Re: [xsl] Are there any free, fully-compliant XSLT/XPath 3.0 processors?

Subject: Re: [xsl] Are there any free, fully-compliant XSLT/XPath 3.0 processors?
From: davep <davep@xxxxxxxxxxxxx>
Date: Sun, 27 Jan 2013 16:02:46 +0000
On 27/01/13 13:03, Michael Kay wrote:

On 27/01/2013 08:23, davep wrote:

IMHO 3.0 has rolled 180 degrees from the one man xslt 1.0 effort and is now the
proverbial horse designed by committee.

I think there are two questions here: is it maintaining design integrity, and are the new features in touch with real user requirements?

On design integrity, I think we're doing a reasonably good job. There are constraints imposed by the XML/XPath syntax split, and of course by backwards compatibility, but I don't think we're producing something that looks as if different parts were designed by different people. As for requirements, I would fully accept that we're perhaps over-influenced by the requirements of the top end of the market; in the battle between rich functionality and simplicity, rich functionality always seems to win. As Liam has been arguing, there are probably more people who would like to see XPath smaller rather than bigger, but the people who want it bigger tend to shout louder, and to win the day.

If you have specific criticisms of the XSLT 3.0 effort, I'd love to hear them.

My prime one is complexity MIke?
<guesses> Few users will get into more than 1/3 of xpath 3.0 with any level of understanding,
primarily due to complexity, which I see as a negative aspect. For every Dimitre there
are 101 others who will stare at the spec, shrug and go back to 2.0 or even 1.0</guess>

It may be biggco driving it, I might suggest biggco as an xquery user?

It seems to have drifted a long way from DSSSL and XSLT 1.0 the transformation language.

Whether some form of subset can be of use to lesser mortals I don't know.
XT played a big part in realising the potential of XSLT. Free as in money helped.
Is there a subset that Saxon could release under the specs conformance section
that might bring that back? Possibly not.



Current Thread