latex title is unreadable, is it possible to display latex title in readable form?

balabi

Here is a picture showing one of my collection

https://pasteboard.co/Hck9shk.png

you can see the first red rectangular shows latex representation for chemical formula. The zotero capture this reference in this way, because currently many journals like Physics Review Letters using MathJax to display math equations as well as chemical formula. latex title is good for using in latex bibliography. However, it is really annoying as a display form in a reference management software. It is really a headache if you search through these titles, and try to find the related paper to the chemical you want!

Besides latex, there is other form showing in the second red rectangular, looks like MathML? I am not sure, but it is just another kind of math rendering syntax, and zotero capture the source code! This math rendering form is used in Nature and related journal. Compared to latex, I think this form is useless, it will not showing right title in latex writing, at the same time, it also make the title unreadable when search libraries.

So my suggestion is that, the latex form captured from mathjax page should be conserved. However, in zotero window, these latex should be displayed in a readable form. It may not be possible to display this latex using mathjax rendering, but at least, display them as readable form. For example

$({\mathrm{Ba}}_{1\ensuremath{-}x}{\mathrm{K}}_{x}){\mathrm{Fe}}_{2}{\mathrm{As}}_{2}$

should be displayed as

Ba1-xKxFe2As2

I am sure zotero already has this conversion tool. For example, the first red rectangular reference is captured from this page https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.101.107006
And in the abstraction, there are also mathjax! like this
https://pasteboard.co/HckhNtR.png
But zotero capture the abstraction as readable form like this
https://pasteboard.co/HckhqqZ.png

And in terms of Nature journal, please provide a way to convert title into readable form. Or directly capture the title as readable form.

Thank you very much! Looking forward to your reply

noksagt

I am sure zotero already has this conversion tool.

Only sort of. The BibTeX translator is able to perform some very basic LaTeX↔HTML conversion and strip some LaTeX tags. The translator you linked to has a "cleanMath" function that strips math or mi tags. But things may go wonky if this is applied haphazardly & your export file won't have any math markup.

The real issue is that some pubishers include MathML and/or LaTeX in export files like RIS that don't support these markup languages. As far as I know, no reference manager converts these, so: Garbage In=Garbage out.

Here's the RIS for the example URL you provide:
TY - JOUR
PB - American Physical Society
ID - 10.1103/PhysRevLett.101.107006
TI - Superconductivity at 38 K in the Iron Arsenide $({\mathrm{Ba}}_{1\ensuremath{-}x}{\mathrm{K}}_{x}){\mathrm{Fe}}_{2}{\mathrm{As}}_{2}$
PY - 2008/09/05/
UR - https://link.aps.org/doi/10.1103/PhysRevLett.101.107006
JF - Physical Review Letters
JA - Phys. Rev. Lett.
J1 - PRL
VL - 101
IS - 10
SP - 107006
EP -
A1 - Rotter, Marianne
AU - Tegel, Marcus
AU - Johrendt, Dirk
ER -

Zotero scrapes the abstract separately and applies the cleanMath function to get rid of the mathjax junk.

Since different publishers have different math quirks, these would likely need to be applied to specific translators, rather than generically.

balabi

@noksagt Hello. Thank you so much for reply.
I think at least you must be agreed it is a bad experience to read and search through in a reference library with such long markup languages all around.

There must be a way to improve the experience, and it will not be that difficult. Because, at this moment, there are actually only two main markup: mathjax and mathml. So only need to prepare two conversion tools. And the rest of the thing is either to provide the user the freedom to link proper conversion tool to specific journal they encountered in their library(In the option setting or as a plugin) . Or setup a database for all the journals in the world(or just mainstream journals probably enough) and make the conversion seems "intelligent" to users

adamsmith

Right, clearly this is undesirable and I think some mathml (and/or mathjax) capabilities in Zotero would be great, but the problem here is really the publisher who is putting something into RIS that has no place there and I'd suspect they're not going to be consistent about how they export, so this isn't a simple problem to solve.

emilianoeheyns

If that page also has a bibtex reference, BBT ought to translate this to a reasonable simile of the math stuff on import.

noksagt

APS has decent BibTeX and BBT translates the particular example title correctly. The BibTeX on that site exports only the abbreviated (and not the full) journal title and lacks the day the article was published.

The second example was Nature & I think they lack BibTeX, so BBT isn't a sufficent work around.

adamsith is correct that there is no solution to the problem in general because publisher export formats and the extent they violate the standard of the export format varies widely.