Electric Josh

Thursday, May 16, 2002

So at work my superb cohort and I have actually been given a week or two in which to plan the next version of our program. Everything is going to be done correctly. We'll rework all our code that hits the database so that it instead hits the same database that the analytical model runs against. We'll insert an extra tier to better manage user state, and allow them to do a bunch of work and then either save it all or roll it all back. Everything will be faster and more scalable, we'll give users more visibility to their data, and we'll integrate with the architecture in the sky that's actually getting constructed around us. It's a great time to be alive.

So of course a pile of mundane deployment details are getting in the way; trivial problems with the current version that the existing customers should just be able to work around, dammit. And we of course go fix those problems before getting too bogged down in our thinking about what things will look like six or eight months out, delaying the fun part for yet another hour or two, but after this year and a half limping along with no design whatsoever, squeezing performance out of code and database schemae that should have died spectacularly three rollouts ago---after making ugly code do some surprisingly coherent things for that long, what's another couple of hours.

We get to be domain experts in front of other people rather than in front of one another; we get to toy with the analytical engine, and to demand that it be designed to take the inputs we feel like providing it; we think about how to get ridiculous amounts of data from A to B via spectacular mathematical kernel C and wedged elegantly and impossibly into web browser D, and we get to specify it properly.

If all I ever thought about or did was work, I'd say that it's a great time to be alive. I'm still looking forward to sitting on the beach this summer, tho. #