Why isn't there an Adobe OS?

Why hasn't Adobe built it's own operating system? I've been thinking about OSs lately, and reading up on SGI's Irix, a system that I would love to get my hands on and play with. It got me thinking about software integration and what I'd want in an OS.

There was a time, maybe 10 years ago when I thought that an integrated OS and software package might be coming around the corner. Avid had a pretty good set up using Avid Studio and related software that showed how one file type could be opened, manipulated and understood by a variety of applications. At the time, it ran exclusively on Mac OS, but didn't work with non-Avid apps.

Likewise, Microsoft was making Office, which is overburdened with too many widgets now but had potential at one point.

The other big software giants at the time, Adobe and Aldus were working on similar projects. Aldus was trying to integrate its two flagship products, Pagemaker and Freehand. Adobe was trying to get Photoshop and Illustrator to integrate. Around that same time, Adobe bought out Aldus, and added PageMaker (and later inDesign, which while branded separately -- and drawing its workflow paradigm from FrameMaker, Aldus' other major product -- seemed to gain it's parentage from PageMaker).

Sadly, these plans went nowhere (probably for legitimate reasons that I just don't understand). Instead, the new paradigm became "browser/desktop" and the battle was between Microsoft and Mosaic -- later to become Netscape)

Now, Irix has been abandoned by its creator, Silicon Graphics, and has not been made open source or freely available (if we're fortunate, maybe Adobe is working on developing it as the basis for a new OS! I dream, I know. Irix was designed to work on specific hardware using MIPS chipsets, but that same hardware now runs Red Hat Linux, so someone knows how to port code to and from that chipset).

What I would want in the ulitmate OS:

Super stability and uptime (hence my appreciation for Irix. Mac OS X comes close, but not enough. Both of these, by the way, are software/hardware combinations No one seems to have built a truly stable uptime OS that can be migrated from hardware to hardware)
A strong foundation in math coprocessing and mutlithread multitasking (points again to Irix)
Intelligent agent software (something like Devon Agent) that will allow you to set goals or priorities and learn from how you react to the results. I haven't had a chance to truly try out Devon Agent, but it sounded like a good start in that direction.
Integration of key applications - The Aldus dream was that a file could be open with a bunch of tabs for the applications that could affect it. So you could do some DTP on the file, click a tab and the toolbox, but not the file, would change and you could work in freehand ( a vector design application), then tab back when you were done.
Greater control over which apps have what permission for a file (in Mac OS X when I right click on a jpg and select "open with" I get about 10 Photoshop-related tools that I have no interest in on the list ahead of Preview and Photoshop - the only two apps that I actually open jpgs in).
True portability of documents - PDF comes close, but a much more editable PDF format, one that was open to word processing and image manipulation, would be ideal (hence the Adobe OS idea). Either the Open Document Format or OOXML might get there yet, but it's 2008 and still the lawyers are arguing over this stuff.
An XML/SGML-based toolkit so that us non-progammers could at least add functionality or personalization to key apps
And of course, while we're blowing smoke anyway, the obligatory "cool yet usable" interface - which I don't think the desktop paradigm is (mostly usable, yes, but not really cool) Yes, the pending 3d interfaces offer promise, but I'm holding out for a gesture-oriented interface where the workspace is larger than the monitor. I also want to be liberated from elements pinned to the sides of the monitor (status bars, application menus). I can't help but think that if Kai, of Kai's Power Tools fame had been properly courted by someone with deep pockets (again given his work with Adobe, why not them?) that we would have much more intuitive, non-linear interfaces by now.

Such a venture might be risky -- anti-trust lawsuits seem to be attracted to closed platforms. But complete devices as closed units aren't unknown - look at RIM's Blackberry. You can make a successful closed product, but to give it broad appeal might be the tricky part. Any takers out there?

— SGP