Between problems 4a and 5a, I'm starting to see trends emerge from my data. Frankly, it's rather exciting seeing the "to date %" numbers fairly consistent between projects for many items; it means to me that my process is getting stable.

Unfortunately, I didn't keep good data for project 1a, and the data from projects 2a and 3a were so out of line with 4a and 5a that I still don't feel that I have enough good data to provide me with good predictions yet. But I'm sure it'll come. I'm already somewhat able to guess how long projects will take me, and arrange to get them done at a reasonable hour rather than having no idea when I'll finish. This, to me as a developer, is very exciting (and, to me as a husband, rather nice as my schedule stabilizes somewhat).

Still, I wish that the PSP came with tools and that the projects were slightly different, if only to keep the process relatively stable throughout; for example, my "plan" phases have grown very dramatically since having to hand-calculate the linear regression parameters and intervals, etc. That's fine (the purpose of learning the PSP is to expand and change your process), but it's a tad frustrating for now. I'll be very pleased after finishing program 6a, which should automate much of the estimation process.