Peter Murray-Rust peter at
Sun Mar 22 14:03:21 GMT 1998

I am trying to create tools for looking at hyperlink structure within XML
documents and would like comment on the role of IDREF (I have never used it
myself). The XML spec says very little about IDREF/ID. (In essence: An
IDREF must match an ID attribute somewhere in the document; no two IDs can
have identical values).

Since IDREF has potentially a strong overlap with XLL, I would like to know
if they are two ways of doing the same things or whether they have distinct
roles (so that it would be conceivable to have a document using both
constructs). (Although XLL is not finalised I am assuming that it will be
widely deployed within the nearish future :-).  Or is ID/IDREF primarily to
ensure SGML compatibility until XLL gains ground?

Are there any examples of XML documents that use IDREF? The XML-REC (in
XML) would appear to do so as there are a number of attributes ending in
*REF, but since I have not yet found the DTD this is guesswork.

The following is how I see the differences between the constructs - I'd be
grateful for additional pointers:

IDREF is formally part of XML and can be used now. Parsers have to honour
it. XLL is not yet here and requires additional software.

IDREF has no semantics in the Spec. XLL has quite a lot (including
attaching ROLEs and BEHAVIOR to the link).

IDREF requires link integrity to be checked by the parser (validity failure
is Draconian). XLL has (at present) no requirement for integrity checking.

IDREF requires DTD information for its identification (though a consistent
naming scheme might be valuable). XLL relies on hardcoded attribute names
(though a DTD can be useful in supplying them).

IDREFs are concise and simple to implement(?). XLL is normally more verbose
and if fully implemented requires a non-trivial investment.

IDREFS (sic) can have several targets. An HREF can have only one. This
might make IDREFS considerably more concise.


At present I neglect IDREFs because of the substantial additional effort in
implementing what I see as a parallel system that may not be widely used.
I'd be grateful to know if this is realistic.


Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS, Virtual Hyperglossary

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list