Simon observed that our hashed filenames:
0000000118-62c65c4f3a826447735efc3fd96db15456fffab0627d389cdc2ab34e8bf1452b
are rather silly due to our use of base16 for the filehash part of the name
Shorter filenames would mean faster filesystem operations.
NB: this won't work if you're using an encoding such as base64 with both lower
and upper case characters (case insensitive file systems?).
Base32 is probably the way to go.
[there was some discussion on using base64 with our own custom encoding, ie.
replacing something other than letters to avoid case problems, but that's
probably going to be problematic]
|