A seminar on TeXmacs I am preparing - your opinions please

Dear all,
I am preparing a seminar on TeXmacs for my workplace, the University of Cagliari (at the dept. of Physics)
These are the slides I have prepared

pdf file
tm file

I plan to first introduce TeXmacs (first few slides), then show how to use it by writing up some text, math, references and so on in front of the audience, helping myself with a software that shows keypresses on the screen; I can post the example document that I would type in front of the audience.

I would appreciate comments/suggestions/criticism. I need to complete the part on export, as I haven’t been able to use a local style sheet for html.


You can check how “Notes on TeXmacs” does it. We use a custom CSS. I forgot how I did it. There are some typos in your presentation (e.g. “mechanismo”), maybe run the spell checker. Otherwise great work Giuseppe! Are you back in Italy?

Let us see how the typing in front of the audience goes.


I think the idea of writing up some text, math, references and so on in front of the audience is really great.

You can also take some documents/presentations from https://github.com/texmacs/tm-forge to show the broad usage and styles already available.

Excellent initiative !

Maybe also showcase the programming sessions and how can be used similar to (without doubt more popular) Jupyter notebooks.

1 Like

The suggestion is good but I think it is better to do it in detail in a possible follow-up seminar, as I think the topics are enough for one seminar. I am going to say that it exists, nevertheless.

In your presentation, you mention that Latex import cannot be perfect. Do you plan at that point to make it more concrete by demoing live the import of a latex paper? They will realize what kind of imperfection they may encounter and that it already provides a pretty usable document. Importing a wikipedia page with math can also be an impressive demo (but that probably means using a build containg the recent mathml import fixes).

You are right, I am vague about the LaTeX import because I do not understand what limitations the Turing theorem places on it. Nevertheless I think that it is important to say it, perhaps I should look for a reference (I think a post by either Joris or Max will be ok).
I think I will prepare a LaTeX document to import.

Now I am trying to figure out how to use a local css stylesheet for the html export; “Notes on TeXmacs” does it through Scheme, I think I need to make it work either through the graphical interface or through assigning a value to an environment variable, so far I did not succeed.
I found out that the list of css stylesheet in the preferences dialog is set up in progs/texmacs/menus/preferences-widgets.scm, but did not understand how to add a local one (we don’t want to do it in the installation files I think). Do you have any suggestions?

For the import of a Wikipedia page, perhaps it is better to use the latest public release, so not to use a build that has not yet been released (and so perhaps not to import the Wikipedia page).

The assignment of a value to the html-css environment variable has to be done in a style file: I got it from Environment variables for customized HTML export do not work

Oh, yes. I forgot about it. It has to be known to TeXmacs well in advance, so you have to put it in a style file. As for LaTeX conversion you might want to look at this paper of Joris for background information: https://www.texmacs.org/joris/latexconv/latexconv.pdf

I would not stress too much Turing in relation to LaTeX, said vagualy is not very clear what you mean. The problem is that LaTeX is not a data format. So for example a user can redefine what tag is used for equation in some macros and for a converter, is very difficult (if not impossible) to extract this information by just looking at the file and not executing every single macros. The problem is even worser because the document can import an obscure package which introduces new macros that the converter knows nothing about. If you want a conversion at a semantic level, then the problems is practically insoluble. It is more or less the same problem which makes PostScript unsuitable as a data storage format and the reason why PDF was designed.

You could point out some statistics about the efforts to convert LaTeX into HTML. It seems the current state of the art is around a 75% success rate:

The remaining 25% may be very difficult to convert due to the reasons @mgubi cites.