Flutterby™! : Jim Gray interview

Mark had a link to an ACM Queue interview with Jim Gray, head of Microsoft's Bay Area Research Center and recent ACM Turing Award winner. Mark quoted Jim Gray on development:

You see this today. Two groups start; one group uses an easy-to-use system, and another uses a not-so-easy-to-use system. The first group gets done first, and the competition is over. The winners move forward and the other guys go home.

And noted the "TerraServer SneakerNet" too, which is just the modern version of "don't underestimate the bandwidth of a station wagon full of mag tapes".

On the second, while I'm diddling with hardware and control systems, my cow-orkers are working on a large system to distribute and synchronize that data. We're talking about polling versus notifications, and net traffic across various bandwidth trans-oceanic links, and what can get scheduled and what needs to be synchronous. It's important to remember that a buck or three a gigabyte means that if you ask with a SOAP message if something's changed every 15 minutes, that's about half a buck a year. If you start doing that with a lot of computers, it's a simple multiplier.

Something the RSS guys need to keep in mind as they babble about the current polled architecture and replacements for email.

I also think it's worth noting his notes about development processes, and the reasons that Tandem failed, might subtly conflict with some of the things he thinks about why the big database vendors are on top, although he cannily refrains from making any real substantive statements along those lines.