« first day (4012 days earlier)      last day (904 days later) » 

8:30 AM
@AlanMunn the amusing part is, I have that game :)
@PabloGonzálezL ooh :)
 
9:09 AM
@JosephWright LaTeX2e <2021-11-15> pre-release-3 at texlive.net:-)
4 candidate nominations posted
2
 
9:34 AM
@DavidCarlisle :)
 
9:58 AM
@DavidCarlisle we have a very good list of candidates, very hard to choose!
 
@CarLaTeX We need Psmith and the !!/choose thingy. :)
 
@PauloCereda :D
 
10:39 AM
5 candidates!
2
 
@PauloCereda yeah!
 
11:00 AM
@PauloCereda not to scare you, but have you read the answer to question 8 ;-P?
 
@Rmano I did. :)
 
@DavidCarlisle why does unicode has so many mathematical number styles? Has it a mathematical sense?
 
11:40 AM
user image
6
 
@Rmano Also the others are no less :D
 
11:56 AM
@UlrikeFischer are but that's a \mathbf{4} not a \symbf{4} :-)
 
@DavidCarlisle ? What "are"?
 
yo'
12:09 PM
@DavidCarlisle yes, 2015. That one was super-cool!
 
12:27 PM
@DavidCarlisle :)
@yo' Yup
@PauloCereda Could be a lot more yet - there is plenty of time
 
@JosephWright ooh
 
@PauloCereda There are a lot of eligible people
 
@JosephWright definitely!
@JosephWright also, take a deep breath. :)
 
@PauloCereda I wonder if we will need a primary
 
@JosephWright ooh that's becoming complicated
 
12:32 PM
@PauloCereda It only happens if we have more than 10 candidates
 
@JosephWright ooh
 
12:55 PM
@cfr yay! Cwac!
 
1:08 PM
@UlrikeFischer the 4 in the comment that you were replying to
@JosephWright presumably not counting candidates who state they don't really want to be nominated
 
@DavidCarlisle (old message) By "format" do you mean LaTeX/context/etc.? Why would changing the format make a difference...?
 
@DavidCarlisle The Powers make it clear that nomination is not binding; I'm not sure how that shows up for candidates if we go to a primary stage
 
@user202729 similar to how we made expl3 faster it used to be that if you wanted to use expl3 code you needed \usepackage{expl3} which had to load that file from the filesystem, interpret a few thousand lines of macro definitions, set up unicode data for the entire unicode range, and stuff. Now that's (almost) all done once, while making latex.fmt and has essentially no run time cost per document. If you load unicode-math currently you see a visible pause in the terminal output
 
Okay so it's just like precompiled preamble. A little hard to lookup.
 
@user202729 yes
 
1:20 PM
By now the font part can't be put there, but the defs can
 
@user202729 sure but a lot of the definitions do not depend on the font eg Umathcodes for all the characters. but my question the other day was really just asking if the slowdown you saw was at the start loading unicode-math or whether typesetting each formula was noticably slower when using opentype math layout
 
I'm looking for ways to speed up xelatex compilation, if possible
Perhaps I should learn to read the source code instead
Although for some reason even with a preamble \documentclass[12pt]{article} XeLaTeX still segmentation fault out
 
@user202729 segfault? it's not supposed to do that:-) You could also try luatex instead (which realistically is more likely the future anyway)
 
@DavidCarlisle Not likely to be faster, though
 
@JosephWright well if xetex is segfaulting speed isn't the issue was my thought
 
1:27 PM
@DavidCarlisle :)
 
 
2 hours later…
3:21 PM
Even in LaTeX3, function arguments are still unnamed, right?
For instance I'm reading through __um_scan_sscript_aux:nnN and it takes a while to understand what #1 #2 #3 are
(and the tmpa family, but this is a separate issue)
 
@user202729 Yes, this is a feature of TeX
 
... bug
Well there are documentation and stuff (although for this one I don't find it in Unicode-math source document)
Also
About the tmpa thing... is it "general poor variable naming" or "reuse variable is faster"?
(not that I could do better but...) (as in "I know most LaTeX3 developers are here)
 
@user202729 Huh?
 
@JosephWright my thought exactly
@user202729 you know that as we post under real names (just saying...)
 
@user202729 You mean the scratch names? We have generally tmp, tmpa/b or sometimes internal (depending on the team's mood) for 'general scratch stuff'
 
3:34 PM
I mean "if it can be not clear what the parameters are supposed to mean they can be explained in the documentation"
Although self-documenting code is better
@JosephWright Yes, I mean
I feel that sometimes it could be named something else better
 
@user202729 It depends on the context I guess - inside say siunitx I use \l_@@_tmp_tl for e.g. stuff recovered from a prop and then used, it can be various things so using up more names seems of little value
 
@user202729 it could be implemented in a language where names are scoped to the function they are used in, but it isn't that means that some conventions used in other languages don't really apply to tex, even in expl3
 
The \l__um_tmpa_tl in the script-scanning part could be "supers or subs" I guess?
 
@user202729 I'd have to look at the code: Will wrote quite a bit of unicode-math several years ago, and we have refined the approach a bit
@user202729 For example, classical TeX code tends to reuse names a lot as they were scarce, so older code often reflects that
 
Okay (wait, names are scarce?)
Are names scarce? I thought any sequence of letters can be used as a name?
 
3:37 PM
@user202729 Not so much now, but they were
@user202729 As @DavidCarlisle ;)
 
As in "the memory is limited and TeX can't store that many variables"?
 
@user202729 when we released latex 2e, article class + amsmath left 50 names left. They are not quite so scarce now but that is the setting in which tex programming conventions formed
@user202729 no as in it's a fixed size hash table and (used to be) a compile time setting. (you can increase it in texmf.cnf without recompiling now)
 
Uh okay
So historically TeX can only store that many variables at once?
Where is it discussed?
 
@user202729 even now. pdftex and xetex will not do dynamic allocation. It's a fixed (but customisable) number
 
@user202729 By 'varibales' we mean csnames, but yes
@user202729 Source of TeX would be my start ;)
 
3:41 PM
Oops (sounds like hard work...)
 
@user202729 look at the end of every log
 
It feels like something as important as that (imagine the user getting "no name left" error) would be documented somewhere
> 21698 multiletter control sequences out of 15000+600000
That one? Okay
 
@user202729 yes, but when I started using tex there were a lot less zeros in those numbers:-)
@user202729 Knuth's original sources in tex.web have @!max_strings=3000; {maximum number of strings; must not exceed |max_halfword|} so you could only have 3000 strings (that's csnames, filenames, ...) and
@!string_vacancies=8000; {the minimum number of characters that should be
available for the user's control sequences and font names,
after \TeX's own error messages are stored}
 
@YiannisLazarides Hi! Long time no see!
 
@PauloCereda More people coming out of the woodwork!
 
3:55 PM
@AlanMunn Indeed!
 
Hi! Paulo, true indeed. Have been busy with a Project now it is done
Hi! Alan
Happy to still see a thriving community. I
 
@YiannisLazarides Hi! Good to see you. Will we see more of you in the future?
 
Sure, I got about 8-9 packages almost ready to go. Managed to color macrocode automatically! So I am excited to be back
3
 
Historically... C, Python and Fortran (← although I don't use this one) seems to be not that much newer than TeX, but they don't get that (unfortunate) convention?

(on the other hand most variables in that piece of code in particular are named correctly like `key` or `chain`)
 
@YiannisLazarides Hello
@user202729 What convention do you mean?
 
3:58 PM
Hi! Joseph
 
The "reuse variable names" thing
 
@user202729 Remember TeX is a macro expansion language primarily aimed at typesetting
 
@JosephWright this :)
 
@user202729 Huh?
@user202729 I suspect they don't have a limited string pool ;)
@user202729 Have you looked at register reuse: they were a lot more restricted than csnames (only 255 of each)
 
Aug 23 '12 at 14:09, by David Carlisle
@AndrewStacey nah nah na nah nah: my pool is bigger than your pool
 
4:00 PM
Yes I knew about that one
@JosephWright Well, every language is limited by the power of the computer at the time the language is written
And (it seems that) nowadays the amount is practically infinite
 
@YiannisLazarides oh no, regex :)
 
@PauloCereda Bruno's regexes in macrocode environment!
 
@YiannisLazarides woo
 
Amazing! slowed down compilation x 3
It is amazing that the group that joined plus minus ten years is still here and going stronger. Time flies
 
4:09 PM
@user202729 You can use namedef if you want named parameters. Though note that it's just a macro emulation of the feature: the engine has no support for that
 
It's okay, comment is enough
 
4:34 PM
Down to the last couple of hours ...
@user202729 I'm intrigued: what are you look at exactly?
 
@user202729 as documented here you mean? Then we search for it in the appropriate property list (\cs{l_@@_tmpa_tl} is |subs| or |supers|).
 
 
1 hour later…
5:57 PM
Hi, I am wondering can someone please help with this question: tex.stackexchange.com/questions/620191/… I am looking for multiline math in an algorithm or if there is another cleaner solution ...
 
One hour ....
 
@JosephWright hm?
 
@PauloCereda Until nominations close
 
@JosephWright ooh
@JosephWright then we can vote?
 
@PauloCereda Er, ask the bot
 
6:04 PM
@JosephWright oh no, it's a terrible bot
 
@PauloCereda Looks like it is one hour to go
 
Still 5
I'm wondering if there will be a last-minute rush
 
@JosephWright hmmmm
 
6:53 PM
@PauloCereda nominations must be in before supper
 
@DavidCarlisle oh no
 
One minute
Ah, no, GMT not BST - one more hour
 
What have timezones ever done for us?
 
Still five candidates: everyone writing their essay?
 
@JosephWright I am replying quack to every inquiry. :)
 
7:09 PM
> How would you deal with a user who produced a steady stream of valuable answers, but tends to generate a large number of arguments/flags from comments?

@PauloCereda: _/quacks angrily at them_
2
 
@PhelypeOleinik EXACTLY
 
7:57 PM
Three minutes
 
@JosephWright will there be fireworks?
 
@JosephWright uuuh, voting!
@PauloCereda they build roads? Oh no, wait, that was the Romans!
I now have 1 gold badge, 45 silver badges, and 90 bronze badges.
 
8:18 PM
Now we wait
 
@JosephWright yes do you think Percusse's should be flagged in some way, most people on the election page will not have seen chat.stackexchange.com/transcript/message/59441161#59441161
 
@DavidCarlisle well, if anyone reads his answer to the questionnaire they wouldn't vote for him (I hope)
 
@Skillmon it mentions ducks. it will win.
 
@DavidCarlisle but it mentions banning every duck content, including traffic from duckduckgo.
 
@Skillmon I know, but ...
 
8:27 PM
@PauloCereda the bot answered my question, I am quite surprised. Maybe it's angry at you.;-)
 
8:40 PM
@Rmano ooh
Yay, voted!
 
@PauloCereda 56?
 
@Rmano ooh
 
@JosephWright maybe people were too afraid of last-minute nominations after all the drama in the SO election? meta.stackoverflow.com/questions/412343/… :)
The one thing this election is missing: election forecasts which will turn out to be complete nonsense :)
 
8:56 PM
@samcarter That seems like a storm in a teacup - there probably should be a gap between nominations and election, but there wasn't, so that's the rules
 
@samcarter per twitter.com/rvf0068/status/1452729141803905024, could you add a halo to tikzducks, please? :)
@JosephWright ooh a gahp in one's 'oop
 
@PauloCereda did you vote for me?
 
@JosephWright SO meta just loves their drama :) Another candidate had to withdraw after meta users discovered plagiarism. (personally I'd just leave the comments open during the voting phase because I assume that many users will only start to think about their choice when voting has started)
@PauloCereda ohh :) "TeXnicus Maximus Vaticani" sounds very important! Can they release edicts to ban all uses of tabu?
 
9:33 PM
@PauloCereda Like this? i.stack.imgur.com/gN5fN.png
 
9:49 PM
user image
2
@DavidCarlisle Constituent badge awarded ^^^
 
@CarLaTeX This would have been the perfect candidate :)
 
@samcarter for sure!
 
why do I get 6 candidates? Is there a ghost somewhere?
3
 

« first day (4012 days earlier)      last day (904 days later) »