Ben Collins-Sussman <sussman@collab.net> writes:
> I should say, that this feature *is* taking longer than I expected,
> though it's nothing to lose sleep over.
>
> I think my estimation of effort was a bit off because I only predicted
> the breadth of this change, not the depth. In other words, I knew
> that this feature would touch essentially every single library and
> codepath we have. But I assumed that because we had nice discussions
> and specs laid out before we started coding, that implementing the
> specs would be trivial.
>
> What seems to be happening is that we implement a bunch of code, then discover that the API needs to be trimmed/rejiggered in some way --
> which means going back over ground we thought was already finished.
> We seem to be making multiple "passes" of refinement. That's the
> "depth" I never predicted. It certainly seems worth it, though. When
> we finally finish this feature, it should be pretty well done, having
> gone through N iterations.
Well, we've got plenty of non-locking issues to work on, so that's
okay. I'll update the status page.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Jan 12 18:36:17 2005