Subject: RE: [xsl] Is letting the browser transform XML to XHTML using XSLT a good choice? From: Didier PH Martin <martind@xxxxxxxxxxxxx> Date: Fri, 03 Mar 2006 10:08:05 -0500 |
Hello Gowri, You heard that it's better to do transform on the server side. I will give you a more balanced point of view. a) if you need to publish your apps or documents to the internet and want to be sure absolutely everybody can access the apps or content, then, yes, it's better to perform the transformation on the server side and transform into a very basic HTML with very little (or preferably no) ECMAScript. In other words target the least common denominator. b) If you target an intranet or an internet niche market where most of the browsers are upgraded to either firefox/mozilla (version 1.3 and up for firefox and version 1.6 and up for Mozilla) or IE (version 6 and up - 5 may be ok for transform but be careful on the HTML/ECMAScrip side) then you can perform the transformation on the client side. I would comment that the IE XSLT engine is less buggy than the firefox one. So if the intranet/niche market has a majority or IE it's even better. If not, be cautious to test with both. For instance a well known bug and still not resolved after more than 5 years is the limitation on xpath constructs found in firefox/mozilla. It is because, the internal XSL engine (transformMix) didn't got any improvement since Keith from Mitre gave it to the Mozilla project. Bottom line, if you target an intranet with a majority of IE browsers (v 6 and up), go for a client side transformation with confidence. In other words, be cautious firefox XSLT engine is buggy and test on both to be sure. Note about testing if you intend to create a more sophisticated AJAX application using XSLT: you'll have more success to test you AJAX application in IE since the debugger (provided by either Microsoft Office or Visual studio) will work after the transformation. Hence, produced HTML/ECMAScript apps can be debugged after an XML to XSLT transformation. It is not the case with Firefox. Firefox debugger (Venkman) doesn't use the result of a transform as source for debugging but instead the orginal document source which in the case of a transformation an XML document. Bottom line, Firefox is a lousy environment to test AJAX applications using XSLT transforms. But IE allows you to do that. Please note that I am in no ways attached to Microsoft and what I am telling you now, is the result of concrete experience but pseudo expertise or religious attachment to a certain technology. Now for XML to PDF transform and this is an opinion, not knowledge gained from experience, I think that the advice given by others about FOP is a good one. My limited experience of using FOP is also very good. But in that case, forget about client side transform unless you include it into an applet. For PDF transform I would also perform it on the server side. Up to my experience and knowledge client side transform is working well for: HTML/ECMAScript (using output HTML), ECMAScript only (using output text), SVG (using output XML) In the last 3 years I was able to create very sophisticated applications and applying model driven architecture using XML, XSLT, XHTML/ECMAScript and SVG. In the last case using the Adobe plug-in/activeX to render SVG in the IE browser. Data is encoded in XML and transformed into XHTML/ECMAScript, SVG components and rendered in the browser preferably in IE (for reasons mentioned earlier) Hope this more elaborate answer gives you enough info to base your decision. If not, feel free to email me, *** real *** experience on client side XSLT transformations is very rare. This is not the case for server side transforms where the corpus of knowledge is well established, rich and diversified. Now let's go back to work.... PS: sorry for style and syntax errors, I wrote this in a rush between two client side transformation debugging sessions ;-) Cheers Didier
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: [xsl] Is letting the browser tr, Robert Koberg | Thread | [xsl] service marks and other speci, Ferreri, Greg |
RE: [xsl] Split the content of a va, Silvia Liberto | Date | RE: [xsl] Is letting the browser tr, Didier PH Martin |
Month |