In which he shoots himself in the foot
There are days when I think that I must be one of the world's best programmers - well, no, I never actually think that. But I do have days when everything just flows, when ideas hatch fully-fledged, when balls juggle themselves, when difficult concepts stand meekly in line to be understood.
There are also days when I think that I must be one of the world's worst project managers, and that if this were a business I would fire myself. Know the joke about the man rearranging the deckchairs on the Titanic? that's me.
We have a meeting in Moneyville tomorrow morning, to sell my database to the Very Large Financial Services Corporation, and after allegedly testing the database all week I have just now discovered a serious
The simplest thing, is to disable the importing routine and tell them that it's not included in the demo. If I input today's currency rates by hand, they will probably not think about updating them for a week or two, which gives me time to find a solution. Make it so.
When this is over, I shall slap myself on the head and write it into my book of Experiences To Learn From. Repeat after me: When preparing for a specific client, ask yourself "what is different about this client?" and spend your time testing against these differences.
posted by Udge @ 21:02
7 Comments:
Just do a lot of regression testing after the fix. They say 50% of bugfixes introduce new bugs & last minute fixes like this are the worst, for that.
All programmers should think they're the best programmer in the world, when they're coding. You just have to stop believing it while you test :-)
(I basically monitor & analyze regression tests on big software projects for a living. Last-minute fixes make my blood run cold.)
That said -- sounds like you're trying to pull off a very demanding job with basically no QA support, so I'm with 3rd and Witho. Give yourself a break!
Like Dale said. (GeekySavta)
I must admit that I had to look up "regression testing". I'm an architect, I got in to programming through the back door and although I have acquired most of the good habits and right ideas (by reading, but mostly by watching and emulating other programmers) I often don't know the names of things. I looked up normalisation the other day, expecting it would be something new, and found that I've been doing it that way for years because it's so obviously the right way to handle data.
damn these tiny buttons. I wanted to say "thanks" to all for the kind words and encouragement. The meeting went very well indeed (I'll post about it).
:-) Yes, "regression testing" sounds like it ought to be some difficult, arcane statistical analysis. All it means is "make sure you didn't break anything."
That's great, that the meeting went well!
Doesn't "3rd and Witho" sound like a good location for a hip coffee bar?
:DDD It does!!!
Post a Comment
<< Home