|Home » Resources » Rants
MS ActiveVapor Revisited
Week of April 19, 2005
Applications do not an operating system make.
Jim Allchin's current pick for top trick in the bag is getting up in front of a crowd of Windows fanatics and running good old VisiCalc on a preview of the coming ('real soon now') Longhorn.
The irony is thick.
For it is precisely this product which first got Microsoft started down the path of actively vaporising the competition in the first place.
The place was Las Vegas; the time was October 1983; and the event was of course Comdex. 'MicroSoft' was at the time a lowly two-bit vendor operating out of a high rise in Seattle and Bill Gates, on hand in his humble little stand for the event, was totally unprepared for what was about to happen.
VisiCorp, creators of VisiCalc, the first-ever 'useful' application for a micro-computer and the program credited with getting the personal computing market going, showed up at Comdex with a bomb: they had a GUI under development for the IBM PC.
It was called VisiOn and it was not finished but that did not matter. It caused a sensation - so much in fact that Bill Gates closed his exhibit to wander over and find out what the commotion was all about. He was caught with his knickers down.
Not to be outdone or trumped by anyone (and a good poker player to boot) Bill did what he's always done best: bluff. He started heckling the VisiCorp people, loudly objecting to their presentation.
On what grounds, one might ask? Simple: VisiOn was not a 'finished' product and according to the increasingly vociferous Bill Gates, this was unethical.
But it was more. With Bill Gates and his Huns at Microsoft, it always is.
His own corporation, Bill Gates shouted out at the VisiCorp exhibit, had a GUI under development too - but they, said Gates, were far too ethical to come to Comdex and pretend they had a product ready for market when said product was not yet ready.
['MicroSoft' had no such product at that time of course, but this part of the story comes later. Ed.]
And Bill stayed on and heckled and heckled until he'd got everyone's attention and basically taken the glory from VisiCorp. And then in a resounding finale he marched demonstratively back to his own exhibit, closed said exhibit in protest against the unethical behaviour of his colleagues at VisiCorp [sic], and took the train home to Seattle.
Where he promptly called an emergency meeting of MicroSoft, gave the delicate assignment at hand to Harvard buddy Steve Ballmer - and 'Windows' as we know it (and wish we didn't) was born.
They probably didn't even have a name. Ballmer probably just stumbled onto it in one of his numerous telephone conversations with curious journalists who wanted to know more about the MicroSoft project Bill had told the world was already several years old [sic].
When Windows finally came out in December of 1985, not just a few wondered how it could have taken so long for such a lacklustre product, but they were missing the point (they'd been suckered in by Gates and Ballmer).
Work on Windows started after Gates' return from Comdex; when it hit market it was just over two years old. Gates and Ballmer had however convinced the media a long time earlier it was over twice that long in development.
Nothing could have been farther from the truth. And with Apple's OS X Tiger right around the corner, we're back there again.
To follow the media development of this engrossing story, one would think we have a new war on our hands - a 'features war' much like the 'browser war' of old. Such is hardly the case however.
Granted that a lot of online resources will compare Windows Longhorn with Apple's Tiger, and granted that CNET as always will make the most out of it to Microsoft's advantage, but this is not about features. Features are, after all, only a bunch of applications and anyone can write applications.
What Microsoft and Apple are supposedly selling is instead operating systems.
Not that the average punter realises this of course, but still the same. And whilst the Apple-friendly media are right in seeing Microsoft's and CNET's lastest tactics as 'vaporware' marketing, they too are missing the point.
That Apple will have a Spotlight magnifying glass in the upper right and Microsoft plan to have the same does not matter. That Microsoft want icons within icons does not matter either. That Microsoft plan to have window translucency - something Mac users have had for five years - even that doesn't matter.
It does matter that once again Microsoft are touting vaporware and they're getting the media to eat out of their hand, but this is not a features war. It's not about Microsoft's penchant for lying to the market.
It's about something much worse.
Microsoft have an absolutely horrid development model. They use a language Alan Kay has described as one of the biggest mistakes ever made in IT, and developers are no less aware of this than Kay.
On top of their poor choice of programming languages Microsoft have the leakiest, worst implemented 'IDE' in the business, and it steals at every turn from that of NeXTSTEP and steals sloppily to boot.
At least half of all bugs and bloat come from the Microsoft development tools themselves. Netscape for Windows became much too wobbly to use precisely because Netscape were too deep into use of this 'IDE' before they realised what a loser it was.
The entire 'windowing system' on Windows is so clumsily hot wired that no one but no one would ever be able to salvage it. It was put together in an ad hoc fashion by a number of green wet behind the ears clueless novices recruited by MS HR boss Charles Simonyi and chosen precisely for these traits.
But that's still not the point.
The creators of BeOS discovered when they set about putting together their network ready OS that you couldn't survive any longer with a single user computer. It does not matter that there is but one user on the machine, that machine has to be multi-user. Things had changed that radically.
Windows was an ad hoc slap-on to MS-DOS that tried to 'vaporise' the market VisiOn would have taken. It was ugly and it was worse than bad. It didn't sell too well at all, but at least VisiCorp couldn't sell either, their product completely forgotten and relegated to Bill Gates' growing junkyard of former competitors.
That Bill and company would later hire on Dave Cutler and his Tribe to create a VMS for the PC changed nothing: NT was dependent on the ad hoc 'architecture' that went before and so could not hope in its wildest dream to be secure.
And all Windows XP and Windows 2000 before it represent are forced rewrites of code Digital Equipment Corporation successfully proved was their own - but sans the help and guidance of Cutler's DEC gurus who were by then long gone, tired of Microsoft.
Windows is a system that by definition - and at least until the New Millennium in practice - allows everything and only as an exception tries to stop anything. None of this can change because the internal architecture - or lack of it - prevents any change on that level. The file system cannot be secure because it cannot be limited to NTFS and even if it were, the permissions system is so unwieldy and archaic that no one is going to be induced to use it anyway. And FAT can never be ruled out, and FAT doesn't even have per-user permissions, so any process - at any access level at all - can access and overwrite anything on disk. At any time.
And it hardly takes a systems guru to see this. Countless are the exploits which simply crash Windows deliberately after having placed rogue files on disk in very sensitive places, scheduled to start up and hijack the machine on reboot.
OS X currently has zero viruses in the wild; Windows has over 100,000. And this is not so much a reflection of market share (Microsoft never had and never will have a market share one hundred thousand times or more larger than anybody) as it is - and everyone understands this intuitively - a reflection of how defenceless Microsoft's flagship is.
For it's one thing to find exploits in really dumb code in web applications such as Internet Explorer and Outlook but it is quite another to piggyback off the exploits and go to root on the disk and corrupt it. In the one case it's flaky code that 101 Comp Sci would be too good for; in the other it's a fatal system flaw that can't be fixed.
So while the media continue to play the tune of Microsoft and while they and Bill Gates himself realise there is nothing to do at this juncture but heckle and try to put vaporware back in the market consciousness, those that aren't impressed by such age old tactics know it all goes a lot farther.
Applications do not an operating system make, and vapor cannot Windows an operating system make. Ever.