Re: Initial draft of XML-Link spec now available

I am very pleased to support the initial draft of XML-LINK and to 
incorporate it in CML as soon as it stabilises.  Because CML is
somewhat of a meta-DTD (and I think there may be other such DTDs in
XML) I'd like to urge that the Addressing (Sect 4) allows a fine 
granularity.  I use attribute values a lot in addressing so that
in 4.5 '[We may want to include ... pointing directly to the values
of attributes...]' gets my vote.

I'm also not clear whether addressing within #PCDATA (TEI's TOKEN) is
allowed.  Again this can be very important where #PCDATA is used in
a highly structured manner (e.g. to hold arrays of primitives) and
where additional Elements would be overkill.

(Personally I would also be pleased to see regexp's allowed in both
content and attribute values ... but that's yet more to implement :-)


Peter Murray-Rust, (domestic net connection)
Virtual School of Molecular Sciences, Nottingham University, UK