Random Musings

Back in January, I read a blog post about Van Jacobson’s talk at Linux Conf.au 2006. (LWN’s coverage and the slides) This is some amazing stuff. I can’t really do these justice - go read. The slides show some amazing numbers in terms of potential speedups through rearchitecting.

What prompted this entry, however, has nothing to do with that. Instead, it has to do with another Van Jacobson talk at Google that deserves attention - namely A New Way to Look at Networking. Where the other paper and talk were about performance, this talk is more about thinking about what’s next - what will the ubiquitous network of 20 years from now look like, and how to make it work. (I’ll spoil a slide near the end, as a teaser - PGP meets BitTorrent and hitches a ride on an airplane. (Not Pacific Air 121, however.))

At work this week there was a very good talk about “Scrum Et Al”, given by Ken Schwaber.

Executive summary: Go watch this, if you’re at all interested in Agile teams and iterative development, it’s definitely worth your time.

What struck me most while watching this, was how much I wished I had gotten around to reading something about it years ago. (I still haven’t, but this video is enough to get me interested now.) At my last job, we were doing something informal, but not too far off of this. Quarterly releases, along with a relentless pace of new features, improving interfaces, improving security, improving developer productivity, etc. Seeing this talk reaffirms my belief that the only good way to build software products is iteratively. Being able to say, “I only need to worry about X right now.” and sitting down and doing X, getting it done, clean, bugfixed, and then being able to say, “Ok, what’s next?”, all the while having a product that can be released at basically any time, maybe multiple times, and a simply relentless pace of improvements? This is great! Unfortunately, I don’t think very many business majors have figured this out yet.