Following up on https://news.ycombinator.com/item?id=27610688, I tested the import of https://arxiv.org/abs/2102.04063 .
My impression at first sight is that there are a few things that could be improved.
-
The centering of the page (I did not find the setting which controls it in the TeXmacs file, in other imports I saw a package named maybe “latex-std”, it is not there)
-
The figures whose width is set in LaTeX with respect to
\textwidth
are set in TeXmacs with respect to\par
(see figure 2 for example)—the widths defined in LaTeX with respect to\linewidth
are set in TeXmacs\par
as well, but in this case it gives an image width that looks correct -
\xspace
does not seem to be taken into account by TeXmacs (compare e.g. “Basalt exhibits” with “Basaltlies” in section 1, paragraph “Content of this paper”)
I am leaving aside the algorithms (maybe is it interesting to support—or improve the support of—the algorithm2e package?), the subfigure
environment and the breqn package (this might be easy to support, it might just be enough turning a dmath
environment into an equation
environment)