Francois Lachance wrote:
> why type the reference string? why not type that to which it
> refers/points? I ask because one could be marking up a string that is not
> in itself a URL
fair point, yes.
> Why use the header and not say <gi>back</gi> or <gi>front</gi> which could
> be a <gi>tei.2</gi> in a <gi>tei.Corpus</gi> ? Didn't someone intimate in
> this thread (Peter, I believe) that they wanted to separate out the
> logical from the physical designators?
whether you use the <teiHeader, or <front>, is really up to
you to decide whether this stuff is part of your text, or is
> Am I correct in believing that it is not
> necessary to modify the DTD so that the <attr>key</attr> could be typed as
> IDREF and thereby formalize the relation to the value of the
> <attr>ID</attr> (and thereby ensure that upon validation that no
> <attr>key</attr> values did not match some <attr>id</attr> value?
making the "key" attribute be IDREF _would_ require a DTD
change. its CDATA at present
> What are the advantages of the proposed approach over one built upon :
> <gi>ref</gi> with <attr>target</attr>
> when the target value is of type IDREF and points to the ID value of an
> element that houses the targetted URL?
hmm. I am getting into hot water here, and I fear Lou and
Syd will come and push my head under; but it seems to me
that <ref> is intended for only one level of indirection,
so that "HERE points to THERE". Whereas the <rs> approach
says "Go to THERE to solve the problem".
I am probably talking rubbish. I wish someone from the TEI
working group on linking etc would dive in and be
But I think we are establishing that the TEI already has
quite a good range of tools to let us avoid the dreaded
Sebastian Rahtz Information Manager
Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431