(2002-02-06) e

Dan Gillmor on ongoing lack of reliability of PCs (and now handhelds) . If my home PC goes haywire, fine. I'll deal with it. When my traveling computer goes nuts, I'm in trouble. I think the 2 pro-unreliability justifications are (a) systems we think of as reliable don't change very much, whereas computers keep changing in pretty fundamental ways, (b) "real-world" systems aren't as reliable as I expect. Elevators and escalators are two of my pet-peeves (if you live in a high-rise building you're more likely to notice this). Phones that occasionally connect you to numbers completely unlike what you dialed are rather amusing. And, can you say "jammed photocopier"?

Also, it's worth noting the exceptions: there's generally high reliability in embedded software, and in projects performed by NASA. So it seems that some ways to improve reliability are: anally specify; and, never change your mind. Oh, yeah, and spend lots of money.

I think of computers at being at the stage of the early automobiles, where they were hobbyist machines, and you had to be rich enough to have your mechanic drive around with you all the time, so he could fix the car as it broke.

Also, here are some Dave McCusker notes that relate to Systems Reliability and Dynamic Language-s.