Cart
[Home] [By Thread] [By Date] [Recent Entries]
(My addendum: sorry, I sent off that last one before it was done.) On Fri, September 22, 2006 10:17 pm, Ben Trafford wrote: > I'd rather not use a transformation when I should just be > able use something like: Either way, the point is that we store the semantics of the content and the link relationship(s) in a device-independent format, and then we use the stylesheet language of our choice to convert it into something appropriate to the medium where we want to output it. I chose XSLT 1.0 to do that (and only 3 of my 10 lines were really necessary to address your example; I just wanted to make it a complete stylesheet so that I could test it) and you chose CSS. If syntax like > xlink-href: ||location; /* where the || indicates an > attribute inspection is more intuitive to you, and works for you, then you'll get your implementation done in a language you're happy with, and that's what counts. If I prefer XSLT, and others prefer DOM or SAX or the latest Java-oriented XML API, and it works for them, fine. If we can all convert the same XML into our chosen output medium with our favorite stylesheet language or API, then that XML was well-designed, and I (and, I believe, Michael Kay) believe that this should apply to linking relationships as well as the roles of block and inline elements of content. >The other problem with this approach is that it still can >only result in unidirectional, single-ended HTML-style links. Your original example had no indication that it was supposed to be anything but a unidirectional, single-ended link. Some feel that all such link descriptions should by default be considered one end of a bidirectional link, but if we are to treat this link as one of those, this has to be indicated either in the markup or in some documentation about the default behavior of the system being described. Bob
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
