ODD and schema info

[ebeshero]

Here's some info to help guide the ODD development after our meeting at White Rabbit on 7 March 2017. (Add to this as we go!)

ODD is currently incomplete. More needs to be added to fully constrain attributes.

See this issue for details and resources: setriplette/golden_age#11

Screenshot of how to "run the wrench" transformation to generate your Relax NG XML syntax:

[ebeshero]

@ahunker @bsf15 @RJP43 I figured out what was going wrong with the Schematron in your ODD! When we go to associate the output Relax-NG schema with each of your XML files, we need to click a checkbox that activates an additional schema association for embedded Schematron rules!

I'm going to do that now on each of your XML files. The Schematron rules work, so that on your <sp who="#washington #jefferson"> or other <sp> elements with multiple speakers, it will now read your personography file!

If it's okay with you all, I'm going to do one giant Git push from here that will add the schema line you need to each of your XML files.

[ahunker]

Go ahead and push

[ebeshero]

I'll ping when I'm about to do it. I'm writing an XSLT so it adds the processing instruction (the complete schema lines) to your whole directory of XML files in Act One (and then Act Two)...

[ebeshero]

@ahunker @bsf15 All of the XML files in Acts One and Two have proper schema associations now, and the schemas are firing on the @who elements. Pull in the changes and check it out! (These files have Audrey's new TEI header info, too.)

One magnificent blunder I made last night, and corrected when I noticed this morning, was to output the updated files from Act Two into the Act One folder. Don't worry--I fixed it, and checked to make sure against the Genius listing of the musical titles in order. But it made me think that it might be wise to include in each song's TEI Header which Act it's part of.

Other thoughts for the TEI Header: Include the author, Lin-Manuel Miranda in titleStmt, and you could probably include other meta-info about the musical as well, though your emphasis should probably be more about the source of the lyrics than performance information. (Did you use the Genius editions? Did you transcribe from an album...?)

These kinds of things become part of the standard material in all the headers, so I'd recommend writing into one XSLT identity transformation file, test it on a file, and then use a batch process to run it over a whole collection. I was just doing that this weekend with repairing your schema lines--and here's how I did it:

1. Test your XSLT identity transformation against just one of your XML files and make sure it's doing what you want.
2. Run a batch process, which you need to set up carefully. I would do this one directory at a time and not over two separate directories at once (which is what I tried and had to undo yesterday). So run a batch process over Act One, and then Act Two. Here are instructions I wrote on how to run a batch process in oXygen: On this page about oXygen transformation scenarios, scroll down to the heading "Batch processing multiple XML files with XSLT". There's also a way to do it at command line, which you'll see on that page, too. If you like we can set up a batch process together in a project meeting.