> The problem is, this doesn't allow for metadata types that we don't
> specify.  Thats one of the main advantages to the metadata field.  Someone
> should be able to put anything they damn well please in the metadata field
> if they want, because there might be instances where our system isn't
> totally what they need (say they want to integrate with a software package
> that uses XML metadata).

Sure you can put anything you want as the file. But I'm proposing a
client-side standard. Which only works if all the clients use the same
format. So the XML-using clients won't interoperate with the FNP clients
(except for clients that support both, of course).

But both your proposal and my proposal are FNP-based metadata
proposals. The only difference is that I want to separate out some of the
fields into a separate data section. The only disadvantage that I can see
to do this is that you can't have a single file which is both a redirect
and a multipart or a redirect and an index, etc.. Which I think is fine.

I'd like to modify my proposal somewhat.

Instead of:
Multipart
Count=2
Data
freenet:CHK at asdf
freenet:CHK at poiu

I'd prefer the format to be more like:
Content-type=text/freenet.multipart
Count=2
Data
freenet:CHK at asdf
freenet:CHK at poiu

That way the code which handles the different Freenet document types can
be in the same place as the code which handles the different general file
types. This allows modular handling of new Freenet document types and
integration of Freenet documents with non-Freenet, mime-based things. So,
for instance, you could store the multipart document on disk but not
download it immedately. Then, if you opened it in a mime-aware environment
(BeOS, web browser) then it would automatically launch your Freenet client
to download the multipart document.



_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev

Reply via email to