Hiya, To the 'processing-model' branch, based on the 'lb42-pureodd' branch, I've added the specs of elements needed for the processing model, modified that for elementSpec, and added the draft prose to the TD chapter. Hurrah. (Feel free to edit and improve the prose.) I don't promise I've not broken anything and haven't yet added the cross-references to the guidelines text in the TD chapter. (But that is straightforward and I'll do so soon.) What I *haven't* added so far are the tests that Sebastian had written for the processing model because a) believe it or not I've never actually added tests to the TEI build system and so now ask for help in doing so and b) I'm not sure the tests Sebastian was doing are specifically the tests we want. In this case what we want are tests that test that our schemas are allowing/disallowing the various model* and related elements in the right place and used only in ways that they are allowed. Sebastian's tests actually test the parsing of these processing models and the output they generate. Is that what we want? Remember, what we're doing here is adding the ability for documenting processing models, not adding TEI Simple, or any particular implementation of doing things with that processing model documentation. For example, in the TEI-Simple repository there are two types of test in the build.xml ant file: The 'tests' target: https://github.com/TEIC/TEI-Simple/blob/master/build.xml#L197 validates the TEI Simple ODD against the processing-model schema, and validates various instances of TEI Simple against the TEI Simple schema. That doesn't actually test the processing model elements are used properly. The 'pm' target: https://github.com/TEIC/TEI-Simple/blob/master/build.xml#L259 first creates some TEI simple XSL from the processing model documentation in the TEI Simple ODD, then creates a variant of this from teipm-ms.odd (which was a test variant requested by Council demonstrating using it for Manuscripts), and then runs various test files through both those XSL in order to produce HTML. Again, this seems to be testing the processing of the documentation of processing models, rather than the elements. (e.g. we don't seem to test 'list' or 'person' in this way?). Could someone with more experience with the TEI test system either help me design some new tests or help me migrate some of these into the processing-model branch? Many thanks, -James -- Dr James Cummings, James.Cummings@it.ox.ac.uk Academic IT Services, University of Oxford