web-native slide decks (beamer.cls in arXiv)

[dginev]

Since @brucemiller and @tkw1536 are working on beamer, here is a nice issue I can intersect you on.

There is a limited use of beamer.cls in arXiv. One such article that I randomly stumbled on today, while fixing Fatals, is 1307.1503.

Funnily enough, its author is from FAU. Indeed, that document is not an article, but rather a slide deck: PDF. There is also interplay with multimedia.sty.

Nothing urgent, but maybe a useful intermediate test -- also for the still-to-come HTML target.

[brucemiller]

Is there anything more to do here?

[dginev]

Is there anything more to do here?

Did I miss a beamer-releated development?

I think the issue is current as per its description. We need a usable HTML target for beamer-based slide decks, when converted with latexml. The arXiv article in question can be a good test case. If we have a better issue targeting that, we can close this one as a duplicate (but I don't think we do?)

[brucemiller]

Yes, there've been several beamer patches since this was originally reported; so maybe this issue has been addressed. Try it out and either close, or clarify what the remaining issues are.

[dginev]

I am attaching a 57 line starter snippet, excerpted from the article in question (arXiv:1307.1503) with two slides - a title slide (no transitions), and a content slide (4 transitions). That snippet converts error-free with today's latexml.

```tex \documentclass[10pt]{beamer} \usepackage[utf8x]{inputenc} \usetheme[pageofpages=of,% String used between the current page and the % total page count. bullet=circle,% Use circles instead of squares for bullets. titleline=true,% Show a line below the frame title. ]{Boadilla} \usetheme{Boadilla} \useoutertheme{infolines} \usecolortheme{default} \usepackage{amsmath,amssymb} \newtheorem{prop}[]{Proposition} \title[Robustification of Opimal Controls.] {On Optimized Feedback Control and the Robustification of Opimal Controls} \author[Martin Gugat]{Martin Gugat} \date[]{ {\bf CDPS 2013}: 8th Workshop on Control of Distributed Parameter Systems, July 1- 5, 2013, University of Craiova} \begin{document} \maketitle %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section[Example]{} \begin{frame}\frametitle{Example} \begin{itemize} \item<1-> Let $y_0(x)=x$, $y_1(x)=0$. \item<2-> We get the optimal control $ u_\ast(t) = \frac{1}{T} \left( 1-t\right), t\in (0,2) $. \hspace*{1cm} \includegraphics[width=8cm,height=3cm]{sprungbild.jpg} \item<3-> Thus if $T>2$, we have a {\bf jump} at time $t=2$! \item<4-> Hence also for continuous data, the optimal state for {\sc Dirichlet} control is in general {\em discontinous}. {\bf Continuity} is an {\bf additional constraint}, \\ {\tiny see \sc M. Gugat; \em Optimal boundary control of a string to rest in finite time with continuous state, \sf ZAMM, 86 (2006) pp. 134-150. } \item<6-> To do this, we need $y_0\in H^1(0,1)$, $y_1\in L^2(0,1)$. \end{itemize} \end{frame} \end{document} ```

Converting to HTML:

• I see transitions are already toggled with classes in the HTML with classes ltx_covered/ltx_uncovered. What remains is using the classes to achieve a slide deck user-experience when browsing (so adding new CSS and/or javascript). There is a similar need for displaying and navigating through each ltx_slide.
• The question I continue to have is whether we want to keep this design and write our own slidedeck frontend framework, or instead pick an existing JS framework and retailor the final HTML to fit a preexisting design. Either way, the "final step" for web-friendly slide decks is missing.
• I think some useful marker for the selected theme may be worth preserving. In this example it's Boadilla, but it is lost by the time HTML is reached. Having themed CSS files for beamer appears attractive to me.

Separately, once we have a slide deck setup working in a browser, going back to the full arXiv article poses some good stress testing of the beamer binding itself. I noticed some math mode errors inside the \frame{} macro variant which may be worthwhile to investigate. Getting the entire article to display correctly would be a lovely achievement, as it uses a variety of beamer customizations, as well as some related packages that aren't yet supported (multimedia.sty).

I hope this is helpful @brucemiller .

[brucemiller]

Ah, I see. We've had this discussion before. The way it seems to me is that there are several competing slide frameworks, all with their own fans, all are quite involved. Worse, they have very incompatible requirements of how the html should be structured and annotated, so that you really can't produce a generalized xml/html that could be used with several frameworks. Thus, without a much more pressing need than a couple of arXiv articles (or slides), it seems way out of scope for us to arbitrarily pick a framework, or take on the complex task of inventing our own.

That said, we would be quite willing to help, advise and adapt if someone wants to take on such a task. So, at the present time, taking beamer beyond where it currently is (other than dealing with specific bug reports) should be either wontfix or future.