Does XSL-FO have a provision for the following scenario:
An application uses an XSL-FO implementation to produce a
PDF or PCL document, but also needs to insert a specific
byte sequence at particular points in the rendered document.
Then another application extracts these byte sequences to
get a count of them, and delivers the rendered document with
an appearance as if the byte sequences never even existed.
Basically, it needs to insert "characters" for the sole purpose
of instrumentation, but the "characters" shouldn't take any
visible real estate on the rendered page.