Back from ArmonkArrived back from IBM HQ late last night. Had no time to blog there, so I'll write some of it up now.

As I mentioned earlier, I had the good fortune of talking with Stu Feldman, head of computer science research for IBM. I asked him what the thought about J2EE vs .NET in the emerging internet operating system. He said he didn't think too many people would standardize on .NET because it would tie them to a single vendor, but saw them both improving interoperability and sharing the space.

I asked Irving Wladawsky-Berger, who's responsible for IBM's Internet and network computing strategy and helped bring GNU/Linux into the company, the same question. He said that when the mainframe era ended, mainframes didn't go away, they just became an uninteresting and stagnant part of the industry. Similarly, he doesn't see Windows on the desktop going away, it'll just become a stagnant and uninteresting part of the industry as (GNU/Linux powered) devices, sites, and services surround it. This seems pretty reasonable -- if I think of the desktop over the past 7 years there have been no significant applications written for it. I don't consider the web browser an application, it's just a piece of infrastructure that renders the real internet applications, like Amazon, Google, and Yahoo!.

I kind of think the recording industry will go this way also. I don't ever expect them to stop banning Napster, but I do expect whatever victory they have in blocking their material being distributed online just leading to them being surrounded by open music that can be edited, traded, and mashed-up.

I also had dinner with Alfred Spector, Vice President of Services and Software. He was mostly interested with privacy and security, and felt that while things like Microsoft's Palladium were necessary they would never forbid users from executing code on their client machines.

Lastly, I met Sam Palmisano CEO of IBM. He was a really nice guy, and argued that since businesses underutilize most of their computing resources, they will begin outsourcing those to companies like IBM. This is kind of like the asp's of old, except instead of farming out top-level functions, they'll be farming out low-level processing. This is what IBM's grid computing initiative is about, and I think they're marketing it under the term "e-utility", which they aren't thrilled with and might change.

I'll close on some thoughts about IBM's purchase of PwC. Firstly, I would argue that IBM has no strategy. I'm defining "strategy" as "ownership of a scarce recourse or position that lets you charge a supra-competitive price and appropriate rent", and IBM's too big to specialize. IBM has major channel conflict issues with its software -- they have ISVs, VARs, SIs, IBM Global Services, IBM sales, and Lotus/Tivoli/WebSphere/DB2 sales, all of whom kind of compete with each other. But they don't care too much about it this, and just try to locally optimize as best they can. IBM bought PwC for high-level client relationships that will let them sell more involved services and systems to large businesses. Overall, I think IBM sees more growth in its services arm, but didn't want to go out and hire 30,000 people. So it bought them cheap.