Open amirouche opened 5 years ago
We can parse the LOSE files (which now also contain the abstracts) to generate Markdown documents like this. Note that there are over 1000 papers, so quite a few .md files!
Maybe would be best to store only the LOSE forms and conversion scripts in the repo. The scripts can convert to whichever formats people want, and the converted versions can be hosted at convenient websites. Let's keep the current .md files until the LOSE files are stable, though.
What do you think about one LOSE file per publication, then we gather with some heuristic those publication into .md
?
In that case there would be more than 1000 LOSE files; we should at least group them into subdirectories by some criteria. Perhaps it makes sense. Each entry already has an (id ...)
property such as (id fuchs1985dreme)
. I guess the filename could be fuchs1985dreme.scm
in some subdirectory.
In practice, I've found it very convenient to work with the current organization so far. Though the page names are non-descriptive page1
..page11
, they are actually organized by topic. I guess we could rename them so the filename describes the topic.
There's a parser in tools/losebib-read.sld. If we don't parse the topic hierarchy from the file contents, we need to parse it from the directory layout.
Use the following template and create one markdown per document: