New member beckfield started two threads on XSLT and Opera, one civil, one less so. That is fairly topical as the recent buzzword AJAX (no relation) includes XSLT and Opera does not.
There is no discussion about XSLT on the server side, transformations is what web servers do, but client-side transformations would put XSLT into another role. XSLT will have to think like a client and is it up to the task?
This has not been an issue due to the dearth of XSLT-enhanced web pages, there is just a handful sites on the web though Google Maps is one of those, but as far as I know nobody has studied this. Where does client-side XSLT really fit in with the Web UI (HTML, CSS, JS) and how well does it adapt to the user’s environment?
Jonny, what exactly do you mean by:
“… XSLT will have to think like a client and is it up to the task?”
I’m not familiar with the XSLT spec, and wonder what particular client-side problems you envisage?
Client and server scripts think differently. A server script is usually written for a specific server, while the client script will have little or no idea of the client it will be residing on.
While the role of XSLT on the server is fairly clear, it would move into a different ecosystem on the client and has to find a different niche. When should XSLT be used and when should it not, what are the fallbacks when XSLT for one reason or another is unavailable. How should it interact with the other client-side technologies.
Existing web experience should give an indication. Say you use client-side XSLT to generate a table of content. This is generated content like in CSS and there are guidelines for good use of generated content. Critical information shouldn’t be GC, but here it is OK as the fallback would be the same document without a ToC. It would be less useable, but it would still be there.
The same rules for semantic markup applies. XSLT allows the pre-transform elements to have more mnemonic names, e.g. ‘author’ or ‘price’, but such ad hoc elements have zero semantic value, even less than ‘span’ does. That can be fine as long as you take that into account. Using XSLT to transform into font tags is no better than putting them in to begin with. You really have to consider the quality of the document both before and after the transform.