> I am sorry for the sake of the Linux project that Linus Torvalds is > wasting his time toying with this naive idea (not that he is the first > one to have it btw, it was popular in systems research a few years ago, > but the interest seems to have faded, probably due to the problems > described above.)
> Tom: Your my-id--archive--etc tuples are right on the mark. You get a > managable, yet collision-free search space. Hashing fails on both of > these accounts. If you need context-addressing, lets use a B-tree or > something similar. Stop panicing. You are just a tiny bit confused about the tech here. It should tell you something that Linus and I, whom you would seem to think are coming from opposing perspectives, are actually in strong agreement. It's confused to think that blob addresses and revision names are alternatives for the same thing. The truth is closer to a revision name resolving to a blob address, in some contexts but not all. -t _______________________________________________ Gnu-arch-users mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnu-arch-users GNU arch home page: http://savannah.gnu.org/projects/gnu-arch/
