On Sat, Aug 19, 2000 at 01:38:00AM -0500, Brandon wrote:
> 
> > In your proposal, the client has no way of knowing whether it has
> > freenet-special metadata, or regular metadata.  Your proposal requires
> > that the client examine every piece of metadata to find out if its
> > freenet-special before it acts on it.
> 
> That's actually exactly my point. In your proposal freenet-special and
> regular metadata are mixed together. In my proposal regular metadata comes
> first and then a trailing field and then the freenet-special metadata. So
> the two are separated. Both of our proposals have to exaxime every piece
> of metadata to find out if its freenet-special before acting. Your
> proposal requires it to look at the DataLength and mine requires it to
> look at the Content-Type field.

Why do we need to separate different types of metadata?  And if we did
separate them, why not just use dot notation to allocate different
metadata namespaces?  For example:

Standard.<name>
Freenet.<name>
User.<name>

Those would define different metadata namespaces.  The first would
represent a namespace for standard stuff which is RDF compatible.  The
second would represent a namespace for standardized Freenet-specific
stuff.  The third would be purely for user use and the content of
fields under is undefined.

In my opinion this would be much better and more flexible than
physically separating different types of metadata.  This would allow
for metadata to be mixed together but still categorized.

> > Having the zero-length data convention (which would be a very very odd
> > thing for a normal document to have) makes it quite simple for the client
> > to distinguish freenet-special metadata from joe schmoe's metadata.
> 
> I think that a convention where it explicitly states hey, this right here
> is freenet-special metadata (Content-Type=freenet/special-metadata or what
> have you) is better than a convention which assumes that a very very odd
> situation signifies something special.

I don't think that this is necessary.  Metadata namespaces take care
of this better than MIME types.

-- 
Travis Bemann
Sendmail is still screwed up on my box.
My email address is really bemann at execpc.com.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 1812 bytes
Desc: not available
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20000819/6d9ade09/attachment.pgp>

Reply via email to