The real reason for requiring LaTeX submissions in academia?

Approaching Scheme takes time, and I agree that it may put off people, but with some experience it becomes fun to program with it. In my opinion the fun comes from the very strong modularity of the code that one can write—it feels like playing with LEGO bricks.

Said this, perhaps having an additional scripting language might help.

@amichail: I think @JoePass knows that.

Yes, I know. The first moment I was charmed TexMacs. But when I tried to make any more then I met a problem with scheme. Here, in the forum there some of my discussions about some problems which I met. I even started to learn scheme and racket. But I think you agree with me, that we would like to see some editor simple but powerful and useful for work, for work but not as a toy for playing. In every way, I support this project. GUI of Texmacs is good but in the end, I need for tex file to send to an editor. Then, I need to communicate with the editor, change the file according to the referee’s suggestion and send another tex file. But it is very difficult to make these operations through the Texmacs. It requires more works to go through this tonnel between Texmacs and Latex. I wrote above my wishes, only.

@pireddag, I think you miss the point: a basic user does not have to learn Scheme to use TeXmacs. Indeed I claim even a moderate user of LaTeX do not know how to use TeX macro system to write macros (apart from very simple ones). TeX macro system is more difficult to use than Scheme.

@JoePass raise a problem which I would like to understand better since it is not clear to me exactly what it is. Which kind of modifications you are referring to? Why they are easier to do in LaTeX/TeX than in TeXmacs in your own experience? It would be useful for me (and the other developers) to understand precise situations so that we can address the shortcomings.

PS: just to be clear, a real TeX macro looks like this

% Subsections.  We don't want a break at the glue if we were preceded by
% a \nobreak.
% 
\def\subsection#1 \par{%
   \par % be sure we're in vertical mode
   \saveskip = \lastskip % will be 0 if no skip
   \ifheading  % Did we just have \section?
      \nobreak
	\fi
   \combineskips\abovesubsectionskip
	\dimen0 = 2.5\baselineskip \advance\dimen0 by -\topskip
	\vskip -\dimen0
	\vglue \dimen0
   %
   {\subsectionfonts
      \leftline{\vrule width.5em height1.2ex depth.2ex \enspace #1}}%
   \writetocentry{subsection}{#1}%
   \nobreak
   % Don't go into horizontal mode within the command descriptions.
   \vskip\belowsubsectionskip
   \ifcommands \else \noindent \fi
   \headingtrue 
   \everypar = {\headingfalse \global\everypar = {}}%
      % only gets executed for \par in horizontal mode
   \ignorespaces
}

If you think that it will be easier for you to undestand this than a Scheme program…

@JoePass I’ve tried to give a look at your previous contributions to the forum, to find the shortcomings you were mentioning in this thread. I found the one regarding the workflow with .bib files and I tried to address on that thread your concerns. It would help me also to have a list of other similar concerns, in particular those which from your experience make the workflow to export to LaTeX for submission problematic. And also I would like to hear your opinion on how these problems would look like in a typical LaTeX-only workflow. I expect some of the problems to disappear but also other to be similar or equivalent (e.g. the extraction of a local .bib file).

A detail on this (also as a suggestion for @JoePass) : in my opinion questions on “how to correct errors in the LaTeX document that I exported from TeXmacs” are on-topic here. @darcy and @mgubi, what do you think? In this way the immediate concern (“How can I send my manuscript to the editor?”) would be addressed.

Following up on both @mgubi post above and on a post of JdvH on the mailing list (https://lists.texmacs.org/wws/arc/texmacs-users/2021-04/msg00059.html), knowing about these errors is helpful to develop a better export filter.

2 Likes

:+1: I see you are an expert in Tex language. I did not use Tex so deep as you do. I’d like to explain my point. In Tex community, there are a huge amount of packages and styles already done. If I need something special I use these styles. It is very useful - I do not need to write something as you wrote. This is the power of Tex and Latex. I need only find the corresponding style and open guide for it. In my system, there is texdoctk to any style. I have to learn only once about using preamble and new write commands. Couple examples. I need slashed letters (Dirac notations) and I use package slashed. Another example - I need hyperreferences - I use hyperref style. A couple of short strings in the preamble and I have the effect which I need. There are a lot of examples more. I already discussed here this question. It will be very powerful if it will be something for Texmacs. Another example about bibliography. If I need to use many references as a singe I use the apsrev4-2 style. The system of BibTeX with many collections of references is very useful for work. I discussed here this possibility in Texmacs. I wrote in savannah my suggestion. Another example appeared recently due to remote education. The package probsol solved all my problems with a database of problems for tests. Texmacs has a useful GUI, but changing something is rather difficult.

Another side of the discussion. At the present time, the communication with editors and journals is very simple. We use a style of journal to write paper. And any suggestions are easy to put in a file and then by using latexdiff it is easy to show editor the changes which I made. The latexdiff is very useful for writing joint papers. I see all changes which made my coauthor. I look at Latex as an instrument that I need to work.

Why do I try to use Texmacs? I see year for the year the problem of students when we start to write something in latex. I remember my problems with writing latex files when I was a postgraduate student. We usually use now with students the overleaf site. For this reason, I started to look into the wyswyg packages which can work with latex. As for me, the ideal to get a package with GUI as Texmacs and a good Latex file without convert. In the best case, it must be as a pen, without programming, but with quality as Latex. I tried to use Texmacs for the last paper but I see that the convert between Texmacs and Latex needs much additional work. And what is worst is that due to these converts many additional typos may appear.

This has not happened in a day. TeX exists since '84 and many people contributed packages to it. Some of these packages can be also used in TeXmacs, e.g. TikZ to generate pictures, others are not needed since TeXmacs provide the same functionality out of the box, e.g. hyperref. Of course we do not claim to cover all possible packages. Nor it is the scope of TeXmacs to reproduce all this large corpus of packages. However among all the examples you gave, only the one for aggregated references cannot be done right away in TeXmacs. Something like probsol could be easily implemented, but the core developers do not have the time to do all this themselves. We initiated the tm-forge repository https://github.com/texmacs/tm-forge to concentrate useful examples for users.

I do not see any problem with this: I write papers with coauthors in TeXmacs: we can see each others modification using Tools->Versioning tool and then we produce LaTeX versions of the documents to send to the editor. We run latexdiff on the various LateX revisions to produce a diff file useful for the referees.

This is not technically possible (or it is very difficult). Nobody managed to do this so far. Many people tried. BaKoMa TeX and TeXture are two successful examples but they also have a lot of limitations and BaKoMa TeX anyway need two windows, not only one as in TeXmacs.

This is a serious critique which we need to address. I do not share your experience, usually it does not take me much to convert my TeXmacs files to LaTeX, maybe ~5-10 min or so. If you experience serious problems it would be interesting to know which they are so that we can do something and improve the conversion experience. If we do not get bug reports we cannot fix these bugs. And bug reports are very much welcome.

Also in 1.99.20 Joris tries to eliminate many bugs. Certainly many more are there but with help of the users we can remove them. This is how an open source project improves.

Let me note that both BaKoMa TeX and TeXtures were commercial programs, and many other similar programs like Scientific Workplace are also commercial. While TeXmacs is free and developed by volunteers in their free time. Is also up to the community to help the developer to make it better.

Here I think extending systematically the documentation on the part of TeXmacs (including tutorials) would help; of course it will take time.

As far as I recall the two-windows view (and editing) is only a convenience for the user, but I may recall wrong.

My impression is that 1.99.20 is considerably improved with respect to 1.99.19; perhaps it is worth for @JoePass to try again converting his last paper into LaTeX (using 1.99.20) and seeing what does not work—I think many forum members would be happy to help with the conversion.

I started to collect problems with converting between texmacs and latex, but left this work. I’ll try to renew this work. With age, there is not enough time for everything. About hyperref. An analog is working in Texmacs by default, but it does not convert in hyperref in Latex. I discussed here in the forum how to include in the latex file in the preamble a couple of strings about hyperref.

1 Like

Thank you, I will try. I understand that this is an open project and it is the fruit of joint creativity, which I welcome in every way.

1 Like

Here something useful might be more possibility of customizing the conversion into LaTeX. For example, instructing TeXmacs that a given tag be converted with a given LaTeX construct.

This LaTeX code is already an easy example. Look at the following short one:

\def\cohtheory#1{
\expandafter\newcommand\expandafter{\csname #1func\endcsname}[1][*]{%
\MakeUppercase{#1}^{##1}}
}

extracted from a SE question.

He meant to convert between TeXmacs and LaTeX files. I suppose that he converts the TeXmacs file to LaTeX, then submit, and then received the comments from referees, and he converts it back to TeXmacs, and so forth.

I think that one should avoid conversion chains like TeXmaxs -> LaTeX -> TeXmacs -> LaTeX, i.e. converting back and forth. The LaTeX -> TeXmacs conversion could not be perfect.

It may be that I miss something about the workflow that you and @JoePass have in mind, because I have in mind a simpler one.

  1. Convert the TeXmacs file to LaTeX
  2. Refine the conversion so that the LaTeX file compiles and does what you want
  3. Submit, receive comments from referees
  4. Apply the comments onto the TeXmacs document (the comments are not tied to the format of the document, except perhaps in very few cases, of which I have none in mind)
  5. Convert again to LaTeX; the second time the editing of the converted file should be quicker as you have learnt from the first conversion.
1 Like

Yes, this is also how I perceive to work. It seems to me that, for him, the referees edit the LaTeX file itself and then send back to him.

He also discussed the collaboration, where he also converts back and forth between TeXmacs files and LaTeX files. I suppose that this might never work perfectly.

Nooo :slight_smile: It seems to me quite out of the ordinary, in all cases that I have information about the referees send their comments separately from the submitted article.

On the opposite, a collaboration with mixed TeXmacs/LaTeX as far as I can see does not work very well, but I do not know how to improve the current status of things. Perhaps having more options for personalizing the LaTeX <-> TeXmacs conversion might help (I have in mind choosing which LaTeX construct should a given TeXmacs construct be translated to … it might be a delicate thing because doing this a user might introduce bugs). I will post a request on the bug tracker (there is an option for the feature requests).

This seems too ambitious. Previously, there was a simple feature request to allow the customization of the conversion of simple constructions (eg. \( and \) for inline math), which has not been implemented yet. I guess that general customization like this would be too difficult.

The HTML conversion already has a similar feature. From the HTML conversion page in the manual:

In addition, given a macro my-tag, you may customize the rendering of the tag when exporting to Html by defining a macro tmhtml-my-tag with the same number of arguments. For instance, by putting the declaration

<assign|tmhtml-strong|<macro|body|<with|color|red|font-series|bold|<arg|body>>>>

inside your style file, all strong text will be exported to Html using a bold red font.