Sun has been notoriously behind with SPARC. UltraSPARC III was behind on arrival and UltraSPARC IV uses the III core (although there is a process shrink in there). While Sun servers have good throughput, their workstations aren't competitive and probably won't be, at least in the near term. Thus, the mass migration in many 3D houses to Linux on Intel (WETA or Pixar anyone?). And if Linux has the high end features and hardware support to do some of the higher end visualization stuff, Sun will become even less relevant in the workstation market (as will SGI and most other RISC workstation vendors).
Once again, you're presumptious that Apple is strictly looking at "WORKSTATION" development (Workstation does "NOT" equal "ENTERPRISE", it has nothing to do with "SERVING"). The truth is, Apple has quite a decent low-priced workstation line in their Pro models, that are comparable to the PC-based Workstations and better in many cases than what Sun and SGI build on the low-end. With PPC 970 coming online, and support for nVidia (in-house) and ATI, Apple has much of the same advantages of what Windows has built over SGI in the late 1980's and into the 1990's when SGI slipped. Of course, Wintel has these advantages in engineering and graphics for workstations (largely because of ATI and nVidia and other cards like Oxygen cards), but Apple has a key software base to work from to compete, and can leverage x11-based applications that Linux can... via Apple's support for x11 and of services like Fink.
SGI, Sun, and Linux lack "MANY" of the key software applications to do "SIMPLE" Office-style stuff, which most "WORKSTATIONS" rely on, as well as the hardcore applications that a "WORKSTATION" is known for. You can't fire up Photoshop, open Excel, Final Cut, After Effects, Vectorworks, Form Z, Archicad, or use Powerpoint. There's no SGI or Sun laptops... etc. etc. Sun is tied primarily to a proprietary hardware/processor roadmap used largely in "ONE" segment (Enterprise), and they've not scaled any processors down to the smaller markets to take advantage (increase development), nor have they had the infrastructure, being a smaller company, to focus on workstations... that is where Apple complements Sun, because Apple works from workstations down. The G4 Pro line is in effect, workstations, and equipped with 970's with OS X... they're a perfect match for anything Wintel throws at them, which is the "REAL" workstation segment. Linux isn't workstation-ready yet for much of anything commercially purchased, and won't be anytime in the "NEAR" future I don't feel, unless Adobe, Macromedia, AutoDesk, and Microsoft, et al. jump on the bandwagon....
Linux has support for doing higher end visualization because the people that work on those machines (those you mentioned) have the programming savvy to develop it for it. They could make their own OS if necessary, but it's highly impractical, hence they have gone with Sun in the past, and now favor an open-source Linux OS because they can take something pre-existing and morph it to their needs (it's not that the hardware is better, but their integration is, and their flexibility, for being so savvy, is immense... plus, no doubt... it's CHEAP!). It's largely in-house and proprietary software that reigns on Linux, made for specific tasks by people that want a more encompassing and flexible need, and have the tremendous resources to get it without lobbying to a company. It's a means to an end...
You can "BANK" on it that Linux won't replace the whole kit and kaboodle anytime soon, or make a "BIG INROADS" into Workstations. They lack the fundamental smaller apps... and I doubt we'll see Adobe making Photoshop anytime soon for Linux (they're tied strongly to Microsoft, with secondary support for Apple for the timebeing... which I expect Apple to launch a Photoshop competitor overtime... following their iDVD DVD Studio Pro philosophies... as well as their iMovie/Final Cut Express/Final Cut Pro philosophies... iPhoto/***/***?)... and GIMP isn't even in PS's ballpark yet. It's good... but not great, although given some work it could rival PS (won't change people's perception over night though), but it's so far from having a "Human Interface Guideline" set for both KDE and Gnome that it's ludicrous to even think it could rival PS.
The problem with Linux is... there's no leadership on the levels that there is in other systems (it's more like a tug of war), as Torvald's is a "TECHNICIAN"-minded opinionated individual, unlike Gates who is a visionary pirate, and Jobs who is a true visionary (with adept pirating skills). If Linux turns out to be a better choice, it'll still trickle down into someone tossing a proprietary system on top, and keeping the "Linux" name out of the limelight on the consumer end. Torvald's believes strongly in a lot of things, but he's also a "technical" man and until we see a dumbed down GUI on the order of what Lindows brings, only 10,000x's more refined, and with more applications and support... Linux will struggle getting to the desktop without someone like Gates or Jobs supporting it, and without an understanding of how to make "easy to use" and "compelling" equate "Linux".
For the giant companies with permanent, on-site, IT staffs and custom-written proprietary softwares... Linux is an option. For smaller to larger SoHo-type stuff and small-medium-sized corporations who pay for services though, small mom & pop local shops doing their own Linux servers will be competing against Sun, HP, IBM, Dell, et al. in terms of x86 equipped Linux servers that are "SUPPORTED" and "INTEGRATED" via trained technicians from each of the big companies, or via technicians from the local mom & pop computer shop. This is exactly where Apple's x86 rivalling XServes and XServe RAID's, and XServe Clusters fit, along with OS X Server equipped G4 desktops... as they compete in low-end servers with a very comparable, and competent, platform. It also comes without the tremendous needs for support staffs or services, as it's as easy to use as... well... a Mac.
For higher-end... Linux has it's champions (almost every vendor), but it also has a long uphill battle. One where everyone is entrenched, and to their right... they face Apple and Microsoft in workstations/desktops. Not going to be easy, especially without an intuitive GUI and a large roster of apps., games, productivity software, and workstation apps. Server software is relatively easy... the rest is difficult. Server hardware is hard... desktop hardware is relatively easy.
I don't see Sun's lead slipping anytime soon... and I don't see Apple just hopping up into the market and taking on the big dogs. That is why a Sun/Apple merger would work... because even if Apple transitioned Sun's mainstays away from say... Solaris and SPARC (over time, right away would be foolish) you would definitely see the technologies of Sun combined into something else, or see Apple embrace something comparable and worthy... perhaps even in-house PowerPC development (using Sun's team) or a next-generation SPARC or <your name here> architecture (as in, new, never brought forth).
So to expect Apple to just slap OS X on everything, as it is now, is naive. If OS X transitions to Sun, it'd do so based on a merger of Solaris and OS X, perhaps as a future OS that'd share the GUI and API frameworks (compatibilities) of an OS X, along with the API's and frameworks (compatibilities) of Solaris, riding on top of either a BSD or System V kernal... or perhaps even... Linux, or an all-new kernal. Those that figure OS X will be Apple's OS forever without any significant changes along the way are foolish. Let's face it, they're probably the same one's that bough "Apple II Forever" t-shirts, or believed that Mac OS Classic should remain forever. ::buzzer:: Not happening... and having a good future game plan, is a definite plus. Given time to get Solaris and OS X together... I think Apple would have great success. Whatever processor they choose...
Apple could dust off the old "OpenStep" (Cocoa) API's for riding on top of Solaris, update them to the current spec, and deply them on Solaris X (10), adding Carbon API's along with them. Suddenly, a simple recompile of programs like Photoshop, Illustrator, Freehand, Flash, Director, Fireworks, GoLive, DreamWeaver, Final Cut (Express/Pro), etc. etc. would run on Solaris X... and said developers could release the next "VERSIONS" of their software on hybrid OS X/Solaris CD's. XServe could be Sun's new low-end server line, and Apple could benefit from Sun's patents and technologies in ramping up PPC 970 workstations/servers (and lower end G4/PPC 9xx laptops/iMacs/eMacs) in the near term 'til they solidify their future roadmap and where they're going.
Will Apple buying Sun happen? Who knows?!? Yet it'd be a far better investment than Universal Music, that's for damn sure. Apple can license music from "ALL" of the labels and make a minimal profit per song, or they could start their own label themselves, focused on bringing the considerable wealth of Indie artists to the masses... in a way that none of the sharing services has been able to do.
Universal is a bloodletting, mired in so much red tape and bureacracy and the RIAA witch-hunt syndrome that it "COULDN'T" possibly fit Apple's loose principles and policies and loose handling of product. Sun is just as loose... more like a fraternity than a corporation. Apple's investment in Sun and Cafepress.com would return profits (maybe not all the time... but at least a good percentage of the time)... the Universal purchase would eat up a "TON" of cash reserve, and likely not produce a profit for quite some time, if ever.
