Florent Georges has updated his XML serializer so that it now works with Gestalt.

I have just tried compiling and running gexslt under .NET, using Eiffel Software's 5.7 compiler.

To my surprise, it works (although it's terribly slow, due to loading a whole lot of DLLs).

So I guess that means that there is an alternative to Saxon.NET, although knowing nothing about .NET, I doubt if it will integrate with standard XML processing facilities (I am thinking about Java - if I could compile to Java byte codes, then it would work under Java, but would not integrate into JAXP - I don't know if there is anything similar to JAXP in .NET, but I presume there is).

Last weekend was the best time for English Bluebells on Penwortham golf-course, but they still looked pretty impressive this weekend.

Alas, people will selfishly insist on planting Spanish Bluebells in there gardens, and these are escaping into the woods and hybridizing. I fear the English Bluebell has only 10 years left around here.

Gobo has long included a (tailored, to make it XSLT-compliant) version of the 1.5 reference implementation of Schematron.

But thanks to David Carlisle, a version of ISO Schematron that works with any pure XSLT 2.0 is now available from him.

I have just fixed a couple of minor bugs (not yet committed to subversion), and so now Gestalt can do ISO Schematron too.

David Carlisle (on his blog) says:

The first stage is to check out the code from the Gobo subversion repository:

svn co https://gobo-eiffel.svn.sourceforge.net/svnroot/gobo-eiffel/gobo/trunk gobo

Now you must set the environment variable GOBO to point to that directory. So on my Linux system, I would add:

export GOBO=/home/colin/gobo

to my bash profile.

These are available from gestalt.sourceforge.net.

The only current differences are the URI schemes supported for input and output of XML documents.

But in general, anything that is not part of Gobo is in Gestalt alone. Gobo does not include any networking code, so resolvers that access the network are only available in Gestalt.

From the outset, Gestalt has had an extension attribute on xsl:function, gexslt:memo-function, to improve the performance of a user function by caching the result of the function for a given set of arguments. This was shamelessly copied from Saxon ("Imitation is the sincerest form of flattery").

But for a long time I have realized that this is non-compliant, as it can change the result in some circumstances.