HN Explorer

Ask HN: is HTML5 ready to takeover LaTeX in academic publishing?

2 points by tunnuz ↗ HN
Being a Ph.D. student and an HTML5 enthusiast, I have been thinking about this matter for a while; I am now turning the question to you as I haven't been able to find a convincing answer yet.

HTML5 is the lingua franca of web, it is lightweight, easy to modify, presentation is separated from content (i.e. the same article could be published on conferences, journals, websites without recompiling the code, by just changing the CSS, and the appearance would fit nicely in the media), has support for math, images (most raster and vector formats) video, audio, links, tables, extensive support for typography and can be consumed on almost every device on Earth. So, apart from the obvious inertia, why hasn't HTML5 superseded LaTeX in academic publishing yet? What's the key feature that is still missing? Also, how a mass switch from LaTeX to HTML5 would affect academic publishing?

Have your say.

12 comments

[ 2.8 ms ] story [ 35.6 ms ] thread
exim
I think no. Can you as easily type e.g. math formulas in HTML5 as in LaTex, in the plain text editor?
tunnuz
I'm thinking about http://en.wikipedia.org/wiki/MathML
exim
It is good for viewing, but not for producing, I think. You can compare LateX vs MathML examples on that page. MathML xml is not something you can type as quickly and easily.
tunnuz
I somehow agree with you, LaTeX formulas are simply easy to type, while typing MathML syntax is a bit involved. However, as I somewhere pointed out, I'm thinking at HTML5 as a product, not necessarily to be authored manually. A Textile/Markdown variant could easily handle LaTeX-style formulas and translate them to MathML for HTML5 publishing.
lumberjack
In my opinion it would make more sense to simply embed LaTeX in HTML5.
Ralith
LaTeX is built for typesetting, and is very good at that. HTML, of any variant, is not.
tunnuz
I want to break the "LaTeX is the only possible alternative" point here, so your comment is very important to me. LaTeX is BUILT for typesetting, this doesn't necessarily mean that is GOOD for that. Sure, as far as we are all accustomed to its intricacies, LaTeX is very good for typesetting on PAPER. But what I am arguing is that:

  * maybe it's no more necessary to publish on paper,
  * maybe a new tool is now better for publishing even if it's not meant for it (but a lot of the new tags in HTML5 suggest that it might soon target book publishing).
So, could you please elaborate on your point? Why is LaTeX better for typesetting than HTML5?
Ralith
You make an interesting point; LaTeX is certainly less relevant if you no longer target paper. At the very least, though, I have yet to see any text renderer or layout engine that can make a wall of text quite as aesthetically pleasing. Perhaps I'm merely biased.
rprospero
I've never seen a good way of handling footnotes in HTML. Most systems pile up all the footnotes at the end of the document. However, if the footnote references an image, you're now scrolling back and forth trying to understand what's happening. With LaTeX, the footnote is placed in context.
tunnuz
Several thoughts. The first is about footnotes (I assume that you refer to \ref{}s and \cite{}s). In LaTeX you need to define references in advance, e.g. in a .bib file most likely, and then reference them. Similarly you can do in HTML5, with the difference that the references could be easily clickable and allow navigation (you can obtain the same by using certain bibliography styles in LaTeX but it doesn't come built-in). Textile (I didn't mean that HTML5 should be used natively) already does so. Also, the position of the footnotes can be handled by CSS separately from their content, e.g. you can have a CSS which overlays references at the bottom of the page while your are scrolling, or a CSS for print which pile them at the bottom of the document. It seems pretty to me. Second, imagine if academic papers were published in HTML5, how difficult would be to navigate from a paper to another by following hyperlink references? No more search through Google Scholar or publisher's websites.
rprospero
I completely agree with you that adding cross-linking to referenced papers would make research hundred of time easier. On the other hand, I think that that would be better served by adding link support to the bibtex parts of pdflatex than switching entirely to HTML5.

As for the rest, I was referring to the literal \footnote command. I'm not sure that the CSS approach is really the best, simply because you don't view webpages the way that you view printed documents. As you scroll down, you'd have footnote bars at the bottom of the screen populating and depopulating depending on which marks just entered or left the screen. That would grow nauseating very quickly. It might be better to include some sort of mouse-over method, but that still feels hacky.