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: Liam R E Quin <liam@xxxxxx>
Date: Mon, 28 Jan 2013 00:24:35 -0500
On Sun, 2013-01-27 at 08:23 +0000, davep wrote:

> If we have 100 people in the world using XSLT. How many
> are restricted to XSLT 1.0 for some reason.

We don't have numbers, and you have to be careful to avoid the
bridge-over-the-chasm [1] with numbers like this: there are people not
using XSLT who would not answer a survey, but who would use XSLT 2, for
example, if they could.

The most common technical restrictions I've seen are
1. XSLT in the Web browser is XSLT 1 with some "1.1" and exslt and/or
James (xt) extensions.
2. XSLT embedded in larger systems is sometimes 2.0 (e.g. IBM has some
products that embed 2.0) and sometimes 1.0, and you can't upgrade it
yourself of course.
3. XPath embedded in systems is almost always 1.0 except in an XQuery
context. E.g. MySQL supports a small subset of XPath with massive bugs,
and which is not namespace aware (the prefix is part of the element
name); Microsoft SQL Server has a subset of XQuery; Oracle has a product
with a confusingly-documented subset of XPath 2 with extensions...
4. on a Web server (a strong use case for XSLT), hosting companies often
restrict users to programs packaged for their OS (usually CentOS or
Debian GNU/Linuxb"), and often forbid Java. This means you can use
libxslt (it's in C) but not a recent Saxon. I haven't tried to see if
mono can run the dotnet Saxon.

I run across people using XPath and/or XSLT 1.0 on a daily basis because
of these. Zorba and BaseX are both helping a little, and so is the work
in Eclipse.

Every now and then a Web programmer discovers XPath in the browser,
although the lack of class() makes it trickier than it need be, and most
these days use jQuery. Extending XPath in the browser to support class
attributes better, and maybe to support pseudo-properties like focus and
hover, would make it more popular, I expect.

On the back end, adoption _is_ increasing, especially of XPath 2 and
XQuery. But it's taking a long time. I think we would actually do better
to write more function libraries, e.g. for finding the size of an image
or for calling a Web service, than doing the core language work and
leaving libraries to others, but we have only limited resources.


[1] the story goes that there were two thriving towns, one one each side
of a steep ravine with a strong river running rapidly down the middle of
it. People approached the mayor of Town A and asked for a bridge. The
mayor said, since no-one today swims across the rapids there's no demand
for a bridge. (You can continue the story at leisure...)

Liam Quin - XML Activity Lead, W3C,
Pictures from old books:
Ankh: freenode/#xml

Current Thread