Missing Old-style figures in DOCX compile

I had precisely the same thought. I intend to continue, but so far nothing has come to light.

It’s a concordance, so there are roughly fifteen thousand documents, some a single line, some hundreds of lines; the total number of pages once compiled is over two thousand. So even going letter by letter is molassaic, so to speak. I’m actually more surprised by the whole massive thing coming together with only so minor an issue.

‘molassaic’ – Duly cubbied away for future use. :wink:


This is the amazing thing about Scrivener. With fairly minimal to no changes to a document, you can transmogrify it in a [potentially bewildering] amount of ways! The key to almost all flexible compilation strategies is to use Styles for chunks of content that may carry semantic meaning (i.e. think of italic as emphasis, figures a figure blocks etc.). This gives you the flexibility to then assign that chunk of stuff a meaning that the workflow can handle.

For DOCX I personally feel Pandoc offers the “best” output. Documents are semantically clean and well marked-up, and can easily be tweaked to use specific typographical features like old-style numbers. You can see this in the sample Scrivenener compiler output here: scrivomatic/sample-output/workflow.docx at master · iandol/scrivomatic · GitHub

The font is the Word default, Calibri, which in a new document does not have old-style numbers enabled, but in this direct output from Scrivener you can see numbers are old-style…

Yes, thank you, the Scrivomatic process was something I ran across. I will require more familiarity with its various moving parts before attempting its implementation. I find all the TeX-related stuff fascinating and promising, if daunting.