"Compatibility" seems to be a frequent word in all software and programming languages.The question of how to view this in TeX/LaTeX may be taboo, and it may be out of place to ask here, but please let me.
One way of looking at it is that the composition result of processing the same source will be exactly the same no matter what distribution (*1) is used in any era.This is probably the toughest one.
Another way to look at it may be that "any distribution from any time period, one same source will be compiled" (no errors, no warnings will be added or decreased, but the results of the composition will not necessarily be identical).
There may be other ways of looking at it.In short, when you use the word compatibility,
I think it's hard to understand thatDon't you know any reference materials to see how TeX/LaTeX users' thoughts are distributed?(It is best to gather such answers, but even if they do not fall under this category, I would like to allow subjective answers because it is my own source.)
I think there are various opinions from the background such as .Please let me know if you have any comments or answers.
(*1) "Distribution" is hereby distinguished by the differences between TeX Live and W32TeX, as well as the differences in release/updated annual/date.There may be platform-specific differences such as Linux and Windows, but we exclude this difference because it seems almost impossible to absorb it completely on a binary basis.
(Please correct the questions accordingly so that they can be properly shaped/meaningful.You can close it if it is not appropriate.)
latex tex
It's the publisher's position.As for TeX/LaTeX compatibility, as an OSS user, I believe that the only important thing is to meet the following two points:
In general, I think software development incorporates regression tests to achieve these two mechanically.In fact, for TeX engines, the TRIP test referenced by @yuw (the current TeX engine is e-extended, so it's actually e-TRIP test) is a de facto regression test.
A regression test for formatting and class behavior on top of that is LaTeX3 team introduced when migrating from LaTeX 2.09 to LaTeX2e.This is to compare the difference between .log when hundreds of test files are processed by \tracingall.Even now, LaTeX2e's development is still working.However, I heard from a 2014 TUG that the current regression test system has evolved to eliminate noise (such as differences between OSs) mixed in .log with texlua.
As for the validity of the change that developers think is "better," I think it is up to developers to make autocratic decisions, just like other OSSs.As this application is also used by commercial publishers, it is easy to imagine that some users will insist on maintaining traditional behavior, but I think the final decision on 1 above is the right of OSS developers.Furthermore, as a user, we believe that they should either defend themselves by containerizing (Please refer to) or ask experts to provide technical support with that in mind (many other OSSs seem to have similar structures).
The nuance may be different, but the compatibility of TeX (or LaTeX) based on the material may be that TRIP test passes.
http://texdoc.net/texmf-dist/doc/generic/knuth/tex/tripman.pdf
Instead of writing, I would like to express my opinion mainly from the point of view of editing (i.e., the source is myself).
When I look at and edit TeX/LaTeX files, I ask and think about two things:
As for compatibility, I believe that it is acceptable to allow the compilation to some extent without any problems, even if it is not the same form as the author intended.
For example, even if $\times$ has changed to * instead of で, you should know that it is a "binary operator representing multiplication."
Of course, I'd like the writers to write commands and sentences with purpose instead of making ad hoc arrangements.
© 2024 OneMinuteCode. All rights reserved.