The project involves combining a tablet computer with a touchscreen; the idea has been in my head for a long time, so i had pretty clear thoughts about what i wanted to do.
The insane last-minute schedule is not exactly the sort of thing i want to repeat –
I arrived back in Berkeley on January 2. On January 3, i bought an Acer Tablet PC on eBay. The Tablet PC was shipped on Monday, January 5 and arrived on Wednesday, January 7. While waiting for it to arrive, i figured out how to read the touchscreen in Linux, and solved a bunch of equations in order to work out how to calibrate it. On Thursday, January 8, i installed Debian GNU/Linux on the Tablet PC, then got diverted into doing some Dean stuff. On Friday, i bought a USB serial adapter so that i could connect the touchscreen to the Tablet PC. But the stock distribution of Debian didn't have the right driver, so i spent the afternoon recompiling the kernel and patching the driver code. That evening i put the two pieces together, got the calibration code working, and wrote a little demo program to test it. On Saturday, January 10, i wrote a second demo program, and started writing the paper. I worked through the night until Sunday morning, when i took a two-hour nap.
On Sunday, i continued writing the paper, took some photographs of the working system for the figures, and implemented a third program to demonstrate another interaction technique. That evening i tested the gadget by asking some people in my house to try it out. This morning, i completed a first submission of the paper by about 9 am. I walked home to get some breakfast and then ran back to campus. From about 11 to 12, i shot some footage of the working system. From about 12 to 1, i revised the paper and submitted it again. From 1 pm to about 2:30, i hacked a little more code to help explain things in the video, and shot some more footage of the working system and an illustration of how it worked. Then for the next half hour i tried to get the DV files from my iBook onto the Windows machines in the student labs that ran Premiere. But i hit a setback when i found that Premiere couldn't open QuickTime DV files. So i edited the movie in iMovie from about 3 pm to 4 pm, cursing at its incredible sluggishness all the while. I recorded the audio between 4:00 and 4:40. Told iMovie to export the movie at 4:40. Uploaded the movie file at 4:50. And uploaded a final edited version of the paper at 4:55.
It's a new record for me. When i did the Peephole thing in three weeks (roughly, one week for implementation, one week for the user study, and one week for the paper and video), i thought that was pretty crazy. That one had a real user study though.
I already knew that when it gets late and i haven't slept for a while, i have trouble concentrating. But what i discovered is that i overestimate my mental capacity in that state, because i can still feel motivated and physically able. In that state, i can do detail-oriented things, like editing figures or even programming, but creative things like writing become impossible. The words just fall out of my brain as i read them and i don't have the slightest idea what to write next. So, for maximum productivity, the best time for me to write is during the day, and the best time to program is at night. Programming is more fun, so i've often ended up doing it during the day when i feel like i have lots of energy and time to spare, and then found myself in the terrible position of trying to hit a writing deadline in the wee hours of the morning when i just can't come up with words.
Anyway... i feel like i've done my big productive thing for the year. (I've been trying to keep up with a rate of one major project per year.) Now i get to slack off. :)
Soon, i'll be in Iowa.