One -- admittedly ugly -- solution for evaluate would be to have XSLTC 
invoke interpretive Xalan (eg, via XPathAPI) to process that request. But 
I think I agree that this is a very low-priority item.

Re the BSF scripting hooks... Again, I agree that this isn't a top 
priority, but it may be worth thinking about. Some of the need for inline 
scripting in arbitrary languages will probably go away when XSLT 2.0 comes 
along, since it's expected to add the ability to write functions in XSLT 
itself.  But BSF is convenient for integration with specific environments, 
and for users who don't yet know Java very well... and it shouldn't impose 
any run-time costs unless it's actually invoked. I think I'd be willing to 
assume that if someone does use an extension they really needed it and 
make them manage the performance issues themselves... so the big issue 
here seems to be whether you're willing to make the BSF jarfile part of 
the XSLTC environment. The best answer might be to code the extension 
elements to support it, but set them up so they fail semi-gracefully if 
bsf.jar isn't available. (Just do a try/catch and issue an error in the 
catch?)

(Disclaimer: I'm one of the original developers of BSF, so I'm somewhat 
based.)

______________________________________
Joe Kesselman  / IBM Research

Reply via email to