>>>>> "Michel" == Michel Schinz <Michel.Schinz@epfl.ch> writes:
Michel> Ok. Would by chance the string be encoded using some standard
Michel> encoding, like base64? If this is the case, wouldn't it be nice to
Michel> provide only "md5-digest->number" and then separate base64 (or
Michel> whatever) encoding functions? (I would certainly see the use of fast
Michel> base64 encoding/decoding in scsh).
Sure, base64 would be a win, but it looks more suitable for the sunet
package. Having thought about this once more I agree to drop
md5-digest->string.
[...]
Michel> Now that I think about it, wouldn't it be nice to provide a general
Michel> "digest" interface, which could be used to obtain digests with various
Michel> algorithms, be it MD5, FNV or others? For example, your
Michel> "md5-digest-for-file" function would be transformed to something like:
Michel> (digest-for-file algorithm fname/port/fd)
Michel> where "algorithm" would be a symbol among "MD5", "FNV", etc. (or some
Michel> object returned by a function to create an encoder).
Let's return to this when we have more then one algorithm.
Michel> For low-level functions this might not be a good idea since they might
Michel> depend heavily on their algorithm, but I think that for high-level
Michel> functions it makes sense.
Maybe some sort of init-update-digest sequence?
--
Martin
|