I finally got console applications compiling in RB 6. I ended up sticking my fingers in over 30 files spanning four different projects just to get this one little feature done. I ended up having to modify the plugins sdk, modify our plugin converter plugin (boy, isn't that a brain-fuck? An RB plugin that converts RB plugins into things the IDE can grok), the IDE itself and of course, some more compiler tweaks. All in all, this "3-day project" ended up taking almost two weeks! Yeeps!
I guess that just goes to show that a WAG (wild-assed guess) really is nothing more than a WAG when it comes to figuring out how long something will take to accomplish. ;-)
What's the worst WAG you've ever made for an estimate? How far were you off by?
We get WAGs right now in our project because we're switching a LARGE project from the waterfall technique to an agile methodology.
With the clients changing their requirements , the transition leads to HUGE WAGS for features and changes.
I still think my initial networking WAG was my worst. I think I said it'd take about a week (5-days) to rewrite our TCPSocket for Mac Classic. Well, that ended up turning into a rewrite for Classic, OS X, Windows a Linux implementation, UDPSocket and a ServerSocket. All-in-all it took about a month. heh, oops.
Oh yeah, and that month doesn't include time spent fixing the random bugs that have cropped up over the years with the new socket code. Thankfully, it was a pretty solid initial writing and so there haven't been too many bugs with it.
College: thought it would take 4 years. Silly me...