| 1:55 pm on Aug 4, 2004 (gmt 0)|
We have many sites running XML/XSLT on the backend, but I don't see what benefits a client-side transformation has over generating and caching normal HTML/XHTML output on the server side except for the disk space. I would not design a site for client-side transformation except as a novelty or academic exercise.
| 4:29 pm on Aug 4, 2004 (gmt 0)|
Well, I don't think that the technology is there to accuratly render XML+XSLT on the client side. But if it was possible, here are the benifits it would provide:
1. You could use your own XSLT to render a site how you want it. While this may not seem that important, people who use ebay all day long or something like that would sure like to be able to customize the layout for themselves when working.
2. You save the time/resources of re-rendering the file. While this might not seem important if you batch render your files, but many people have sites with interdependancies that force them to re-render the entire site every time they make changes. Which means they have to wait for the daily CRON job or whatever to render. With client side rendering, just change the data, and changes are immediate and for everybody.
3. The website is available as data, which means that you could create some nifty tools... for example, imagine a search engine that instead of finding websites that contain the information you are looking for, simply returns the information, compiled from multiple site. You could have all kinds of intelligent web agent tools and things like that impossible from simply parsing HTML.
Those are just the things I could think of off the top of my head. There has to be many more good reasons.
| 10:27 pm on Aug 4, 2004 (gmt 0)|
Well if you are talking about client side, then I believe only IE supports it. (Plus it's buggy at that.) I use xml/xsl quite often, but only server side, that way complete html is sent to either the engine or user.
| 12:47 pm on Aug 5, 2004 (gmt 0)|
You are correct... there is no real way to render it on the client side with any sort of reliability. I was just discussing in the hypothetical... there are big benifits for doing it on the client side if you can get it working properly.
| 9:45 am on Aug 14, 2004 (gmt 0)|
Erm, Mozilla now supports XSL transformations on the client side.
| 11:32 am on Aug 14, 2004 (gmt 0)|
I know that if designed right, xml + xsl can work in ie, mozilla, firefox... most of the browsers that people care about.
The problem for client side xml + xsl is the fact of search engines. If you search through google, you may find a a few xml and xsl pages. The problem is that google does not seem to be picking up the links that are in the xml... they will list the index page of an xml site but no pages below it.
If someone can prove me wrong, I will be very greatful.
| 3:19 pm on Aug 14, 2004 (gmt 0)|
I'm not entirely sure about that.
Google follows the attribute HREF, I don't really think it depends on whether this is in HTML or XML, so I would presume Google infact does follow HREF attributes in XML.
However, with XML comes along extensibility, which results in users of XML no longer using the HREF attribute, but instead using custom attributes and elements, for example <link url="http://localhost/">My Site</link>.