Written in response to TAG discussion and specificallytag issue 23
When we are designing a new XML language, and we refer to something on the web by URI, should we useXLink?
Three possible answers:
The short answer in my humble opinion, is that (2) is right.
When an application uses functionality which is within the scope of Xlink, it should use xlink. To do otherwise breaks the principle that we are trying to make an interoperable web.
The third extreme case, that xlink should always be used, is not tenable, as URIs are used generally as identifiers for everything. Paul Cotton and David Orchard pointed out on a TAG call (2002/6/17) that the scope of Xlink is hypertext linking. A motivation for XLink was to give to languages for human documents a much richer form of hypertext than HTML, with features which had in fact been used in hypertext products for many years before the web.
A counter-example is the speech grammar specification, which uses a URI parameter to refer to a piece of grammar in an external file. This logical information is not intended to be browsed by people as a document. There is no need for the hypertext functionality of Xlink. There is no need to clutter the language with xlink:href syntax.
The idea that all uses of URIs are formally hypertext links does not use the termhypertext in the sense I use here, or in the sense in which hypertext link functionality is the scope of Xlink.
The XHTML specification does not use xlink, as (I understand) the working group felt that it was too clumsy to use a different namespace, and they wanted it to look like HTML, which uses href=. The group is (2002/06) looking at schema annotation ways of declaring html:href to carry the significance of an xlink. These are known as "hlink". The pros and cons of schema annotation in general as a means to add semantics or style to a langauge are currently under debate in the community.