NOTATION/MIME (was Re: Recent XML WG decisions)

[Rick Jelliffe]

This is off topic for XML-DEV. Apologies.


> From: Peter Murray-Rust <Peter at ursus.demon.co.uk>
 
> In message <199709141028.UAA14599 at jawa.chilli.net.au> "Rick Jelliffe" writes:
> [...]
> > The standard way to stick a MIME type into a system identifier is
> > given as part of HyTime '97.

Sorry, maybe I should have capitalized "standard" to be clearer.  XML is certainly
neither standard (common) nor Standard (adopted by a reputable open not-for-profit 
body whose job is to set standards without undue proprietary influence) at 
the moment. 


> Being picky, this is not valid XML since prod [74] requires a SystemLiteral
> as well as the PubidLiteral.

Yep.  And do the < and > have to be entity references too in XML?


> > <!NOTATION gif SYSTEM "<mimetype>Content-Type=image/gif">
> 
> This is fine for my purposes, but I'm not clear how it fits with the XML spec.

Yep, XML does not support "formal" system identifiers as I understand it. 
I think it is a shame, since there are things that are not URLs that would 
be nice as identifiers, even in web systems.  But support for FSIs can be 
retrofitted at some later stage to XML.  I hope there is no chance 
of them being added to XML 1.0.  But I hope people keep FSIs in mind as 
a good way to ramp up the power of URIs and other identifiers in the
near future, in particular for selecting particular system identifier 
notations (schemas).

For example, assuming hrefs could be FSIs, you could have

<a  href="&lt;xml publickey='ASDASKDKJHDFKSJH#(@#$HAHAJSDLKASHD'&gt;x.txt" />

in which data about the transfer and unpacking of the resource (e.g. here a
public key for encryption) is also marked up as a part of the system identifier. 



Rick Jelliffe

xml-dev: A list for W3C XML Developers
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To unsubscribe, send to majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa at ic.ac.uk)