that thing, but I'd seen enough hangs and crashes. I can believe that DR-DOS had been close enough to give more or less the same frequency of crashes when combined with that thing, but in the place of MS masoch^Wdevelopers I would try very hard to filter out all bug reports of that kind. OTOH, I can't imagine being desperate enough to work on such project in the first place...

Said that, my memories of that story had been of "nasty message" variety; it had been a long time ago and having seen the wonderful stability of aforementioned Windows I never had been interested in it - *DOS had been able to run text editor, C compiler and uucp clone, so lousy as it had been it was usable for some work and adding a crash-prone multitasker and GUI had been very low on my wishlist... I had dealt with DR-DOS (they had the damn thing installed in a school where I taught an after-hours optional class for a couple of years) and from what I remember it was just as lousy as MS-DOS. TBH, I'd rather forget both sorry excuses of an OS, along with all the shite spawned by CP/M...

SCO v. IBM reopened

Posted Jun 17, 2013 21:35 UTC (Mon) by dlang (guest, #313) [Link] (11 responses)

for those starting from a *nix capability, I'm sure that DOS and Windows seemed like toys.

But for those of us who did not have access to real *nix machines DR-DOS had numerous advantages over MS-DOS and Windows 3.1 was a significant step forward.

Also remember that at that point in time, Windows was not an OS, it was a shell running on top of DOS, including on the pre-386 x86 machines that could not do a good job of providing the OS safety you are used to.

A license for Xenix of Unix cost several thousand dollars (I remember being thrilled at being able to pick up a copy for "only" $1000 at a clearance sale when I got my first 386 computer around '94). I discovered Linux shortly after that, but was not able to make really good use of it for a couple more years due to video limitations. There's only so much you could do with Linux in those days without a graphics display or an Internet connection.

This wasn't checking a hash of the kernel, this was a simple check for the string MS-DOS at one particular address. And this wasn't in a beta, this was in the finished retail version. (although, you could argue that Windows is still in an extended public Beta, but that's a different discussion ;-p )

SCO v. IBM reopened

Posted Jun 17, 2013 22:36 UTC (Mon) by pboddie (guest, #50784) [Link]

Indeed, there were a number of DOS variants in the 1980s, and for a while there was even something like a degree of competition around providing a better product, although it doesn't take much imagination to realise that those apparently striving to offer the better product were not the "incumbents" in the personal computer business. Meanwhile, it has since been documented that Digital Research were most likely able to offer a DOS product without restraint because of the somewhat "contested" origins of MS-DOS.

The response from Microsoft was to tie two separate products and undermine any remaining pretense of product interoperability on the PC platform. One can always claim that a vendor has no obligation to make its products work with anything other than ones it prefers, but when that vendor indulges in anticompetitive behaviour in other areas, any tying of products should be scrutinised properly and in a timely fashion, not years after the damage has been done.

(And in case anyone cares, I never used DOS in any serious way. A lot of people who grew up using microcomputers regarded DOS as a complete joke, presumably because they only ever got to see the Microsoft version and not the ones that tried to evolve with the availability of increased computing power and hardware functionality.)

SCO v. IBM reopened

Posted Jun 18, 2013 12:23 UTC (Tue) by nye (subscriber, #51576) [Link] (8 responses)

>And this wasn't in a beta, this was in the finished retail version.

Wikipedia disagrees: http://en.wikipedia.org/wiki/AARD_code

Larry Osterman has a blog entry from 2004 that has a little more: http://blogs.msdn.com/b/larryosterman/archive/2004/08/12/...

SCO v. IBM reopened

Posted Jun 18, 2013 12:27 UTC (Tue) by nix (subscriber, #2304) [Link] (1 responses)

To me, the fact that the code was deliberately obfuscated suggests that they *knew* that they were up to no good. Sure, they were trying to avoid reverse-engineering, but they must have known that would be futile.

SCO v. IBM reopened

Posted Jun 18, 2013 15:17 UTC (Tue) by nye (subscriber, #51576) [Link]

I completely agree, but the point is that the check never made it into any released version of Windows, and was not a fatal error even in the beta - the one screenshot anyone's managed to dig up prompts the user to 'Press ENTER to continue'.

SCO v. IBM reopened

Posted Jun 18, 2013 13:51 UTC (Tue) by pboddie (guest, #50784) [Link] (2 responses)

"The AARD code in Windows was code to detect when Windows was running on a cloned version of MS-DOS," says Osterman. Already the terminology misleads the reader into thinking that anything other than MS-DOS was some kind of rogue product riding Microsoft's coat tails, potentially implying that they copied or licensed MS-DOS and made their own variants, when in fact Digital Research's products (upon which SCP/Microsoft had apparently infringed) were separately developed and delivered on successful hardware products in their own right.

Certainly, if by "clone" it is meant that something behaves like something else, then things like DR-DOS were clones, but the negative connotations of the word denies the origins of MS-DOS and the viability of the other DOS products. It almost asserts that the only legitimate DOS product was Microsoft's own.

Given that microcomputer technology stacks were often tightly integrated, with the hardware, operating system and applications often originating from or being delivered by the same vendor, the reaction to the AARD code is an indication of how attitudes were gradually changing, leading to what I'd like to think is a more enlightened perspective on such matters that many people have today.

SCO v. IBM reopened

Posted Jun 18, 2013 14:49 UTC (Tue) by anselm (subscriber, #2796) [Link] (1 responses)

MS-DOS started out as a »clone«, or, more exactly, a reimplementation of CP/M for the 8088 processor. (CP/M was originally for the 8080/Z80 series of 8-bit microprocessors.) This actually made reasonable sense at the time since the 8088 assembly language was close enough to that of the Z80 to allow for automatic transliteration of existing (Z80) CP/M programs to 8088 assembly; the only thing needed was a way of making the system calls work, and that was exactly what MS-DOS provided.

Various advanced features like a hierarchical directory scheme and stream-oriented IO (both arguably inspired by Unix) only came along with MS-DOS 2, which is why many software packages at the time stipulated a minimum requirement of »MS-DOS 2.11«. This is why we have »\« as a path separator on MS-DOS/Windows, since »/« was already spoken for through MS-DOS's heritage from CP/M (and further back to the DEC operating systems that inspired CP/M).

Digital Research (the company behind CP/M) actually came out with a version for the 8088 and 8086 processors called CP/M-86, but that never really went anywhere because of the success of the IBM PC (with PC-DOS a.k.a. MS-DOS). There were various 808x-based computers in the 1980s that weren't actually compatible to the IBM PC, and these generally used MS-DOS (rather than »PC-DOS«, which was the same but for the IBM PC) or CP/M-86, among other less important operating systems.

SCO v. IBM reopened

Posted Jun 18, 2013 16:34 UTC (Tue) by pboddie (guest, #50784) [Link]

Indeed, the origins of MS-DOS are no longer in dispute.

I'm sure you have more hands-on familiarity with CP/M and DOS (from different vendors), whereas my experiences were with other microcomputer operating systems, but it's interesting to note that it was Digital Research DOS Plus that was shipped on the Amstrad PC1512 (which was rather successful in the UK and Europe), and Concurrent DOS was also offered for some systems.

Certainly, Digital Research were a viable competitor to Microsoft in a market that itself had only developed because of cloning of the original hardware platform.

SCO v. IBM reopened

Posted Jun 18, 2013 15:16 UTC (Tue) by dlang (guest, #313) [Link] (2 responses)

and we all know that wikipedia is never wrong...

I never had access to any beta software from Microsoft, but I did buy and run DR-DOS and Windows 3.x

I ran into this bug personally and had to patch to work around it.

SCO v. IBM reopened

Posted Jun 18, 2013 15:21 UTC (Tue) by nye (subscriber, #51576) [Link] (1 responses)

I'm very sorry, but your memory of this must be imperfect. There are loads of citations, and they all seem to agree that it was not in any released version. See for example http://www.drdobbs.com/windows/examining-the-windows-aard... by Andrew Shulman who analysed the code in question at the time.

SCO v. IBM reopened

Posted Jun 18, 2013 19:24 UTC (Tue) by Wol (subscriber, #4433) [Link]

I tried to run DR-Dos. Likewise, I couldn't successfully run Windows on it. But with me it was different - the hard disk kept getting corrupted.

And like the GP - I wasn't using beta software - I had no access to it. I fail to see how you can accuse him of mis-remembering something as big and unusual as manually patching a commercial binary.

Cheers,
Wol

SCO v. IBM reopened

Posted Jun 22, 2013 9:02 UTC (Sat) by nim-nim (subscriber, #34454) [Link]

However, you'd be mistaken in thinking that this "toy" impression stemmed from any deep technical analysis. I fear a huge part of it was (is) only due to the scars left trying to find the budget to buy "real *nix machines", and refusal to envision that all this pain would be wasted.

You get the same phenomenon when someone buys an expensive car. It's night impossible to get him to admit any serious car default while the budget scars are still bleeding.