The only reason Windows only runs on x86/x64 now is that customers weren't interested in it on other platforms. Microsoft really tried to make it cross platform.
They just forgot to port Office. IIRC, the then-current release of Visual Studio would run on x86 but compile to supported RISC platforms. That turned it into a server-only platform. No Office and no development tools before the dawn of the web application means certain desktop doom.
Lack of Office killed the desktop Windows on RISC.
On PPC, I remember SQL Server had some weird network bugs too.
This isn't really about technology, it's about convincing these enterprises that it's better for them to scale vertically on a single machine than to scale horizontally on commodity hardware. From a technical point of view this has long been proven false, but for certain companies it may be true from a business point of view: if technology is a cost center (i.e., it's called "IT" not "R&D") and you don't have the talent needed to operate a cluster of commodity hardware (e.g, there are either no operations engineers who can program in your geographic area or they simply won't want to work for you) leasing/buying big iron (especially if it comes pre-configured and with a support contract) starts to sound attractive.
Even if SPARC is superior hardware, the customer simply wouldn't care. It would be more profitable for IBM to cut SPARC off and continue only with POWER. The professional services involved in this is the lucrative part for the vendor.
People who do care about technology, have in-house talent and do HPC (scientific computing, machine learning/data mining, Internet companies with scalability problems) are best served by x86_64 (in almost all cases excluding some types of computation), ia64 as well as IBM's own Power-based 1U/2U servers (what used to be the RS6000 series).
Brian Aker says "I'm sure everything else Sun owned looked nice and scrumptious, but Oracle bought Sun for the hardware." http://radar.oreilly.com/2010/04/a-mysql-update-from-brian-a... which means the SPARC survives.
[edit: proper quote]
That would overlap with their POWER lineup. It's unlikely they would keep SPARC because it would eat away their POWER market share.
Don't know if it would, in the end, be bad for IBM, but the managers at the POWER side would pocket smaller bonuses. They would never let that happen.
Apart from that, it would be a good thing.
Not me, at least directly, because I live in Brazil and we have no such software patent nonsense, but tech business would be quite impossible where they are valid.
Also, not only were you "long on slogans and short on insight" but your argument doesn't fly. (IBM usually keeps the patents of the companies that it buys.)
IBM would want to keep the patents for itself, but Microsoft would probably wrap them in a do-not-hit-us term and sell them to trolls. This way they could undermine the competition while staying completely out of it.
Oh really? How about some supporting evidence?
The best evidence would include examples. After all, Microsoft buys companies all the time. If this is something that Microsoft will "probably" do, surely they've done it several times before.
Well... It's better than Microsoft, I guess. The thought of a "Windows 7 Server for SPARC Enterprise Edition Plus" is frightening.