« first day (3127 days earlier)      last day (695 days later) » 

2:53 AM
@HenriMenke I reported him. At least such comments make unnecessary and (very) annoying noise to my inbox
Anyway if we think positively, probably he had some serious mistakes in copy and paste, maybe he tried to paste to his own repo but he opened the TikZ-PGF repo
4 hours later…
6:49 AM
question about buggy @egreg package I see
@DavidCarlisle ?
ä is in the extended latin block and n is in the basic latin block so there is a transition there, but you would have hoped \setTransitionsForLatin would have not inserted any code at that point as both those blocks are listed as part of the latin block, but apparently not.... — David Carlisle 12 secs ago
7:07 AM
@DavidCarlisle Why is the font switch only after the ä and not also before?
@DavidCarlisle Not my fault, nor my package.
7:45 AM
@egreg you are credited in the file, so you inherit the blame:-)
@UlrikeFischer I was leaving it for @egreg to trace but I suspect the package makes some assumptions about what is safe, it offers the user "enter" and "exit" code for each block but xetex only has a single insert, the interchartoken at a boundary the package isn't clear what happens at a boundary if the exit of the left class and the entry of the right are both specified, nor if anything is inserted at boundaries between blocks that are contained within one of the meta blocks like latin.
@DavidCarlisle which package? (I credit egreg in one of mine)
@Skillmon ucharclasses (see link above)
@DavidCarlisle strange one, indeed.
Why do we downvote to a total vote of -3 or even lower? Weren't we a welcoming and forgiving community with the convention to only downvote to -1 (except for some extreme cases, like e.g., worsening the site design in every possible aspect)?
8:02 AM
@Skillmon people will downvote if they wish and given that the rest of the network regularly downvotes lots of new users will not know or not agree with a "-1" policy, I don't think it was ever really that regularly enforced just that a few regulars regularly voted for bad questions to top them up if they got a very negative score. I still do that occasionally if I notice one.
@DavidCarlisle yes that is what happens. It actually depends on the order of the declaration.
does something different if you switch the declarations.
8:21 AM
@UlrikeFischer I guessed as much-) let's just blame @egreg and leave it at that.
@DavidCarlisle well, when I was new there was like never a question downvoted to more than (or less?) -1. And I liked it that way. My first question on SO got downvoted to -infty before I deleted it and fixed my issues on my own.
@Skillmon are you sure? when you were new you could not see up and down votes, just the total.
@Skillmon take this example vim users pushed that to -10 but fortunately 26 right minded souls have kept the total positive. new users just see a total 16.
8:39 AM
@DavidCarlisle I meant the total. Still the general principle applies, when you're new and your question gets donwvoted too much this might cause the wrong impressions.
@DavidCarlisle oh, subjectively I'd downvote that answer 10 times, but objectively it is not a good answer and might get a downvote from me, as you don't provide any reasoning for that, and I think that there should be a bit of reasoning with the opinion based answers, some objective arguments why this is good. See for example the other Emacs answer (still subjectively a bad answer), that one is objectively good.
@DavidCarlisle and that other one got no downvotes.
@Skillmon I know, I even got a mod-added, non removable banner telling me off:-)
@DavidCarlisle I'm jealous ;-)
@Skillmon yes but many people just join for a while and come form other sites where downvoting is more common so I think it is impossible to expect there is no multiple downvoting, the only way to have a -1 policy is to get people to upvote bad answers more.
@UlrikeFischer even harder to get than a gold badge.
@DavidCarlisle which is quite hard now - before I even see a question someone has already answered it ;-(.
9:05 AM
@UlrikeFischer those @marmot s get everywhere.
9:16 AM
@Skillmon Negative votes are meant to allow community as a whole to remove poor material
9:55 AM
@DavidCarlisle oi
@DavidCarlisle ooh
2 hours later…
11:35 AM
@UlrikeFischer Somehow I suddenly got that today
To my surprise
11:53 AM
@JouleV I'm not surprised ;-)
@JouleV non-round badges don't count (@UlrikeFischer)
12:10 PM
@UlrikeFischer Btw [bug report] Your duck badges don't work on other sites :(
@JouleV they are not mine, you need SamCarter to get alternative pictures.
@UlrikeFischer Well at least I took the CSS code from one of your chat messages so you are the first to whom I reported
12:31 PM
Hello my friends in the U.S.!

Could you help me please: How do you say in formal speech for "Draw a circle..."?

In German we would say: "Beschreibe einen Kreis", that would be: "Describe a circle"....

Complete Example:

· "Sei k ein Kreis, beschrieben um M vom Radius r"
---> "Let k be a circle, described around M of (with?) radius r"

· "Beschreibe einen Kreis um M vom Radius r"
---> "Describe a circle around M with radius r"

Is this correct?
@cis No. 'describe' can't mean this. I can't think of a more formal term (and if there is, I suspect few people would know it.)
@cis I'm not in the US but.... "Describe" while it does have a technical meaning close to what you want is almost always used more casually to mean "talk about", I think I would say" Let k be a circle with centre M and radius r"
@DavidCarlisle I wasn't even aware of the technical use of 'describe'.
@DavidCarlisle @AlanMunn OK, the definition is clear (" Let k be a circle with centre M and radius r") and for the process?

"Draw (BETTER WORD?) a circle around M with radius r" ?
@egreg Writing your book?
12:43 PM
@cis 'draw' would be fine. (btw, the American spelling of 'centre' is 'center'.)
@AlanMunn definitions.net/definition/describe gives a websters definition of to represent by drawing; to draw a plan of; to delineate; to trace or mark out; as, to describe a circle by the compasses; a torch waved about the head in such a way as to describe a circle
@DavidCarlisle Yes, I looked it up, but had never come across that usage.
@AlanMunn you should do more Euclidean Geometry:-)
@DavidCarlisle That ship has sailed. :)
Ok, "draw" is understandable, but not very formal-speechy, I think.

We say
· "beschreibe einen Kreis" --> "Describe a circle" (ok, thats wrong, we clarify)

· "ziehe einen Kreis" --> "stratch / drag / pull a circle" (I do not think that's correct)

· "schlage einen Kreis" --> "hit a circle" (incorrect too. The meaning of "to hit" is an allusion to the compass)
@DavidCarlisle Ah, ok. Due to Webster I can say "to describe a circle by the compasses".
Very good.
12:53 PM
@cis Why do you need formal? I would go with understandable. I think you're stuck with 'draw' unless you assume that your audience consists of geometers.
Actually the better translation for "beschreibe einen Kreis" would be "outline a circle" which, I think, doesn't sound too bad.
@cis No, this is not very good. Using a word that nobody will know doesn't help anyone. Dictionaries don't determine meaning, language use determines meaning.
@cis use "draw":-)
@AlanMunn I would like to write a book, Sir.
So I think I could not talk, like we do, when we go out for a beer... ;)

@DavidCarlisle @0xC0000022L But "outline" is good too, I think.
12:57 PM
@cis it will make people stop and read the sentence twice to be sure of what you mean, unlike draw which will be clear on the first reading.
@cis Just because it's in a book doesn't mean you need to use language that nobody will understand. Stick to draw.
@DavidCarlisle Or they will think 'This is a bad translation of some German word".
@AlanMunn Yes, this danger exists. :()
Actually why not use a placeholder such as this would be a bad translation of a German word/phrase ;)
A suggestion: "Line a circe around M with radius r" OK? "Line out a circe around M with radius r" Better?
@cis No. Please use 'draw'.
1:03 PM
If you are really looking for alternatives to "draw" in "draw a circle" I strongly suggest you hop over to english.stackexchange.com and confirm to create an account there and ask ... at least the number of native speakers of English will be bigger there and the gamification aspect of the site will ensure someone will rush to help you out.
Of course there is also a chance that they will repeat the advice you got here; to use "draw".
@UlrikeFischer Compile with latex+dvips+ps2pdf. Watch the terminal while running ps2pdf:


  \special{ps:: (AtEndDocument: Howdy world!) print}

  \special{ps:: (AtEndDvi: Howdy world!) print}




@0xC0000022L @cis You've got identical responses here from a mathematician and a linguist. And you seem to have an idea that because a word is informal in German, its translation in English is also informal. This is simply wrong. And formality shouldn't be an aim in and of itself in any kind of writing.
@AlanMunn :() Ok, I will think about "draw".

@0xC0000022L @DavidCarlisle Do you know the book "The Bronstein" in English? I think that's a good example of archaic mathematician language. But it is still possible harder. Probably depends heavily on the translation.
@cis heard of it, sifted through the TOC, stayed away from it ever since.
@cis lines are straight and circles are round, "line a circle" sounds like "square a circle"
1:16 PM
@DavidCarlisle Haha, ok...
@DavidCarlisle @0xC0000022L @AlanMunn Somebody wrote me: "Consider a circle(circular line) centered at M and radius R" Mmmhh....
@AlanMunn I am very well aware of the differences of word use between languages (and my limitations in regard to my knowledge and use of English as non-native speaker). In fact words in different (related) languages sharing the same origin is kind of a hobby. Needless to say that more than once the contemporary meaning didn't match a 100%.
However, your point about formality is well made. A book - in my opinion - is first and foremost a vehicle to transfer knowledge. No need to complicate matters by trying to sound ... well, overly sophisticated (?) ...
@0xC0000022L Sorry that comment was mainly aimed at @cis not you (since the original request is theirs not yours).
@0xC0000022L well, we should, most importantly, not conserve the meaning of the words. Like, of the word "conserve".
1 hour later…
2:28 PM
@DavidCarlisle @AlanMunn @0xC0000022L Somebody wrote me: "Circulate a circle around M...." -- not bad... :)
2:42 PM
@DavidCarlisle I blame Claudio, of course.
@egreg two buggy packages in one day!
3:18 PM
@DavidCarlisle ooh
@DavidCarlisle oh where?
Q: Showidx gives "Undefined control sequence" error

Joseph QuinseyThe following MWE with showidx and imakeidx: \documentclass{book} \usepackage{showidx} \usepackage{imakeidx} \makeindex \begin{document} Test\index{xxxx} \printindex \end{document} generates the error: ! Undefined control sequence. <argument> \ifdefequal {\imki@jobname }{\@idxfile }{}{...

@DavidCarlisle oh. And I have \arabic in chessboard and so it breaks with bidi. Who can I blame?
@UlrikeFischer egreg
@UlrikeFischer ooh
3:33 PM
@UlrikeFischer Well the Persians invented chess, so... :)
@DavidCarlisle ;-) even without the arabic problem the outcode under RTL is well interesting. I think I should tell people to wait for version 2.0 ...
can't you just force ltr for the board?
@AlanMunn ooh a conundrum
@DavidCarlisle yes switching to english works ok. But I don't know if I want to correct this now, I'm not even sure if the docu still compiles ;-(.
3:49 PM
@UlrikeFischer open an issue in gnats and come back to it in a few decades
@DavidCarlisle I opened a github repo (but I refrained from calling it "chessboard-fixed").
How can I use ">1" in a s table from siunitx? The table column: S[table-format=2.0]
@MaxR you can put {>1} (including the braces), but this loses the alignment. I'm not sure whether there is anything like this.
4:07 PM
hey, Wisdom of the Crowd
quick question
is there a command-line option to the latex binaries that can set the timestamp used to set the output of \today and related commands?
@yo' I tried that and in my case it doesn't really influence the appearance of the table
say, something arXiv could add to their compilation environment that would fix things like this:
@EmilioPisanty I would guess the date in the header is hardcoded. Especially as default \today uses a different date format IIRC
@yo' the one after Dated: in that image? I checked the source, it uses \today
arxiv.org/abs/1001.0019 if you're curious
but there's examples all over the place once you start looking
@EmilioPisanty ok I see. That could be worth writing to the arXiv webmasters as this is indeed strange. However, it's also possible that the publishing of the paper got delayed; AFAIK the timestamp is only added later to the final PDF.
4:14 PM
@EmilioPisanty I would imagine they have frozen the epoch settings to get reproducible pdfs, not necessarily that helpful here but..., anyway it is better not to use \today in a submission as you want the authoring date not the date it was last run through tex
I'm raising an issue on their GitHub repo, and I'm wondering whether there's some useful compilation-time option that could be used, or whether they need to freeze the epochs as David says
@DavidCarlisle they haven't frozen the epochs.
@EmilioPisanty Oh OK, so it was just an old tex run?
and yeah, it's better not to use \today in a submission, but that's beside the point - a whole lot of arXiv eprints use the syntax and they're starting to get wrong dates
@yo' it's not that the publishing got delayed. arXiv caches the pdfs for several years but at some point they get deleted, and when that happens they only get recompiled when somebody asks for them again
and, when that happens, they get imprinted with the date at which the pdf was requested, which then gets cached
(as I understand things, that is)
hence the November 2018 on that paper
@EmilioPisanty so what could arxiv do (other than never regenerate the pdfs?) if you want \date{\today} to do anything useful?
Does any of you on linux have issues running for foo in *.pdf ; do  pdfinfo $foo ; done in a folder with suitable pdf files? BMy box says pdfinfo does not exist, but clearly do when I run it on a single pdf file.
4:19 PM
@DavidCarlisle what you said - freeze the epochs
i.e. supply a "fake" system time to the latex binaries
or, alternatively, use some suitable compile-time option that will influence the time that latex things it is
@EmilioPisanty that's a relatively new feature, but I think they have a new enough tex, but not everyone will be happy if they submit a paper with \today and it comes out with some arbitrary date like 1st Jan 1970
@DavidCarlisle I'm suggesting that the compilation runtime should happen in a sandbox wherein the system time is set equal to the timestamp at which the paper was submitted
@EmilioPisanty yes I have no idea if their back end set up could easily do that, have per-document run time settings, but perhaps it can.
@DavidCarlisle I'm currently writing it up as an issue on their GitHub repo
I'd like it to be as useful as possible
hence my question
is there a simple solution that only involves options fed to the latex binary at compilation time?
@DavidCarlisle add \def\today{24th May 2019} in INITEX phase and recompile the format daily? I agree, too much overhead. They should simply add "do not use \today" in these guidelines: arxiv.org/help/submit_tex
4:23 PM
@EmilioPisanty you don't need a sandbox and change the system clock simply set an environment variable
@EmilioPisanty well, if you consider pdflatex -jobname myfile "\def\today{24th May 2019}\input myfile.tex a "simple solution", then yes :-)
@yo' I think you're vastly over-estimating the effectiveness of that solution
(and it would not solve the problem with 20+ years of accumulated files that do use it)
@DavidCarlisle sure. I don't know what the environment looks like on their side so I won't speculate. I just want to know whether the solution needs to be on the side of the environment variables, or whether there is a tex-specific solution
@yo' that's unlikely to help with prints where the class itself calls from the system time.
say, arxiv.org/pdf/1001.0017.pdf, which has no \today in the user tex
@EmilioPisanty doesn't matter, \def\today{blablabla} goes before \documentclass or \LoadClass. Or do I miss something?
@EmilioPisanty well th eenvironment vars do more than tex (they affect the internal id in teh generated pdf or dvi and so produce reproducible output, but you could as @yo' showed redefine \today oor teh \year, \month \day primitives on teh command line
@EmilioPisanty you can redefine \year \month and \day which catches a few more things, but same basic idea
4:29 PM
@DavidCarlisle oh right, you need to do these three as well.





@EmilioPisanty ^^
@DavidCarlisle hmmmm. I don't know whether this will be enough. Ideally the solution should work for all the existing corpus on arXiv, and that's an enormously heterogeneous class
@yo' you don't need to redefine \today` if you have changed the promitives
@DavidCarlisle right
@EmilioPisanty the only access tex has to time are those three registers so if you have the data that says what dates you want for each document you can set them
4:31 PM
@DavidCarlisle ah. that's very interesting.
@DavidCarlisle can I tempt you into suggesting that as a possible solution on the GitHub issue?
@EmilioPisanty this discounts luatex (which can get time from lua) and documents using --shell-escape to access the system (which can do anything) but I'm assuming they don't allow that.
I suspect it'll get the arXiv team to give it a bit more attention ;-)
@DavidCarlisle --shell-escape is almost certainly forbidden there's no way to do --shell-escape at all
I'm not sure about lua
@EmilioPisanty possibly simplest would be if they automatically modified incoming tex to look like @yo' example (using teh submission date as values)
@EmilioPisanty then that document would always give the same date
@DavidCarlisle could be difficult with inputted TeX files. It really depends on at which phase they recognize which TeX file is the main one to proceed. And as their workflow is pretty unique, it's hard to tell which way is even compatible with it.
@DavidCarlisle @yo' thanks for your input =).
4:57 PM
@JosephWright yes, but downvoting a question past -1 when there already are two answers?
@Skillmon The fact that there are answers doesn't imply the question is not to be downvoted, otherwise nobody can win the "reversal" badge :)
5:21 PM
@DavidCarlisle @AlanMunn @0xC0000022L

I found out:

"beschreiben", engl. "describe" comes from the math. technical-language of the 16th Century, that means from Middle High German, and means "construct" as much.
And that from the original meaning: describe "making a curved movement".
In the literary style of the 19th to the 20th century and in the GDR, this language is used.

You can have that in englisch too:
score a line on with a pointed instrument, as in metalworking
@cis Yes, as @DavidCarlisle pointed out, there is a very technical mathematical use of 'describe' which is what the German version means too, but we both agreed that people would not know this use, so using 'draw' would be the most appropriate term. This is not about trendiness, just about making language understandable to your audience.
As I said in my original comment it has that technical meaning, but as most readers, even native English speakers will not recognise that usage, I can't understand why you keep coming back to this.
1 hour later…
6:51 PM
@DavidCarlisle Ich bin ein Erpel
7:07 PM
@AlanMunn @DavidCarlisle He,he...

Plan figure. The barrel circle over the median $s_b = |M_b B|$, which holds the angle $\alpha$, also contains an isosceles triangle $M_b P B$ with the base $|M_b B|$ and the angle $\alpha$ at the point $P$.
The altitude of the base of the isosceles triangle bisects both $|M_b B|$ at $ M_ {s_b}$ and the angle $\alpha$ at the top. \par
The centroid $S$ divides the medians in the ratio $2:1$, with the longer part lying on the side of the corner.
The point $A$ lies on the barrel circle and on a circle $\bigodot(S,\frac23 s_a)$ described by $S$ of radius
@PauloCereda Du bist Abendessen
@cis why use \par ?
@DavidCarlisle oh nein
Because when I use `\\` everybody scolds me.....

BTW: It's just about the wording, but picture could be nice.
@cis 99.999% of documents would use a blank line (\\ would be wrong, \par is the same as a blank line just odd markup).
@DavidCarlisle das Abendessen.
7:12 PM
@cis I would never use "described" there but if you want to ignore advice that's OK, it's your text.
@UlrikeFischer are you suggesting that google translate can be wrong?
@DavidCarlisle never, perhaps the input was wrong.
@UlrikeFischer, @egreg, @DavidCarlisle I wonder what Bruno is going to do with the 'FPU Brexit votes'
@UlrikeFischer surely you are not suggesting that there could be an error in my typing?
@DavidCarlisle In Germany, all professional TeXer say: You may not use '\`, but must use`\ par' ... otherwise there will be trouble .... xD @UlrikeFischer for example....
@JosephWright maybe we will have to depose Frank and have a new strong leader who will sweep aside discord?
7:14 PM
@DavidCarlisle :)
@cis I doubt she said that.
@DavidCarlisle Right, who's blond and annoying ...
@cis certainly a paragraph is better than \\, but if you input it with a \par or an empty line is your choice. I would use \par in code and empty line in a document.
@DavidCarlisle ooh an uprising
@JosephWright ooh
@PauloCereda Do you get the reference? I worry only @DavidCarlisle will
7:23 PM
@JosephWright well 4 out of 8 put B in first place ...
@JosephWright yes :)
@UlrikeFischer :)
@UlrikeFischer Seems that way ...
@PauloCereda Of course, the problem is I'm loth to describe anyone on the team as annoying
@JosephWright I used to be blonde (now as my husband says I'm mouse-blonde)
@JosephWright You could flip a coin. :)
@UlrikeFischer Sure, but clearly you don't meet the second criterion :)
7:31 PM
@PauloCereda @DavidCarlisle "PauloCereda Du bist Abendessen"
Wow! You are German Spokesmen?! I am surprised!

"PauloCereda, Du bist (beim) Abendessen?" (Paulo, are you eating dinner [at this moment]?)

My tip: Say "Abendbrot" - That does not necessarily mean eating bread - it is more classical for dinner.
@cis I don't know German, just a few words to make myself silly in a foreign language. :)
7:50 PM
@PauloCereda Enten sprechen Deutsch
@CarLaTeX Enten landen im Suppentopf.
@cis and they are very good! (Don't tell Paulo)
Translation: Ducks land in the soup pot. (I do not think, that is correct.)
Oct 23 '17 at 8:34, by CarLaTeX
@HaraldHanche-Olsen I ate bigoli al sugo d'anatra (a kind of pasta with duck sauce) yesterday with @egreg... but don't tell Paulo ;)
@CarLaTeX Not bad, he can invite me for dinner.
@DavidCarlisle Mr. Carlisle, I mean, you have a very complicated household in terms for triangles: https://en.wikipedia.org/wiki/Triangle

Many of the words are not self-explanatory.
PS: @CarLaTeX I do not know, how to translate "Begriffshaushalt". :(
8:08 PM
@AlanMunn And yet, in Pennsylvania there is Centre County. (I spent a year there, on sabattical leave.)
8:20 PM
@CarLaTeX I never did tell him, so I'm sure he never found out.
I can tell you something about ducks in Germany: There is no Duck Moo Shu in any Asian restaurant. :( That is an impudence!
8:39 PM
@cis Neither do I
@HaraldHanche-Olsen Indeed, it's a secret :)
@HaraldHanche-Olsen I guess Webster hadn't quite caught on for naming counties at the time.
1 hour later…

« first day (3127 days earlier)      last day (695 days later) »