Re: [tei-council] processing the processing model

In your testprocmod.odd there are 3 issues: - one more or less your fault: link behaviour expects link param, not uri - one a bit tricky, atm note behaviour expects to have label param, I wonder if we should have a default there, see https://github.com/TEIC/TEI-Simple/issues/28 - one definitely a bug on the cunning stylesheet side, figure behaviour produces invalid xslt template and so far I don't know why... On 6 February 2016 at 15:25, Lou Burnard wrote:

I've made a start on testing this, but I think it's going to be a long job.

I made a new git repo for it since it's not clear where this stuff goes (if anywhere) :

https://github.com/lb42/lb42-testprocmod

This includes a sample ODD that is meant to exercise all the behaviours currently defined, and copies of Magdalena's cunning XSL scripts for converting that to XSLT that should produce HTML or maybe LaTeX from a TEI XML file (probably using a subset of TEI Simple, but that's not what this is trying to test)

It doesn't work of course, but that's probably my fault. However, I now have to go shopping or there'll be no bread for breakfast tomorrow.

In the proc mod branch

On 06/02/16 12:56, Lou Burnard wrote:

...

By "integrate properly" I didn't mean any more than find a corner to put them so that people who download the Stylesheets library get the ability to process proc mods as well. In much the same way as the Stylesheet library now also ODD processing built in. But your third way is also possible, if the Council wants the chore of maintaining yet another package.

On 06/02/16 12:49, Magdalena Turska wrote:

...

There is a third way as well: disentangle from Simple, but not necessarily integrate with TEI Stylesheets. Frankly, I don't know what the latter would mean? I can imagine making TEI Stylesheets use procmod for generating a significant portion of itself in the future, but not really the other way around. Unless you just mean, lets carve a small corner of TEI Stylesheet repo for procmod, which is easy enough to do and shouldn't cause any headaches.

On 6 February 2016 at 12:24, Lou Burnard wrote:

Thanks, that clarifies my question below.

...

There is quite a big policy decision here though. Do we want to disentangle the processing model processing tools from simple and integrate them properly into the existing stylesheets library?

If we do, that's work we ought to do in tandem with defining the proc mod spec in the Guidelines.

If we don't then the proc mod spec ought simply to say that an experimental implementation of tjhe proc mod exists, refer to Simple, and avoid giving the impression that the proc mod is ready for prime time use by anyone else!

On 06/02/16 11:05, Magdalena Turska wrote:

Poligon is no longer relevant, as I did reorganize the repo a bit, so

...

procmod xslts ended up here https://github.com/TEIC/TEI-Simple/tree/master/processingModel

The simplest way to use xslt version is to take your ODD and run simpleoddtoxsl.xsl on in

saxon -xi -s:my.odd -o:simple.xsl -xsl:simpleoddtoxsl.xsl

Which should generate a stylesheet that one could use for transformation of tei files.

The fragments that 'knows about behaviours' are

https://github.com/TEIC/TEI-Simple/blob/master/processingModel/html_function... (for transforms into html) and latex_functions.xsl (dunno how complete this one is). So these files are where the behaviours definitions live, while actual magic that controls the whole process happens in polygon_lib.xsl

Therefore the easiest way to do the tests of type 3. would be to take current teisimple.odd, generate output xsl as shown above and store it as expected result.

M

On 28 January 2016 at 11:42, Lou Burnard

...

wrote:

Hi Magda

...

OK I will carry on producing some simple test files for (1).

On (2) I'm not sure why you mention sample Simple files here: I think that's a separate exercise. Obviously we will need to revisit them when we start work on the TEI Simple Exemplar: but you shouldn't let that stop you working your magic!

How exactly do we do number 3? You havent answered my question about the Stylesheet implications: is there somewhere a fragment of XSLT which knows about the proposed "behaviours" which we can integrate into the Stylesheet library?

And thanks for explaining the poligon!

L

On 28/01/16 10:48, Magdalena Turska wrote:

Hi,

> as for test files, I guess we should test roughly three classes of > things > 1. as you pointed out, that sample ODDs including <model> elements do > validate against TEI schemas > 2. sample TEI Simple files can be converted into output using > processing > model and this output is as expected (test files for this are in > https://github.com/TEIC/TEI-Simple/tree/master/tests, let me clear > them > up > a bit and provide you with magic commands for executing the tests) > 3. tests that sample ODDs get converted into expected XSLTs/XQueries > by > processing model > > 2 is already in place, or almost, 1 most probably straightforward for > Lou > or James and I guess even I could do it if I had an afternoon to > kill. > We > have nothing for 3 as yet, though what we could easily do is take the > ODD > we have and results of application of current processing model > implementations as our expected output. > > To satiate your curiosity: "poligon" in Polish means military > training > area, where you can shoot and run soldiers around without causing > collateral damage. The closest English word (in looks, not meaning, > though > I believe the roots of the Polish and English words are indeed the > same) > is > polygon, so that's how it got to be named like that. Sorry if it > tells > too > much about my mental processes. > > I am trying to keep afloat and not to sink under pressing work > obligations, > but if I took to it early next week would that be acceptable for you? > > Magdalena > > On 28 January 2016 at 10:22, Lou Burnard < > lou.burnard@retired.ox.ac.uk> > wrote: > > The next step in getting procmod into the Guidelines ought to be to > > produce some test files. However, as James mentioned earlier, this is >> not >> entirely straightforward. >> >> I can dream up sample ODDs which include <model> elements no >> problem, >> and >> can then validate them against the schemas generated from the >> current >> source. I can also re-use the ones in the Polygon >> directory in the original tei-simple repo, written by Magdalena. >> (why >> is >> it called "Polygon" btw?) >> >> However, should we not also be producing a test implementation or >> two >> showing the actual output generated with the aid of those models? >> That >> would require adding, presumably into the Stylesheets tree, >> some version of Magda's pilot version of a stylesheet for >> transforming >> <model>s into XSL. Or is that already there somewhere? >> >> Or are we just planning to keep quiet about the whole thing? >> >> -- >> tei-council mailing list >> tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >> >> PLEASE NOTE: postings to this list are publicly archived >> >> >> -- >> > tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived