> So, is my understanding correct when I say that TagTime is for tracking
> averages over an extended period of time?
Yes, this can be pinned down precisely but a rule of thumb is that
once you have something like a week's worth of full-time work,
TagTime's estimate will be within 10% of the truth. Maybe it was two
weeks. Something on the order of weeks, though to me it always feels
like it's at least in the right ballpark after a couple days or even
hours.
> If I am a developer billing by the hour, TagTime is not the right tool for
> me - correct?
I actually think this is fine. There are so many other sources of
randomness anyway, like an unlucky bug that takes an extra day to
figure out, or happening upon the right forum question where someone
saves you days of work. TagTime adds more randomness (for example, if
you work for 5 minutes there's a 90% chance that you'll bill for 0
time which is about right, but a 10% chance of billing for a full 45
minutes; if you work for 3.5 hours there's a 1% chance you won't be
able to bill for any of it) but it's perfectly unbiased and thus
entirely fair to the client you're billing. They may get a bit lucky
or a bit unlucky.