It would be great, but is far from trivial. Sometimes a page has multiple suttas, or a sutta starts in the middle of a sutta. Also, when it’s a long sutta, say in the DN, a page reference is more specific than a plain ID, but it would be hard to identify the exact paragraph. And so on. For tens of thousands of cases.
Rather than converting such references, this is why I suggested at the meeting yesterday that we build a comprehensive model to map these references. The conversion mapping remains on the programming side, and we don’t try to correct the text. That way we can direct people to the appropriate place, without having to commit to actually determining the exact reference. If we send people to a text that is a little before or after the right one, oh well, they can adjust.
What we can do is to make the references machine-readable, which means two things:
- Mark them up with
<ref>tags or similar. - Ensure each reference is complete and consistent.
I did this as best I could during the several months work it took to prepare the DPPN.