October 18, 2005

Lessons for today

This blog is about history, and the parts I’m most interested in may not be the ones with the most immediate consequences. But here’s a list of some things I see from a historical perspective that can maybe provide insight for at least some people looking at today’s industry landscape.

On-premises and time-shared software used to be highly synergistic. There’s no good reason they won’t be again (although multi-enterprise bottlenecks may have crept into the real-time architectures that didn’t arise in the batch and partly-batch cases. Which reminds me …
A big part of the history of software development is clearing bottlenecks: Find one; clear it. Find the next; clear it. That’s a big part of the reason product evolution is so seemingly slow and incremental. Modularizing is a big part of the bottleneck-clearing process, but not actually a replacement.
Similarly, a big part of the history of artificial intelligence — and of software that has anything to do with text — can be described as “Start with a bad system; keep improving it.”
Oracle always had a NIH (Not Invented Here) syndrome about system software. Although there are increasingly many exceptions, that still seems relatively true as compared with other system software vendors. However, Oracle never had an NIH syndrome about applications.
Microsoft people used to be utterly driven. If it’s not that way any more – and I get the impression it isn’t — this is a big change.
There always are important conferences. Their identities just change. Mid-level marketing people take over the previously good ones, and newer more exclusive ones emerge in large sectors. In particular, new industry sectors commonly have great trade shows, conferences, and other gatherings. (Example: 2005’s inaugural Text Mining Summit.) A huge part of what I learned about the software industry was at conferences, and occasionally one is still very worthwhile.
The software industry is quick to use some technologies, slow to adopt others. Tracking which is which can be good guide in predicting which will eventually succeed.