On Tue, May 08, 2001 at 12:12:18AM -0500, Brandon wrote:
> 
> > > The benefits are: 1) the first inserter can't determine in a totalitarian
> > > way what the metadata of a file is. 2) The network will not store multiple
> > > copies of the same file with differing metadata (because metadata will be
> > > in a different file).
> >
> > Problem #1 won't exist in 0.4 anyway.
> 
> Stop it already with the lame 0.4 allusions. If you're going to say
> something that like, you should but a "because..." clause in it.

Sorry, I was lame enough to assume that the fact that we discussed the
reason a few weeks ago on this list was good enough.

The problem in 0.3 is that the metadata-length is stored in the storables,
yet the CHK routing key is independent of the storables.  Since in 0.4
the CHK routing key _is_ the hash of the storables, problem solved.


and in anticipation of the obvious questions:

* the storables now contain a field called Initial-digest which
  contains the hash of the first part, i.e. what we currently use
  as the routing key
* the metadata length is no longer present unencrypted in the
  storables, instead it is now a subfield of the more general,
  encrypted, Document-header storable

-- 

# tavin cole
#
# "The process of scientific discovery is, in effect,
#  a continual flight from wonder."
#                                   - Albert Einstein


_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to