From Kristopher Brown to ~jonsterling/forester-discuss
I'm trying to upgrade my forester on a Macbook M1 Pro. When I run "nix run sourcehut:~jonsterling/ocaml-forester" (or develop a flake.nix file), I get the following output: trace: warning: `overrideScope'` (from `lib.makeScope`) has been renamed to `overrideScope`. error: builder for '/nix/store/3xh5bisv9bkqvg5d9p252p5s1k7nq548-cf-0.5.0.drv' failed with exit code 1; last 10 log lines: > 29 | (enabled_if > 30 | (= %{system} macosx)) > 31 | (deps > 32 | detect.c
From Kristopher Brown to ~jonsterling/forester-discuss
I like Kento's ideas a lot! Sorry to digress from implementation talk, but a concern I felt is that the forest structure is being treated as a kind of binary relation on trees: pairs (A,B) where tree A has an internal link (or transclusion) to tree B. However, I think it would be helpful to think of the forest structure as a collection of binary relations on the set of trees. Forester already has special syntax to declare other binary relations, `\author`, and `\parent`. I believe there could be a unified syntax for arbitrary relations, parameterized by a string (or a tree itself, ideally!). For example: `\relation{parent}{...}`, `\relation{author}{...}`, `\relation{tree-xxxx}{...}`, where `tree-xxxx` is a tree that describes some relation, e.g. one particular notion of "abstraction". In queries (within a tree, A), we could introduce syntax:
From Kristopher Brown to ~jonsterling/forester-discuss
Often I'd like to have a symbol in a math expression to have a hyperlink, which seems like a supported feature of KaTeX (https://katex.org/docs/supported.html), which requires passing an option "trust: true" to KaTeX. My real goal, though, is to have capacity for internal links, something like "#{a \forest{\otimes}{math-0012} b}". The present alternative seems to be "#{a \otimes b}, where #{\otimes} refers to a [monoidal product](math-0012)"... which doesn't scale with large expressions. A related issue is the KaTeX rendering of formulas as the text of links (rather than links within formulas). "[#{otimes}](math-0012)" renders as the otimes symbol. Hovering over it with the mouse makes it
From Kristopher Brown to ~jonsterling/forester-discuss
It occurs to me that the same functionality could be achieved with a new frontmatter directive "\relation{Examples}{xxx-NNNN}" rather than tied to a particular link. In this case, rather than partitioning the Backlinks/Related lists, forester just renders some additional lists of trees at the bottom, such as "Examples". On Tue, Apr 9, 2024 at 2:16 PM Kento Okura <kentookura@gmail.com> wrote: > > Since the link taxon would be a property of a specific link, it is > conceivable that two trees might be connected by multiple links with > different taxa. >
From Kristopher Brown to ~jonsterling/forester-discuss
Hi, I have encountered a number of cases where it would be useful to be able to characterize what _kind_ of relation a pair of trees has. One practical effect of this is, rather than Backlinks and Related sections being unstructured lists of trees, these lists could be partitioned by 'link taxa'. A general concept might have lots of trees referring to it, and an important subset of these to easily identify are "Examples" (Backlinks would then group things which are examples of the concept, whereas Related would group what the concept is an example of). If I have trees representing perspectives / opinions, some links are notably "Supporting" or "Opposing" the argument. A tree for a particular paper or book might have, among its backlinks, trees which are "Reviews" of it, as distinct from trees which merely reference it.
From Kristopher Brown to ~jonsterling/forester-discuss
Hi Jon, linking to subsections is a nice-to-have but not necessary from my perspective. Thanks for looking into this! Best, Kris On Mon, Feb 5, 2024 at 5:58 AM Jon Sterling <jon@jonmsterling.com> wrote: > > Hi Kris, > > As it happens, I believe that the simplest way to implement this > in fact does not have the limitation I described — so you would > be able to assign an address to an inline tree and refer to it, > and even transclude(!) it from elsewhere.
From Kristopher Brown to ~jonsterling/forester-discuss
Hi, I see that one can get a TOC subsection within a given note via a transclusion or a query. It would also be nice to be able to simply put text within a \scope (or \block) and have it render, e.g.: \scope{ \put\scope/title{A TOC section} \p{Here is some subsection text.} \scope{ \put\scope/title{A TOC subsection} \p{Perhaps even nestable?} } } I still want to respect the philosophy of evergreen notes (i.e.
From Kristopher Brown to ~jonsterling/forester-discuss
On Sun, Feb 4, 2024 at 2:05 AM Jon Sterling <jon@jonmsterling.com> wrote: > Please let me know if this is a very high priority for you. Hi Jon, it's a high priority for me insofar as it blocks porting a large amount of pre-existing notes (written in a markdown dialect) into forester. (There is just one other blocker, too, which I'll create a thread for.) Thank you for creating the issue! Best, Kris
From Kristopher Brown to ~jonsterling/forester-discuss
Hello, I'm curious if it's possible to support sidenotes and/or footnotes (possibly with pop-up when hovering over), since it can be a bit too heavy duty to make a new tree for a parenthetical remark or phrase. Best regards, Kris