My husband works in the field of software safety, so I get to hear tons of scary stories about poorly-designed software and what the tragic results were (or could have been). I think I remember him talking about software designed for hospitals once, and how the biggest problems tend to crop up when the software designers don’t spend enough time talking to the actual people who will truly be using the software day-to-day–i.e., the nurses.
Thus it was with great interest that I read an article in a local magazine this morning that happened to mention software designed for a hospital system. A nurse made a suggestion that the programmers incorporated into the design, and as a result errors in giving patients the wrong prescriptions plummeted 80% overnight.
Software isn’t inherently dangerous, although it can be easy to take that message away from the number of software-related problems out there. What’s dangerous is the way many people approach the design and implementation of such software. Sometimes the company they work for pushes them too hard, and there’s just no way for them to do adequate design and still meet deadlines. Sometimes people just don’t want to take the time and effort necessary. Sometimes people have too high an estimation of their own abilities and don’t want to admit they could stand to learn new or better methods.
Software is really complex stuff, with tiny, hidden bugs that you can’t predict the effects of, and sometimes it’s a little frightening how much trust we’re putting in the unnamed folks who design the software that operates everything we use from day to day. I’m willing to bet most folks have no idea just how much software goes into the operation of, say, their car, these days. It looks like a big, mechanical thing; I don’t think it occurs to most people how much software is in your average car now.
Anyway, it was so nice to see a story of how people did the right thing, and because of it the software at issue probably saved lives.
Aha… an internet search turned up a nifty article on medical bar-coding that I think includes some of the same story, although it’s much wider-ranging.
Just a quick note that I found this blog entry from jputnam linking to this entry. Since livejournal doesn’t do trackbacking I thought I’d manually add the link myself.