@barbarabeeton Oh, they know things much better than I do (and I'm just thinking out loud). Equation numbers to the left is probably the smallest problem in the question. I am also hesitant to the logical structure in ending proofs with a square in a displayed formula, flush right. It looks a bit like tagging the equation. Maybe the real problem there is ending proofs with displayed formulas.
I expect the answer to the following is no, but I'll ask anyway. Suppose I defined something like \newcommand{\dd}{\ensuremath{\delta}} as I use this symbol often, both in formulas and in isolation. Now if I write foo \dd bar, then it will be typeset without a space in front of "bar", i.e. "δbar" instead of "δ bar".
At the moment I use foo \dd\ bar. Is there a simple way to define \dd in a way so that this is not necessary in text mode, yet it still works fine in math mode?
When I insert a short maths section in a sentence, it renders fine with a space each side as expected:
...returned $\mathcal{P}$ from...
If I define a command to do the same thing, it also renders with a space each side:
\newcommand{\pc}{\mathcal{P}}
...
...returned $\pc$ from...
However, i...
I'm certain I have read that xspace can cause more problems than it solves. But I cannot find anything to that effect either on this site or the web. Are there any drawbacks or risks associated with it?
I have used xspace for a long time to take care of adding space after usage of some macros. Usually, it is with a macro to format a word consistently, as in:
\documentclass[]{article}
\usepackage{xspace}
\begin{document}
\newcommand{\matlab}{\textsc{matlab}\xspace}
\newcommand{\enterprise}{\emp...
@mickep For example, a list can be seen as needing two templates: a generic block object and a generic item one, and the way the list looks depends on the settings that are applied to those two objects
@mickep texdoc.net/pkg/xtemplate, but my point was really that after about 25 years of working on them, we are now looking at getting the ideas moved into the LaTeX kernel
@mickep I expect a 'stable' version to be in the Fall 2022 LaTeX release - I'm working on the agreed changes now
@mickep Sort-of - ConTeXt doesn't try to pre-evaluate keys as far as I know (taking 'key = val' and converting to \def\foo{val} for speed at point-of-use), and I've not seen anything like objects (e.g. 'generic block')
[This](https://chat.stackexchange.com/transcript/message/61248416#61248416) is what I want to do. Right now, subplots are centered in the figure env. Would it be possible to right align all the subplots?
@JosephWright I can't say if it is new, I forgot to check the old version on ctan (it should be in the doc folder). But I uploaded then a version where I exchange the files manually.
@UlrikeFischer Trying to reproduce I encountered some issues running the tests locally. (Mostly getting \marks1` instead of \marks4 in various places.) Do I need some special package versions to run these?
@MarcelKrüger that sounds like latex-dev version problem. But I can't check now what I have on my laptop, we are visiting Trient currently. You could simply update the tests locally.
I managed to avoid these failures by removing all local latex-dev/latex3 installations so I now only have failures due to Windows line endings which are ... more interesting to deal with. But that's an issue for another day.
@UlrikeFischer I think the issue is that PNG files have to be marked as binary data, I send a PR for tagpdf.
@JosephWright @UlrikeFischer That also poses the question if the default for binaryfiles in l3build should be extended. Currently we have .pdf and .zip. I guess we might want to add most image and font formats used in LaTeX.
@MarcelKrüger quite probably. And as it is wrong in texlive too, I guess it was wrong for quite some time. ctan seems to have improved their tests that they now complained about it (they complained about duplicated files as both png had the same content).
@mickep I generally do not understand these users "solve problem, but with no packages". It might be for some mathjax solution, but they know we tend to close mathjax questions.
@mickep -- Actually, it's exceedingly rare to have a proof ending with a numbered display (and it's quite a nasty problem when it does), but I've seen some proofs where it would have weakened the effect by artificially adding text at the end. It's not frequent, but also not unknown.
@yo' -- Oh, I answered before I read your comment. (Hi.)
@barbarabeeton (and @yo') I think I support the suggestion from Knuth - Mathematical writing: "Paul Halmos introduced the handy convention of placing a box at the end of a proof; this box serves the same function as the initials ‘Q.E.D.’. If you use such a box, it seems best to leave a space between it and the final period."
@mickep -- That's a valid point of view, but what if the last line of a proof is so filled with text that there's not enough space for the box or "Q.E.D."? It gets lost visually if it just starts the next line. In that case, the end of the next line seems the appropriate location. Just something to think about.
@barbarabeeton Yes. I was thinking about the box (not "Q.E.D."). I wrote an analysis book with a colleague, and we used this. There were one or two occations where the situation you describe happened, and then we reformulated slightly.
@mickep Indeed, Halmos used a black rectangle a quad or so after the final period. A bit awkward when a display ends the proof, but such situations are best avoided. The main problem with this style is when paragraphing sends just one word or, worse, a word part to the next line (not just the box, of course). Rewording is needed in order to get a good typographical result.
@barbarabeeton Ending a proof with a numbered equation is one of the worst things I can imagine: what purpose does the number serve? It possibly happens when all displayed equations are numbered, which is a bit silly by itself anyway.
@egreg -- Yes, I think that the "number all displays" is the usual situation when a numbered display ends a proof. Why referees don't suggest improvement is beyond me, but the current editorial policy at AMS doesn't encourage rewriting. And I don't anticipate improvements any time soon ("too expensive"). Some of us old-timers find this very discouraging.
@DavidCarlisle -- I'm determined to get to the bottom of that decision. It appears to be older than the ISO diktat, but I've got a lot of library/historical research in front of me. (@mickep has already helped with that.)
@barbarabeeton I will be interested to read about your conclusions. (As far as I remember, Lansburgh merely writes that the upright d is a "förfinat skrivsätt".)
@mickep -- in the current AMS style guide, it's shown as acceptable if the author used is consistently, which rather points at it being considered an individual choice. I still have to dig into Chaundy and the first edition of Ellen Swanson's Math into Type to see what, if anything, is specified. (I suspect it may not even be mentioned.) When I do get to the bottom of the matter, it'll be published in TUGboat. (But I'll let you read it before it goes to press.)