Replies: 1 comment 1 reply
-
This would have been a more flexible choice to make at the beginning (and people have requested both of the things you mention -- see the issue tracker). But this kind of change to the pandoc AST would be very painful to make, as it requires revision in virtually every pandoc module, plus many subsidary packages, as well as the whole ecosystem of filters &c that depend on pandoc. Given the pain it would cause, and the effort it would take, I wouldn't count on this happening any time soon. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'm looking at the differences between pandoc and djot document models.
In this message I'm focusing on (foot)notes.
In pandoc, a Note is an Inline that contains a list of Blocks and has no attribute.
It's the only Inline element that can contain Blocks.
In djot, footnotes are split into a "Footnote reference" that has a "reference label" (a string) and a "Footnote" that contains a list a blocks and has the same reference label.
Djot's model opens the way to:
multiple references to the same footnote; e.g. in ConTeXt there's
\footnote[label]{text}
for the first reference, and\note[label]
for the next references to the same note;multiple types of notes, e.g. footnotes and end notes: by convention, you could give the same suffix to footnotes -- say "fn" -- and to end notes -- say "en" --, so that specific filters/writers could discriminate between the two kinds.
Is pandoc going to cover djot's model?
Something like this (in terms of Haskell constructor, as in
Text.Pandoc.Definition
):NoteRef label
(a newInline
)NoteText label [Block]
(a newBlock
)Or like this:
Note label [Block]
(the currentNote
Inline
, with alabel
string added)NoteRef label
(a newInline
for other references to the same note)Beta Was this translation helpful? Give feedback.
All reactions