xproc vs xslt

[Patrik]

Hi,

after some (more or less successfull) experiments with xproc I'm wondering if it wouln't be better to implement everything in xslt and add the missing steps (for instance fop and schematron validation) as custom exention elements written in java.

My argumentation:

• calling xslt from xproc you can only use string parameters (trying to fill in for instance a boolean prameter will result in conversion error)
• The integration of xslt into oxygen "feels" much more advanced than that of xproc.
• Direct output of messgages from xslt (for debugging and logging with a link to the code that caused the output). (Running an xslt from xproc the xslt messages are not visible at all.)
• I will have to add some custom extension written in java anyway. (So learning how to do it will be necessary in both cases.)

Does anyone have more experience with xproc and could maybe add some arguments for or against xproc?

Thanks and regards,

Patrik

[adrian]

Hello Patrik,

I see you are also member of the oXygen-user mailing list.
You may want to also ask this question there:
http://www.oxygenxml.com/mailinglists.html#oxygen-user
You will definitely get more feedback regarding XProc from the list than from the forum.

Regards,
Adrian