On Wed, Dec 12, 2001 at 04:37:48PM -0600, Mark J Roberts wrote:
> toad:
> > an MD5 checksum (don't ask me why they aren't using SHA-1) of each package
>
> MD5 is twice as fast as SHA-1 (25MB/s vs. 50MB/s on my 450MHz), and
> SHA-{256,512} are even slower. Disk IO is the bottleneck here, but I
> imagine they still wanted the optimization.
>
> > e) this could be used for looking up the decryption key for a CHK in a
> > datastore. This use is strongly discouraged as it is bad for freenet, but it
> > is possible.
>
> If a court said that nodes had to honor file deletion requests
> signed by the MPAA, that would be bad. Node operators looking for
> known keys in their stores is just a silly waste of time.
>
> > Note that this is spammable.
>
> ...hence useless.
>
> > b) Looking up SSK@<P>,<R>/<md5>-<length> for a known insertor. We keep an
>
> ...this is fine, but wouldn't it be nicer to just download a list
> of keys once, at least in most cases? Also I don't see why you've
Hmmm, what do you mean? In what context? I agree that with apt-get over
freenet, the most elegant solution is to change the Packages file to include
a CHK field as well as an MD5 field... is that what you mean?
> included the <length> parameter at all.
--
The road to Tycho is paved with good intentions
_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech
