On Tue, May 08, 2001 at 08:21:53PM -0500, Brandon wrote:
>
> > > Really, the only reason not to put the embed the metadata key is that it
> > > will make CHKs longer. I don't see this as being a serious problem, or
> > > even noticable at all.
> >
> > No, but then what is the difference between doing this and having one file?
>
> Well, as has been said many times, the difference is that files with the
> same data part will have their data part collide on insertion, meaning
> that many people can try to insert the same file, but fill out the
> metadata differently, and yet there will still only be one copy of the
> file in the network. There will then be different keys pointing to the
> same file with different metadata.
>
> This is cool since if someone inserts a file and you find the metadata to
> be erroneous, you can insert corrected metadata and distribute the new key
> without having to insert an entire new copy of the file.
>
>
I can't see the likelihood of two people working with -exactly- the same data.
That's what it would take to create colliding CHK's wouldn't it?
And if metadata<----->data references aren't bi-directional, which is what
it would take to have changeable metadata files, then metadata is not verifiable.
You can't be sure what it points too, and you just said you wanted metatdata to
be verifiable.
David Schutt
_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/devl
