Slashdot

Thuds has been delayed, but not discarded. I'll update when I can. Complicated and boring work and home situations are the cause - I don't have as much free time. If anyone cared, sorry, but I will get back to it - probably in the next two weeks or so. Here's where it is:
> I've simplified/sped up process a bit, by conglomming more work into each iteration.
> I've failed to successfully develop the habit of "test first, always", but am still "testing first, mostly". Mainly, it's hard to figure out where to draw the line on what to write tests for. I've come up with a sort of fuzzy definition, that I need to work into a more formal spec. Basically, if it takes longer to write the test code than to bench check for a bug condition, then I skip the test code. Formalizing this will be: Any skipped test condition will be documented in the design spec as a manual test step. Importantly, this is to be used only to skip stuff that is too simple to test for, not to skip complicated tests cuz I just wanna get on with it. The complicated tests are where the money is - starting to write a complex test forces me to rethink the function, and usually, to break it up or simplify it.
> Thuds have a world now, but all they do is stand there in it and grow old and die. Sad.