On Mon, Apr 02, 2001 at 06:21:50PM -0500, Brandon wrote: > > > The point is that the encryption key ought to be generated from the whole > > file, > > not part of it. A side effect of this oddity is that any CHK control > > documents > > have the same encryption key. As for metadata in CHKs, we should probably > > just > > get rid of it. The metadata should be in another key. > > That would mean two requests to get any files with metadata (most of > them). Is it worth it? How do the advantages of this method compare to the > advantages of putting control documents in the data?
Well, I think moving the metadata into a separate key is going to be the only sane way to associate metadata with a CHK and yet still have that CHK collide with other inserts of the same data but different metadata. So yes, it's worth it, because it's the only way that will work. Besides, if we do our jobs and make a reliable network here, we really oughtn't to be penny pinchers about an extra (parallel) request here and there. If our reliability is so bad we get all scared about making an extra request, then the system is useless anyway. -- # tavin cole # if code is law, then Freenet is a crowded theater _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
