Let the user set a "sample" size for a file, like I
would set something like 1000 bytes starting offset
into a file and 10K sample.
Now you hash that part and save the hash and use it to
detect dups (a option you can turn on) so that people
who change a file a little bit to make it seem new will
not be able to do it anymore. Or those damn jpgs that
look the same but some ahole put his web site on the
bottom.
The reason for the user setting is so that noone knows
where you might take your samples from a file, so there
is no set standard place.
Of course you can only check files that are larger than
your sample size, that's a given, so you tell users to
stick to the first part of a file.
If they stick to the first part of a file depending on
their download speed, meaning like under the first 1
meg, it doesnt matter if they get a few 1 meg dups, it
doesnt waste all that much bandwidth.
It would also be nice to have a preview of the file
while downloading, meaning if I was DLing 150MB MPG
movie, I would like to view it while DLing so I can see
if I want to cancel it or not.
