Is it only me or did they uglify the voting buttons further?

@Skillmon see samcarter comment above (yesterday). (I don't know how to link to a post...)

@Skillmon The second you said. Awful :(

@Skillmon I'm wondering if one could modify meta.stackexchange.com/a/389568/237989 to bring back the beautiful pen nibs voting arrows?

Do we have any css wizzard who knows how to replace the voting buttons?

FWIW, Ubuntu 22.04 just pushed the lualatex security update today...

@SylvainRigal FYI to link to a post, you can click on the downwards triangle on the left hand site and copy the "permalink"

@Skillmon It is surprising that they sell this change under the umbrella of improved accessibility. Previously, the information of an engaged vote was conveyed by both a change in colour as well as a change in luminosity (light gray compared to the saturated theme colour), but now only the colour changes and it is also in a much smaller area, this makes it harder to notice (the background colour of the circle is so faint, that I don't think it helps at all).

@samcarter Thanks !

@SylvainRigal You're welcome! (and thanks for your work on the beamer2ppt answer, a very detailed overview!)

@samcarter I’m glad to have some feed-back. Thanks ! I’m not very confident about the utility of such a long answer…

@SylvainRigal It is certainly useful as a duplicate target - similar questions come up regularly.

Particulary I don’t know if the list of related post is a good practice or not. I wonder to ask about this on meta but I don’t want to pollute to much with all my questions…

@samcarter would be wonderful, the old pen nibs were so beautiful. And they apparently don't stop at "looks boring but ok", they have to push further!

finally ...

@UlrikeFischer Yeah! Will the Bär let others have a look inside or do they have to wait until he finished reading both volumes?

@UlrikeFischer congratulations on your achievement!

@samcarter I'm not sure if I should let him (or anyone else) look inside, I fear he will find a typo ...

@UlrikeFischer ask the Bär to not tell you :)

Welcome back old friends:

(the alignment is a bit off...)

// ==UserScript==
// @name         SE tex pen nib vote buttons
// @description  modified from stackapps.com/questions/9443/…
// @match        tex.stackexchange.com*
// @match        tex.meta.stackexchange.com*
// ==/UserScript==

(function() {
    let allbtns = document.querySelectorAll("button[class^='js-vote-']");
    let downsvg = document.querySelectorAll('.js-vote-down-btn .iconArrowDown');
    let upsvg = document.querySelectorAll('.js-vote-up-btn .iconArrowUp');

@Skillmon ^^^

Hi all. I discovered that this problem is quite common in several fonts (for example, Palatino). Do you think that using for the minus sign $\vphantom{+}-$ by default can give some unexpected problem?

@samcarter -- I think by "accessibility" they're referring to people who insist on using a dark background. It's pretty clear that many changes have been directed at solidifying the "bottom line", even though service to participants suffers.

@barbarabeeton In dark mode, their new icons are even worse with respect to contrast to the background

@samcarter -- Hmmm. They sure weren't very specific about why they made that design choice, so that was just my best guess.

@Rmano Globally? I learned some days ago that the minus in CM has the same height and depth as the plus.

@mickep Yes. Given that the minus and plus in circuitikz are positioned independently, in different boxes, I'd like that they align always --- I made another example in github.com/circuitikz/circuitikz/issues/721

@barbarabeeton They are no longer specific at anything. My best guess: because they don't know themself why they do random stuff. I found meta.stackexchange.com/a/389576/237989 a nice summary

@s.harp -- An undefined \newcommand ?!?!? This is just a guess -- is it possible you have placed these lines of code within another definition? It's bad form to define commands in a local environment, though I wasn't aware it might be prohibited.

@barbarabeeton alterring the position of the copy pasted snippet fixes it, thanks for hte hint!! If the snippet is after "\usepackage{thmtools}" then I get this error, if it is before that command i dont get the error

which is weird?

@mickep -- Yes, and this is intentional, for the reason illustrated by examples like this one. There have been complaints about the illusory height, but that can use \smash in a workaround. A bug should be reported for Latin Modern if this is really the case.

@s.harp -- Wow! Please add this at least as a comment on the posted answer. It may help someone else.

@barbarabeeton See ^. I'm not sure it should be considered a bug, though.

@mickep -- The bounding box only references the baseline. In situations where there is no clear baseline, that becomes irrelevant. The CM alignment is based on the math axis, which is more appropriate and foolproof. I'll see if I can devise an example that "works" without tikz.

@Rmano Since all OpenType fonts I've looked at behave as Latin Modern, I guess another approach could be to add a reasonable strut to all labels to enforce consistency..

...even if switching to Times.

@Rmano Oh, cm math with times text? Doesn't sound as a good choice. Or is it just those symbols?

@mickep I think they are just those symbols, but I have not explored more: tex.stackexchange.com/questions/687173/…

@s.harp not really thmtools redefines \newtheorem and it now expands its arguments. I wouldn't load thmtools together with such code, the probability that something breaks is high.

@mickep Yes, \strut solves the problem, but it also moves the signs away from the line, which can be somehow unexpected (although I think it looks better...). maybe using the \vphantom{+} is like inserting the minimum possible strut there...

@Rmano I think you might have access to differently aggressive struts.

@Rmano -- If it's just the plus and minus, I'm sure this is the reason. @mickep -- my simple-minded attempts, which isolate the minus in common positions off the baseline, have failed {i.e., the alignment is what is normally expected). I'll keep trying.

@barbarabeeton IEEEConf uses mathptmx, and I really can't find any special treatment for plus and minus, so I am not sure about that. Notice, I have no problem in normal math or text, my problem is when the two boxes (with just $+$ or $-$ in it) are aligned with a common anchor --- like in tex.stackexchange.com/a/609342/38080

So I think that for my (very limited) scope, using the \vphantom{+}- trick works (unless there is a font where the block for minus is taller than the one for plus, which I hope does not exist...)

@Rmano -- Ha! I think with considerable certainty that you're safe there. It's quite unusual for a bounding box to enclose so much empty space around a glyph. (I still think it's a bug in lm, or at least a misfeature.)

I have added an example to play with on GitHub.

site passed a quarter of a million questions

@Rmano You can always also do \vphantom{-}+ if you are into that solution. :)

@DavidCarlisle Which means that some people have more than one point per question on average. Some even four-five points...

@mickep I have more than one point for all of my questions

@DavidCarlisle Oh, empty set again. But one day you will be brave enough to ask questions. :P

@mickep "why don't people use context?"

@DavidCarlisle Oh, that was your newly created fake account? :P

I'm disappointed you did not choose a better name for it.

@mickep This must be his sock account: cooking.stackexchange.com/questions/68259/…

@samcarter Haha!

@mickep ;-) defensive programming...

@Rmano Well, I would go with the struts. :)

People who use numeric citation schemes: do you ever refer to authors by name at all? E.g. Knuth wrote TeX. [1]. And if I want to refer to all of Knuth's algorithms, I can't say "All of [1]'s algorithms". So do you just avoid it, or are there ways that people do this?

@AlanMunn This depends on personal preference. In CS papers I have seen pretty much everything. Personally, I do write things like “X and Y propose Z [1].” and if I did mention the author names beforehand I would also use “All of X's algorithms [1]” instead of “[1]'s” algorithms. But as I said, this is personal preference. And at least in CS you would get away with pretty much any position of the numeric citation.

@TeXnician Ok. Thanks. My linguist head doesn't like using references as literal parts of sentences, so your version is what I would also do. We generally use an author year style but for abstract writing going numeric can save a lot of space, which is why I'm asking. :)

@AlanMunn -- The AMS doesn't permit numeric (or any \cite usage) in abstracts, Abstracts are provided to abstracting services in a form that may be redistributed without the accompanying bibliography, so any specific reference must be provided in full. Regarding the use of any citation in a paper proper, I'd input the author's name explicitly, and use \cite just for the link. (And check the spelling double carefully.)

@AlanMunn -- On a totally unrelated topic, do you suffer from this phenomenon in your part of Michigan? twitter.com/ellenbacca/status/…

@barbarabeeton This is abstract writing for a conference. Not the abstract of an article.

@barbarabeeton In our neighbourhood there's only one cottonwood that I'm aware of. So we don't really get blizzards like this.