feature creep at the unpaid gig

[sistawendy]

As I've done four or five times by now, I spent from 9 to 11 last night talking with the director & IT guy at Lambert House. I'd worked pretty hard over the last couple of weeks on all the features they'd previously asked for, and I was pretty proud of the results that I'd been able to forcibly extract from MS Access. I'd written up a nice little deployment doc, and I was ready to bring it to the customers, i.e. the other volunteers. But neau.

See, the scenario for most of these changes is keeping people who've been banned or suspended out of the house. The DB at present isn't nearly as helpful as it could be about who's been naughty and who's been nice, even though the data is there (all over the schema). We're all agreed that the UI needs to be a bit bossy - but not inflexible - and as idiot-proof as we can make it, because we can't rely on the volunteers' being trained and able to cope with anything else. How to do that was the subject of most of last night's meeting.

Plus, the volunteers that work the "floor" have found out about this project, and some of them have some (admittedly very sensible) suggestions. Basically, they want to go paperless.

The director, Ken, asked me whether I'd rather do one big deployment or several little ones. I said the latter because a) it'll be better for my morale and that of the floor volunteers, and b) there's the bus effect: if I get hit by a bus after the first small deployment, they still get something. There's also the benefit of testing by volunteers for future releases.

Ken was also chagrined to learn that Access isn't really a client-server product, with all of the negative security implications that entails. Maybe I'll be rewriting this thing in SQL with a web front end after all.

Comments

[staxxy.livejournal.com]

I figured you were. I am suggesting you divide the time you are already spending on the Access stuff in half - and spend part of that time doing this instead. Or maybe consider doing that once you get a more reasonably functional Access one rolled out.

[sistawendy.livejournal.com]

I think after the first Access release, which is going to entail a buttload of schema additions, would be a good time to do the SQL thing. There's no sense in doing them twice.