- From: Shaun McCance <shaunm@gnome.org>
- Date: Wed, 13 Jun 2012 19:43:04 -0400
- To: public-multilingualweb-lt@w3.org
On Mon, 2012-06-11 at 16:35 +0100, Yves Savourel wrote: > Hi Jirka, all, > > > Personally I tend to be conservative in adding features. > > So unless there is clear demand supported by use-cases > > I wouldn't do this. If decide to do this we need mechanism > > similar to AVT in XSLT, see: > > http://www.w3.org/TR/xslt#attribute-value-templates > > So far I haven't seen any use case where the variable would be in > something else than an XPath expression. > > So if nobody else can come up with a use case I agree that we can just > limit the parameters requirement to enabling the use of XPath > variables. I agree adding something like AVTs is too much without a real use case. It's only necessary to enable dynamic content (variable or not) in otherwise static attributes. And if this is even useful for some attribute, there's precedent in ITS for just having two separate attributes, i.e. locNote and locNotePointer. As for variables in general, I don't see anything in the ITS 1.0 recommendation that explicitly precludes implementations from binding variables to the evaluation context. But if you use any variables in your selectors, then you have an interoperability problem with any implementation that doesn't explicitly bind the variables you reference. If there are good use cases for variables, I think the simplest and most robust thing to do is to introduce something like the stylesheet-level xsl:param as a child of its:rules, and then explicitly any variables not mentioned by a global param. -- Shaun
Received on Wednesday, 13 June 2012 23:43:30 UTC