Gamall Wednesday Ida

Re (email): External academic review and publication of Wikipedia pages

@Evolution and evolvability:

Dear Thomas,

Thank you for contacting me; the wikijsci project, which I didn't know about, seems quite interesting indeed.

Unfortunately, as I understand it after a cursory look, you can only submit pages edited online in the Wikimedia format, as opposed to the LaTeX sources submitted to regular journals.

If that is so, then it's a non-starter for me, as it means that any serious work invested into the journal suffers from the same fundamental problem as work invested into the wiki: it is non-interoperable with my existing workflows and tools, and can neither benefit from nor add to my lecture notes, slides, drafts, papers,... or those of my colleagues. At least not without going through the unadulterated tedium of rewriting everything in a different syntax (including the LaTeX maths, because of packages and macros...).

The scope of the journal and its stance wrt original research are also unclear to me. [1] states that "The journal publishes both review articles and original research in various formats", but the link on "formats" redirects towards the wikijournal of Medicine, which is not pertinent to my field (th.compsci).

That said, again, I find the concept interesting, --- perhaps writing a small article for it could be a good objective for a research internship, for instance --- and I thank you for bringing it to my attention.

Cheers, — Gamall Wednesday Ida (t · c) 12:36, 5 February 2023 (UTC)

Hi Gamall - Glad it was of interest.

The LaTeX aspect is tricky. We have accepted submissions in that format previously. This one was originally submitted in Latex and stayed so during peer review, but was basically copy-pasted into mediawiki at the very end of the process. You can see from the PDF that it's in a somewhat different format to the other papers. However I appreciate it's far from ideal process for many fields. One of the long-term goals is to have a working parser to automatically convert latex to mediawiki, but that's apparently quite tricky.

In terms of scope, I the formats link sent you over to the sister journal, WiikiJMed, by accident (now fixed). It should have linked to WikiJSci's formats. The journal generally aims to be as flexible as possible in publication formats, so long as it would make sense to a potential peer reviewer, so that list isn't exhaustive (e.g. this paper is a teaching method).

Anyway, have a think on it and you're welcome to send others in our direction if you you reckon they might find it useful. - T.Shafee(Evo&Evo)^talk 23:13, 5 February 2023 (UTC)

@Evolution and evolvability:

Dear Thomas,

working parser to automatically convert latex to mediawiki, but that's apparently quite tricky

It's quite tricky indeed as, in full generality, TeX parsing is Turing-complete. That said, ignoring the TeX underneath, LaTeX itself is more amenable, and there are good LaTeX compilers targeting HTML, such as HEVEA. I don't know whether or how they handle things like TikZ figures, though — I only use HeVeA on rather straightforward documents. Presumably they could be rasterised (or even rendered as SVG).

Interestingly, HEVEA seems to only work at the lexical level, no parse tree involved, exploiting the similarity between LaTeX and HTML structures. MediaWiki format does not seem too dissimilar.

Adapting HEVEA to generate MediaWiki syntax would be a bit time-consuming but far from impossible, I think, given some funding and good-will. I assume I'm not the first to think of this. If you have a link to the relevant Talk pages where plausible solutions were discussed, I'd be interested in taking a look sometime.

by accident (now fixed)

Nope. 'j' != 'J', so the link was broken ;-) Fixed it.

Cheers, — Gamall Wednesday Ida (t · c) 11:14, 6 February 2023 (UTC)

addendum: Pandoc is also capable of LaTeX-to-WikiMedia, but I'm not sure how good it is for maths. Generally speaking, the problem is that translating basic LaTeX is next to useless: many depend on a large ecosystem of packages and personal macros — I know I do — which must be supported to a large extent for conversion to be practical. I know HEVEA supports macros to a decent extent, but I've no personal experience with Pandoc. — Gamall Wednesday Ida (t · c) 16:07, 6 February 2023 (UTC)

Good points with TeX vs LaTeX. We've done some tests with pandoc and it's one of those things where a parser can probably get the basics right 90% of the time but the remining 10% can be pretty frustrating. We've had similar discussions around MS word, where converting the references is the main problem. Thank you for fixing that capitalisation, how embarrassing! T.Shafee(Evo&Evo)^talk 00:04, 7 February 2023 (UTC)