On Sun, Jul 20, 2008 at 10:03:58AM -0700, Lan Barnes wrote:
No, Tracy, they really did screw the pooch. You can spend years making mud pies, but they're still mud pies[0]. The fundamental architecture is IMO deeply flawed. - no change sets
What's fascinating to me is that SVN seems to have made the same mistake that P4 did. Both have everything in place to treat things as changesets. It's just obvious that the developers don't see _why_ history should be viewed that way. With some discipline, both allow a user to view something resembling a changeset. I think that P4 is much worse in this regards, at least in practice. It is pretty much normal practice (even expected) that a given functioning "version" consists of diverse versions of files, with an unversioned rearranging of the directory tree to boot (the client view). With some discipline, it doesn't have to be used this way, but the lack of branching means that the random versioning is usually much easier, even to get right. I think the curse of SVN was that they tried to make a better CVS. They tried to make the limited functionality of CVS work a little better. They didn't seem to be able to look at other solutions, or even what else has been developed in the past 20 years. It's actually worse than that. Their discussions seem to indicate that they are aware of what other things there are, but don't see value in them. Willful disregard of knowledge counts as stupidity in my book. David -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
