Re: [Gnu-arch-users] Arch Versus CVS Versus Subversoin

[Florian Weimer]

* Thomas Lord:

>   Another point is that we have seen zero (`0') evidence that any
>   significant saving would arrive, under any real-world usage 
>   conditions, from the use of binary deltas.

If you use binary deltas for text files, too, it's likely that you
won't need revision libraries anymore because constructing revisions
from the archive itself would be fast enough.  (Yes, I know your
arguments why it can't be done.)

>   Who is editting these BLOB files that change in such reliably small
>   xdelta increments?

You could not even use xdelte because it tries to store both source
and target in memory.

>   I have heard suggestions like "people editting sound files" but,
>   supposing that the on-disk format of such files is a compressed one,
>   is that really true?

During development, no one hopefully uses compressed file formats.
However, you'd probably need such a huge window size to detect those
edits that existing algorithms wouldn't be able to compress even
simple editing operations efficiently.

(I believe there audio systems that already store intention lists,
which appears to be a good idea anyway.)

>   Don't get me wrong: I can think of plenty of applications for
>   non-textual diffs.   I just don't see any huge wins for generic
>   binary diffs.  Does anyone else?

Some people use version control systems for backup and file sharing
purposes, and might use it to distribute compiled binaries. 8-)