Arxive Gets HTML Running | Hacker News Reader

Arxive Gets HTML Running

info.arxiv.org

59 pointsschafsam19 comments

Why editorialize the title? Just to add the typo?

> HTML as an accessible format for papers - arXiv

It is almost philosophical that cutting-edge research is published in a manner that poses an accessibility challenge because converting Tex to HTML can be difficult.

Can pandoc help?

Pretty fitting for academia

In fairness, having proper math rendering in html and web browsers is very recent.

Mathjax, mathml and stuff are less than 10 years old, whereas postscript and pdf are like 30 years old and generally even predate HTML and web browsers.

And if you have ever had to deal with some (tenured) university professor... You know how hard is to even get them consider something new.

> Mathjax, mathml and stuff are less than 10 years old

MathML is from the 90s and MathJax is from 2009.

Getting arbitrary latex content to render well in HTML using mathjax is not trivial. There is a lot of valid tex that doesn't render correctly in mathjax without various little tweaks (adding newlines mainly) and of course this doesn't include package dependencies etc. I encounter this all the time because my personal notes in obsidian include a lot of equations etc. Mathjaxx will very frequently fail to render valid tex as html when you have something like $$\begin{align} and I have to go through and change this to
$$

\begin{align}

...to get it to work. On another site I try to post equations to, mathjaxx will stochastically mangle equations if I post them on my ipad vs working fine (except for the issue above) if I post from my laptop.

They didn't work well at all back then. And they still work worse than LaTeX, although the deficiencies have become acceptable.

Yeah but pretty much no browser implemented mathml until fairly recent. A spec is worth very little without an usable implementation.

Several folks got together and created the "RASH" HTML dialect a few years ago for academic papers:

Silvio Peroni, Francesco Osborne, Angelo Di Iorio, Andrea Giovanni Nuzzolese, Francesco Poggi, Fabio Vitali, and Enrico Motta. 2017. “Research Articles in Simplified HTML: A Web-First Format for HTML-Based Scholarly Articles.” PeerJ Computer Science 3, October. <https://doi.org/10.7717/peerj-cs.132>.

One of the authors (Vitali), did his thesis on an editor they created, RAJE, intended to be used for writing RASH papers.

This looks like it might be the GitHub repo: https://github.com/arXiv/arxiv-view-as-html

If the HTML conversion code is open source, they should make that more public! Their GitHub account link is pretty buried on their website, but they seem to be active there so they might get lots of help working out the bugs that way.

It seems like this is focused on automatic conversion of a PDF paper to HTML. I wonder if there will be the option for uploading HTML directly. It seems like generating two formats from the source will often result in a better result than "chaining" the conversions.

It doesn't convert from PDF, it converts from the LaTeX source.

Ah, I missed that. That is probably better and should work well for many papers. It would still probably be nice to have the option to provide HTML directly but a much lower priority than if it was translating from the PDF.

Apparently that's to be avoided, in the interest of keeping a single source of truth

Dupe: https://news.ycombinator.com/item?id=38724665

Seems to only work for newly submitted papers. Checked an older paper of mine and it doesn't have an HTML link.

Will it undo the two column layout?

Yes it does, e.g. compare the pdf [1] and corresponding html [2].

[1] https://arxiv.org/pdf/2312.12451.pdf

[2] https://browse.arxiv.org/html/2312.12451v1

Reply on news.ycombinator.com