Thursday, September 15, 2005

Two for the Show

The second day at was even more exciting than the first.

This is a great conference for exhibitors--the best I've ever seen. I've been in exhibit halls before where the number of exhibitors seemed to be equal to the number of attendeees; in those conferences, exhibitors are practically dragging people into the booths. Here, there are about 10,000 attendees (or so I've heard) and probably around 100 booths. The aisles in here are frequently packed. We've had so many people come by the booth. Surprisingly, many of them aren't ISVs--most seem to be "in-house" developers (or, as we think of them, potential customers).

Yesterday we had some great visitors that weren't potential customers. Scoble stopped by (that guy always seems to have a crowd around him: first of all, he knows everyone; secondly, everyone who doesn't know him wants to know him so there's always a gaggle of people around him. He's a rock star.). Anyway, he stopped by briefly and was glad to hear that our SDK is coming out. He may stop by with his camera today. That would be awesome.

Then, at the end of the day, Jim Gray stopped by. Jim Gray is positively a luminary in the field of distributed computing. His writings are very instructive. He happened by the booth when I was the only Digipeder here, so I had him all to myself. It was a great 15 minutes or so. I told him about the system in great detail (we've talked to him several times before, but not when we had a computer with us and could show diagrams, etc). I did a demo of "Digipede-enabling" the code behind a spreadsheet. He loved it. He talked about how chip manufacturers will soon be putting 4 cores to a chip (or more), and then manufacturers will be putting 4 chips in a box. Now you've got 16 or more processors; how do you keep them busy?

He went on to say that "no one" is thinking about how programmers will take advantage of that many processors--it's a large enough number that it's difficult to manage that many threads. However, he said that he thought that our Framework is the perfect programming model. It allows developers to think the way they already think--in terms of objects. We already have the ability to tell a machine to "start as many objects as you have CPUs."

Suffice it to say, in a world where he feels that the development tools are at risk of falling behind the hardware trends, he called Digipede "one of the only points of light" in this area. It was high praise, and it meant a lot coming from him.