On Mon, 2004-08-30 at 16:42, David Steuber wrote:
> On Aug 30, 2004, at 6:14 PM, William Harold Newman wrote:
> > Oh, and by the way, thinking ahead to when sf lets me release: Does
> > anyone have any clever ideas about crypto signatures or other
> > replacements for the old MD5 signatures I used to post? Now that MD5
> > collisions have been reported, MD5 signatures aren't all that
> > reassuring any more...
> I hadn't heard about this. Are valid files causing MD5 collisions?
This came up in a recent crypto seminar. From what I understand, they
basically just said that it was more theoretically possible than
previously thought or random chance would have you believe. In other
words, if you're designing a new protocol, you should probably look at
something other than MD5. If you're running a protocol that uses MD5,
don't panic yet. You're still pretty safe, but less safe than we
> How about using SHA? Isn't that 168 bit hash rather than 128?
As far as I know, SHA-1 is still good.
Dave Roberts <ldave@...>