Marked up the Lingard article, using my new strategies with FODT and XSLT. The bibliography is still missing, though.
With my new familiarity with fodt, it's been very straightforward to do a basic transformation from FODT to TEI for the body of the document, and cleanup and extra tagging with regexes is pretty efficient too. Bolduc is now done; Lingard is the next (and presumably last for vol 21).
Now there is another review article in the pipeline, so we have to move quickly with Bolduc.
I've set up a new SVN repo for all the ScanCan materials, and added svn keywords to XML, XSLT etc. This will help manage the project better going forward. I sometimes work on this project at home, and SVN is obviously preferable for multi-location editing.
Following advice from AH, I've widened the columns for the PDF parallel text rendering and drop-capped the chapter numbers; it's much more readable now. I also fixed a rendering bug in the XHTML view of review articles.
Just to test the waters, I've implemented the same approach with a table for PDF as I did in XHTML. It looks not-bad to me, so I'll leave it up to the principles to decide if it's better than serial rendering. I've asked on the XSL list about approaches to two-page rendering, and had some intriguing suggestions, but none are straightforward.
I completed the encoding of the text today -- the actual edition text is very simple. I've devised an encoding strategy for parallel texts (very simple -- just @type values for
<div1> elements), and I've written the XSLT and CSS required to render a side-by-side XHTML version that works pretty well. The results are up for proofing, and I've let the others know.
Meanwhile, the problem of PDF rendering on facing pages is daunting. I confess I can't (yet) imagine how it might be done, short of manually crafting each individual page, which is not practical. I'm going to post some questions to people who might know better.
I took a new approach to tagging this large edition article: I saved the Word doc as odt, pulled out the content.xml file, and then wrote XSLT to convert the bits I needed to TEI. The results were remarkably good, and I've now completed checking the introduction as well. The bodies of the translation and source text may need their own markup, but in many ways will be much simpler than the academic text. This approach is more effective than I thought it would be, and much less boring than manual tagging.
Meeting this morning with JT and HT to discuss publication and rendering for the future. Decisions (still to be run by the board):
- For the current volume, I will attempt to render the edition using side-by-side original and translation (in PDF, on facing pages). It's accepted that this will take a while to code.
- A preface to Vol 21 will need to be written, explaining the inclusion of the edition.
- Volume 21 will ship as "2012-2014".
- The existing proofed articles will be published now, with Vol 21 (2014), with individual pagination. Their print pagination will be added when the print volume is ready.
- In future, we will publish articles and reviews as they are ready, with each year constituting a separate Issue of the current Volume; when a volume reaches printable size, at the close of that year it will be produced.
- Looking forward, we will create new guidelines for submission of editions, in consultation with MD, NvD and others, and actively solicit such contributions.
- A standalone PDF/print rendering will be available for editions, so that they can be purchased in print, eventually forming a series.
Next is the intro.