On Tue, 23 Apr 1996 18:12:32 (BST) Jon Knight said:
> ...  Just one minor point struck me when reading
>through the first example in section 2.1; why does the "scheme"  part of
>otheragent (transcriber) get separated from the name by a colon whereas
>the "scheme" part of date, form and language are surrounded by brackets?
>Surely we're attaching some internal syntax into the value of the name
>attribute to the <META> element and so we should be consistent?

I don't believe that 'transcriber' is a SCHEME; it specifies the role
of the other agent being named, not the rule set governing the
identification and/or transcription of the name.  Hence the variant
improvised syntax.

>However I'm a little unclear how the last part of the proposal will work
>still (the SGML DTD for the Warwick Framework).  Am I right in thinking
>that all the contents of an instance of the WF would have to be "SGML
>friendly" in order to be included?  If not, how would a non-SGML metadata
>format (say PICS or IAFA templates or even a binary file) be embedded into
>a document conforming to the WF DTD without breaking SGML parsers?  Is

The SGML DTD for the Warwick Framework provides a method of naming
a lot of packages in a single container.  Since the packages may not
be contained physically within the SGML document entity, the DTD per
se does not address the problem I think you are concerned with,
namely packaging it all up in a single data stream for shipment over
a network.

There are a number of ways people go about packing up the entities
referred to from an SGML document; unlike the Warwick Framework DTD,
they do address the problem of shipment over the net.  There is the
SGML Document Interchange Format (SDIF) defined by ISO a few years
back, but I don't know whether anyone uses it or not.  The SGML Open
industry consortium has also been working on this problem, but I
don't know the state of play.  I hope Lee can tell us all a bit more,
or point toward the right documentation.

As far as I can tell, MIME could also be used as a packing tool for
packaging sets of entities.  (But I don't know enough about MIME to
say for sure, or to know what would still need solving.)

>there a way in SGML of including variable value boundary markers (like
>those found in MIME) so that the content of the metadata packages can be
>distinguished from the WF stuff that surrounds them?  Using external
>references might not be an option (for example if you want to drop a
>whole load of WF containers onto a laptop to work on during a flight).

SGML itself abstracts away from transmission media and does NOT
attempt to deal with this, or require a particular method of dealing
with it.  Any software that can map from a system-dependent name to
a data stream and can interact with the SGML parser can, in principle,
serve as an SGML entity manager.  (Which is not to say making the
parser and entity manager talk to each other will necessarily be
easy.)


-C. M. Sperberg-McQueen
 ACH / ACL / ALLC Text Encoding Initiative
 University of Illinois at Chicago
 [log in to unmask] / u35395@uicvm

All opinions expressed in this note (except those I have quoted with
a view to refuting them) are mine.  They are not necessarily those
of the Text Encoding Initiative, its executive committee or other
participants, its sponsors, or its funders.  Anyone who says otherwise
is wrong.


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%