August 8th, 2002



I am behind here at work. I should have stayed late last night and attempted some semblance of catch-up. Instead I went to Aurifice and to the Mercury, where I had a good time until about 1:30. However, I forgot to set my alarm and still was up at 6:30 this morning. I just ate a breakfast sandwich, although I wasn't hungry. No appetite. Little sleep. Yep, I am stressed. I think I have to bail on my brother Joe's "bachelor party" dinner tonight, although I may catch up with them later. I have a deadline of 10 pm Sunday night to get a whole truckload of work done, and with the wedding rehearsal & dinner tomorrow, and the wedding itself Saturday, those days are out for extensive work. I can probably get in a bunch of time Sunday, but that's pushing it too close to the deadline. Bleah. This sucks. Limited live journal and crack bored for me today.


Work update

So after 3.5 hours here (minus meetings & other stuff), I have gotten the one thing done that i thought would be really difficult. This leaves four tasks before my next milestone:

  • Must figure out why "Save values and move to next month" followed by "Save values and move to previous month" seems to overwrite data. Shouldn't be too difficult once I break into the debugger and see what I screwed up.
  • Must complete/unit test the "Commit these changes to the database" functionality
  • Must fix the layout
  • Must do final walk-through of entire feature as user will

4:33 pm - Figured out that the middle tier is not recording one of the first pieces of information I send it regarding this new feature. So items in red can't be completed until the guy doing the middle tier figures out why.

6:26 pm - Got layout pretty much done. Could be better, but it's an internal tool. Our employees will just have to deal with things not being exactly pixel correct



Since I couldn't finish the stuff for my other milestone because the middle tier is broken, I went ahead and skipped to the work for the last milestone for this feature. The dev lead on the team that originally owned the feature had estimated 2 days. I had bumped the estimate up to 3 days because my group (i.e., me) was unfamiliar with this part of the code base. The previous milestone (the one I am waiting on other people for) was originally estimated at 3 days. I bumped it to 5 for the unfamiliarity factor. It took me about 9 days. Some of that time was context-switching though. That's time I lose because I have to go do other things. For example, I have to drop the work and go fix a live site bug where $1,500 items are being sold for $1 (true example, and the problem was hard to diagnose but easy to fix). I spend 3 days working on that. Now I have to switch back to my real work. But it takes me a half a day to get back to where I left off, for various reasons. My schedule gets bumped 3 days. But the estimate for this feature now is a half day longer and the end time is now a half day further out.

Now, my rule is that I always assume that if the early part of the work goes longer than estimated, that later work will also go longer than estimated. So I re-estimate that this second feature will go at about the same percent over. So the estimate for 3 days is now 5 days.

However, for once the gods were looking on me with favor! When I looked at the existing code in depth, I figured out I could extend that code, rather than write a bunch of new code to handle the new feature. Feature written and partially unit tested in about an hour. Won't be able to completely unit test until after the middle tier stuff is fixed. Basically, all I have to do now is some tweaking and set up some stuff so the international team can translate the text.

Stress level down, although caffeine level is still way high (shouldn't drink caffeine to deal with stress and lack of sleep, it's really bad for me). Should be able to sleep tonight and eat tomorrow.