I have a quick implementation question...
Where does Author's Affiliation (or, Company) go, using the HTML META tag
method?
I am more and more coming to like the idea of supporting separate files using
<LINK REL="metadata/dublinCore" href="theFile.meta">
<LINK REL="metadata/IAFA" href="theFile.iafa">
and so on. For a non-HTML document, you could also have
<LINK REL="instance" type="application/PostScript" href="file.ps">
and simply link everything with an HTML `document' containing only the
metadata.
This is partly because I think that the HTTP people are also working on ways
of shipping bundles of stuff, and probably we shouldn't be trying to solve
compound documents.
It's very rude of me to be attacking a document with my name on it, I know.
But I am coming to think that the approach I am suggesting is much easier
for WWW robots to handle, and allows binary formats to exist, and doesn't
_require_ a special unpacker -- you can leave the files loose if that's
convenient for you, and software that uses (say) IAFA templates to build
an FTP archive index (or whatever) can find them without being modified.
Having said that...
SGML can easily point to external files.
A recent modification to SGML (the `corigendum') allows you to include
foreign objects, too, although they mustn't contain the 2-character
sequence "</" in them. MIME base64 is probably OK, but I haven't checked.
(the restriction is for backards compatibility)
Sorry to include so many issues in one letter.
Lee
--
Liam Quin, SoftQuad Inc +1 416 239 4801 [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|