LWN: Comments on "What's new in TeX, part 2"

Hyphenation is overrated

nix — Tue, 10 Nov 2015 15:55:34 +0000

Because e-readers, phones, etc, have small screens. You don't always want to carry a huge screen around with you!

Also, the recommended (for good reason) line lengths of single lines is not many more times than the average length of a word -- so for longer words, a way of splitting them will *always* be needed for good appearance.

Hyphenation is overrated

Cyberax — Tue, 03 Nov 2015 19:17:20 +0000

I've been thinking about this article. And more and more it seems that trying to move TeX to web is a wrong idea.

End-of-line hyphenation is severely overrated, it actually slows readers down and looks fugly. And it's not even needed in modern browsers!

Scientific articles are most often formatted as two columns of text, as it is necessary for real dead-tree journals. You have to use small fonts to actually fit enough content into limited journal space and it's really hard to read long lines of small text.

But hey, we're now in the world where we can actually _rescale_ the text on-demand. Why should we follow the two-columns style?

Most beautiful equation

kingdon — Tue, 03 Nov 2015 16:34:12 +0000

Correct, it uses the (Apache-licensed) MathJax library so if javascript is blocked or malfunctions the math won't get rendered.

Most beautiful equation

johill — Tue, 03 Nov 2015 14:44:03 +0000

I think it's java-script enabled to render that way? At least it does for me on chromium.

Most beautiful equation

paulj — Tue, 03 Nov 2015 12:28:34 +0000

Interesting text, but jeebus, it's not very readable. Could they not have found a way to render the TeX markup? (and I know Tex math markup!).

Most beautiful equation

ballombe — Tue, 03 Nov 2015 10:34:11 +0000

This manifesto omits all the instances where tau is used as a variable:
Should we rewrite the nome
q(tau) = exp(2*i*pi*tau)
as
q(pi) = exp(i*pi*tau)

What's new in TeX, part 2

droundy — Fri, 30 Oct 2015 21:13:17 +0000

I was also disappointed by the lack of inclusion of latexml. True, it converts to mathml, which is a limitation, but it is under ongoing development, and I think it shows some promise.

What's new in TeX, part 2

leephillips — Fri, 30 Oct 2015 12:02:09 +0000

Thanks for mentioning this - there was no space to list all related projects. Note that HeVeA works on a subset of LaTeX, but it's very capable.

What's new in TeX, part 2

karkhaz — Fri, 30 Oct 2015 01:57:31 +0000

A nice and fairly complete LaTeX to HTML translator is HeVeA. It's mostly used for writing documentation for software, but can be used for other tasks. The manual is here:

http://hevea.inria.fr/doc/index.html

(created using hevea).

What's new in TeX, part 2

michal_h21 — Thu, 29 Oct 2015 23:28:03 +0000

It is actually possible to use tex4ht with Open Type fonts, both used in the TeX document and in the generated HTML file. It isn't supported by default, due to a bug in tex4ht DVI processor, but some tips are given in [1]. In the HTML, you can either include local fonts [2], or use some service such as Google Fonts [3]

[1] http://michal-h21.github.io/samples/helpers4ht/fontspec.html
[2] http://tex.stackexchange.com/a/166061/2891
[3] http://tex.stackexchange.com/a/247479/2891

What's new in TeX, part 2

oever — Thu, 29 Oct 2015 17:44:13 +0000

A small correction. .odt is the file extension for OpenDocument Format which is supported by more software than just OpenOffice.

Commenting on my experience

gwolf — Thu, 29 Oct 2015 17:08:20 +0000

Thanks for a nice read, Lee. I will surely take a look at some of the tools you mention.

I have published two books using LaTeX (both available online, although both in Spanish), and both are in some way related to this article — The first, non-technical in nature, is «Construcción Colaborativa del Conocimiento» (http://seminario.edusol.info/), an overview of the permissive-licensing creative landscape from eleven authors from different disciplines, while the second is technical (although not mathematically heavy), «Fundamentos de sistemas operativos» (http://sistop.org/).

As for the first, the authors submitted their content via a Web platform (fundamentally based on Drupal). The editting work was basically converting that to LaTeX, for which I used gnuhtml2latex (which I maintain in Debian; discontinued upstream for good reasons). The experience was... Nice, it made me learn a lot... But I would not go down that path again if possible. Fixing all the markup missed by gnuhtml2latex was painful, and once I started the editorial process, explaining to all of the authors the text was effectively frozen was a topic by itself :)

The second experience has been very positive, although there are some things I would like to polish. I am using a stack controlled by Emacs org-mode, which allows me to use a very light markup and export very good quality LaTeX code. The sources for the book are very easy to follow with no knowledge of its syntax (i.e.. https://raw.githubusercontent.com/gwolf/sistop/master/not... yields chapter 1 "in the raw", and can be easily converted into LaTeX or into HTML (as an example, https://github.com/gwolf/sistop/blob/master/notas/01_punt... shows that same chapter converted by GitHub).

To finish my work with this second book, I only need a bit of time for polishing the result. I have most of the book exportable to HTML chapter by chapter (which is enough, say, to craft an EPUB), but I need to incorporate the needed changes for it to incorporate my bibliographic references done with Biblatex; writing a post-processor with this functionality cannot be too hard, but I have to get some free time to do it :)

What's new in TeX, part 2

leephillips — Thu, 29 Oct 2015 13:57:31 +0000

Original Markdown is useless for the kinds of documents we're talking about here, with footnotes, citations, equations, etc. That's why I mention it in the context of Pandoc's enhanced Markdown, which is useful for these purposes.

What's new in TeX, part 2

dskoll — Thu, 29 Oct 2015 13:57:16 +0000

We used to use LaTeX2HTML, but I find tex4ht produces far superior output.

The HTML produced by tex4ht is not quite as readable as that produced by LaTeX2HTML, but it's still decent enough that it can be post-processed reasonably easily.

What's new in TeX, part 2

jnareb — Thu, 29 Oct 2015 13:47:24 +0000

One of simpler way of generating HTML out of *LaTeX* sources (no plain TeX) is LaTeX2HTML. It is configurable, and produces quite readable if uninspired results.

BTW. why AsciiDoc was hyperlinked twice, but Markdown wasn't hyperlinked at all?

What's new in TeX, part 2

jezuch — Thu, 29 Oct 2015 13:43:42 +0000

> Another approach is offered by the pdf2htmlEX project. This tool translates any PDF (not just one created by TeX) into an HTML page.

Recent versions of Chromium and (I think?) Firefox can open PDFs directly. Works quite well, I think, at least for the documents that I tried.

Most beautiful equation

brugolsky — Thu, 29 Oct 2015 13:11:20 +0000

Ha, good luck to tau advocates! I was "traumatized" decades ago, at the age of 10, when I realized that my hero Euler had stained mathematics by implicitly giving credence to the ancient but wretched concept of "diameter"! The "5 constant presentation" of Euler's equation is a consolation prize for the errors of history. ;-)

Tau Manifesto

fmyhr — Thu, 29 Oct 2015 12:57:36 +0000

Thanks for that link, I thoroughly enjoyed it.

What's new in TeX, part 2

leephillips — Thu, 29 Oct 2015 12:30:52 +0000

True. I meant to say "can process...".

Most beautiful equation

leephillips — Thu, 29 Oct 2015 12:29:45 +0000

You are right, I should have written it that way. The more ordinary explanation for the formula's beauty is that it combines the five most important numbers.

What's new in TeX, part 2

leephillips — Thu, 29 Oct 2015 12:27:30 +0000

Eitan Gurari, the creator of tex4ht, died unexpectedly in June 2009, and nobody has come forward to advance this rather amazing piece of software past the DVI age. So, no fontspec nor OpenType in the document at all.

(La)TeX is awesome

dskoll — Thu, 29 Oct 2015 12:09:30 +0000

Thanks for this article. We use LaTeX and tex4ht to produce documentation for our commercial products. They do an amazing job. We have both beautiful PDF manuals and nicely-accessible HTML versions and because they're generated from exactly the same source, they're guaranteed to be semantically identical.

We also do a bit of pre- and post-processing so that from any page in our product's web interface, there's a help link that takes you to the exact section of the manual describing that page. And finally, the documentation process fits extremely well into our git/make workflow.

Sometimes the old ways really are the best ways.

Most beautiful equation

kingdon — Thu, 29 Oct 2015 11:08:47 +0000

Well, if we want to quibble about Euler's identity, I suppose I need to mention e^{i\tau} = 1, and point to http://tauday.com/tau-manifesto which has a rather lengthy discussion of this.

Most beautiful equation

ballombe — Thu, 29 Oct 2015 11:02:54 +0000

You are confusing zero the digit with zero the number. Zero the number was know by ancient Egyptians, alongside negative numbers. I do not know why people are relating invention of positional notation with the invention of the zero, since before positional notations there were no true digits, only numbers.

Source: https://en.wikipedia.org/wiki/0_%28number%29#Egypt

What's new in TeX, part 2

leipert — Thu, 29 Oct 2015 10:55:59 +0000

The article contains a mistake regarding the KaTeX library, the following is stated:

"It processes mathematical markup on the server, rather than in the browser."

KaTeX however supports in-browser rendering via JavaScript.
Server side rendering is possible, the JavaScript library does not need to be included.
In both cases inclusion of the CSS and fonts is necessary.

What's new in TeX, part 2

Seegras — Thu, 29 Oct 2015 09:54:34 +0000

> Tex4ht has, indeed, made realistic singe-source authoring workflows [PDF] based on LaTeX markup
> a reasonable approach—if the author is willing to prepare source files with multiple output
> targets in mind.

You nearly had me. But then comes this:

> There are some limitations, though, including the inability to use OpenType fonts.

I do understand that OpenType probably can't be used on the Web, but does this mean I can't use OpenType fonts in the document at all? Of course, all my documents use OpenType fonts. Ligatures, you know..

Most beautiful equation

ncm — Thu, 29 Oct 2015 09:50:39 +0000

Off-topic, maybe, but it's amusing to note that Europeans were using abacuses -- which implicitly use place notation, and zero -- for centuries before they began transcribing the results that way.

Limited MathML support

pr1268 — Thu, 29 Oct 2015 07:17:55 +0000

The paucity of Web browser support for MathML is depressing. It's an ISO/IEC standard, and yet the reasons Chrome, IE, and Opera/Safari dropped support (or never supported it fully) are lame. Sigh...

Thanks for the article, Lee!

What's new in TeX, part 2

matthias — Thu, 29 Oct 2015 07:11:09 +0000

I seldomly see orphans and widows in the web. The reason is that there seems to be no good solution for breaking text down into columns. Looking at standard 16:9 monitors, there is easily space for 2-3 columns which would make reading much easier than those overlong lines, one often sees. Once this support is there, of course we have to avoid orphans and widows.

However, I am not sure what would be the best solution for pages that do not fit entirely on screen, even with columns. Breaking them down into smaller pages would be one option. Horizontal scrolling (to see more columns) another. Vertical scrolling probably would not work that well, as obviously a single column should not be higher than the screen.

Interestingly, for small devises like phones, the existing one "endless" column approach works much better than for bigger devices.

Most beautiful equation

pr1268 — Thu, 29 Oct 2015 07:04:06 +0000

Should I assume you like this form better because it includes 0 (zero)?

Zero was such an abstract concept that the Romans didn't have a clue. (How do I write 0 in Roman numerals?)

Thanks to 9th Century India (by way of Persia and Arabia) did 0 find its way to Europe. (Source: Wikipedia [which also states the profound importance of 0 in mathematics].)

What's new in TeX, part 2

pr1268 — Thu, 29 Oct 2015 06:52:00 +0000

Hey, Knuth's creation was/is so profound it caused the Earth to tremble!

;-)

What's new in TeX, part 2

Cyberax — Thu, 29 Oct 2015 06:41:20 +0000

"What's new in TeX" sounds a just like "Exciting developments in plate tectonics!"

Most beautiful equation

brugolsky — Thu, 29 Oct 2015 04:21:04 +0000

Quibble: the "most beautiful equation" is better written as e^{i\pi} + 1 = 0, as that presentation highlights how the various operations and their unit elements delicately combine and relate elementary arithmetic, algebra, geometry, & calculus.

What's new in TeX, part 2

leephillips — Thu, 29 Oct 2015 02:21:32 +0000

You are absolutely correct. Also, Firefox supports MathML, which Chrome does not. Firefox and other browsers that support CSS hyphenation still, however, do not control for such things as consecutive hyphens, widows, orphans, and other things that make paragraphs look bad and hard to read. Thanks for pointing out my omission.

What's new in TeX, part 2

roc — Thu, 29 Oct 2015 02:11:31 +0000

It's too bad you didn't try Firefox for your "typography on the Web" example. Firefox supports auto-hyphenation (if you enable it with CSS "hyphens:auto" and set lang="..." in the HTML). It also enables ligatures at all font sizes. Chrome != the Web.