On Sat, Aug 19, 2000 at 03:12:33PM -0500, Brandon wrote:
> 
> > Now, at this point, the client *has* to decrypt the metadata, parse it as
> > a FNP style message, and find the content-type to find out if its a
> > special freenet metadata section.  
> 
> The client has to decrypt the metadata and find the content-type
> anyway. That's what clients do, they fetch the file, decrypt it, find the
> content-type, and then deal with the file appropriately based on what type
> of file it is. If the metadata is in an unreadable format, it won't be
> able to determine the content type and so won't be able to do anything
> with the file except for display the undecipherable metadata to the user
> and save the data to disk. So if the file is a redirect formatted in XML
> then the redirect won't work. But in your system it won't work either. You
> just specify that redirects, etc. have to be FNP.

The only problem with encrypted metadata is that it will have to be
changed to unencrypted metadata once searching comes along.  One
solution would be to have two sections of metadata, one public section
which is unencrypted, and one private section which is encrypted.  You
would probably put potentially incriminating stuff in the private
metadata.

> I will also specify that if you have a special file, then it must be
> formatted in FNP in order for FNP clients to read it. But it can also be
> formated in XML in order for XML clients to read it.

Yeah, but I don't see why we should really bother to support XML.  We
should standardize on one or the other or otherwise clients will end
up having to support both.

> >  Either way is a tremendous waste of time, since 
> > 
> > ( Metadata-length - Data-length == 0 ) 
> > 
> > takes a nanosecond on a processor.  
> 
> I think a better solution would be explicitly state what kind of metadata
> was in the metadadata field.
> 
> Metadata-format==0 (0 for FNP, 1 for XML, etc..) is faster.
> I'd prefer something more explicit such as Metadata-format=="FNP" which
> shouldn't take much more time and would be much more explicit.

Yeah, using a short string to indicate the metadata type would
probably be easier in the long run than using integers, even if it
incurrs a small performance cost.  However, as I said above, is
different metadata types really necessary?

-- 
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: 2027 bytes
Desc: not available
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20000819/35d44212/attachment.pgp>

Reply via email to