@PauloCereda The man sitting beside me on the bus has "qua qua qua" as ringtone of his smartphone!

Some one here who can answer a question about typesetting a statistic tabular?

@CarLaTeX My husband has currently a "quack"

@UlrikeFischer Oooh

@AlanMunn He uses this, but just as a quick \bigskip for restructuring the document. Still I called him a barbarian for this of course.

@yo' Do you know if one should write $N=10$ or $\mathrm{N}= 10 $ (in statistic)?

@UlrikeFischer Shouldn't variables representing scalar quantities be always set in italics?

@AlexG "always" is a large word. The question is if statistic author follow this rule or not.

@UlrikeFischer As a Math discipline, the same rules should apply (I think).

i think authors are trying to distinguish between vectors, in bold, and scalars.

@s.patroller This is what I think as well.

There would be no need if there were no vectors around, right?

or even matrices

Perhaps, stat.se is the place to ask :-)

@UlrikeFischer for which N? For the number of variables? Certainly mathnormal, not mathrm

@s.patroller this is not quite the case, it's a large oversimplification and more the physics point of view

@UlrikeFischer if it were typeset roman, a variable could be mixed up with an operator (sin, cos, erf, ..) by the reader

@AlexG I know this. But since when have typesetting traditions been logical and consistent? I want to know what the field does, not what logic says that the field should do ;-)

@UlrikeFischer If statisticians do it differently, they should be educated to to it the correct way :-) .

@UlrikeFischer It should certainly be $N=10$, also for statisticians.

@UlrikeFischer this.

@CarLaTeX ooh

@UlrikeFischer :)

I'm currently eating lunch in my office a student just came by with a message that his LaTeX (or what ever) says that the log file is "too large, do you want to proceed". I told him to come back at 1pm, such that I can eat my lunch. Anyone ever heard of log files being a problem? I haven't looked into it yet.

@daleif well you can produce very large log-files if you use e.g. \tracingall. And it could be quite possible that an editor like texstudio which analyses the log-file gives up in such cases.

@UlrikeFischer that was what I would suspect as well, that this is not a latex error, but from an editor or similar.

@daleif I never assume that people know the difference between latex and the editor ;-)

@UlrikeFischer precisely. I often start debugging a student problem by eliminating the editor, often they have strange build settings and does not know what they do.

@UlrikeFischer What is "T-bar" supposed to be?

@PauloCereda Ohh, that one could end up at my next course web page! :)

@mickep :D

@yo' No idea. I thought that it perhaps should be $\bar{T}$ (which I wouldn't know either) but I asked back ...

@UlrikeFischer ^^ :)

@PauloCereda and buy him a scarf for cold weather

@UlrikeFischer ooh I could find a duck-themed scarf!

@PauloCereda You could knit one ;-)

@UlrikeFischer ooh :)

@UlrikeFischer $\bar{T}$ is certainly common in statistics

@yo' That was my guess ;-).

@UlrikeFischer people just don't know how to do it in W@#$, right? :D

@AlanMunn So, I showed the friend this chat, and he asked me straight: "How did he know?!" :D

@yo' No the source is a latex document, but it shows some missing knowledge about how to typeset math ;-) (I know now 6 different ways to type $R^2$).

@UlrikeFischer :D I understand the source is LaTeX, but I mean that people often believe that something is correct typography because it's how you do it in Word.

@yo' lol

@UlrikeFischer It turned out he had a \left at the very beginning of his file and texstudio never seems to report it as an error. Easy fix. Followed by a 90min consultation of why his latex was of poor quality. It's been a while since I've seen a$_2$=1.23

@daleif Oh, so you leant him that he should use \raisebox instead of a$_2$, right? :)

@mickep He was very understanding and went home with a lot of homework

github.com/YuanshengZhao/Garamond-Math

@egreg, @UlrikeFischer ^^

Wow. A typo in a package option (hratio=1.:1, the period should not have been there) led to a Missing \begin{document} complaint while loading the geometry package. That one might have a beginner stumped, for sure. Good thing I'm not a LaTeX newbie.

@PauloCereda Do you know the Griffon framework?

@PauloCereda Wow!

@TeXnician Yes. :) At least, from name and a couple of tries. I had an outdated material and it looked very good. Sadly there are a lot of new things that I missed.

@egreg :)

@PauloCereda Well, I've never used it, just thought that it might be useful for ArTeXmis. It even makes it easy to create a command-line interface using Laterna ;)

@TeXnician Yes, Lanterna looks cool. :)

@TeXnician If I can find a newer documentation, I might dive into it, it's very interesting. I thought of using Pivot, but it's insanely slow, even for Java standards.

@PauloCereda But that's nothing we should focus on when starting our development. Else only vim users will migrate :D

@PauloCereda I would favor using JavaFX (it's really flexible and not too slow) and maybe we could back it by Griffon, so that we could interchange it with something else (like Pivot) if needed.

@TeXnician LOL

@TeXnician Do you have any experience working with JavaFX in non-Windows environments? That's my main concern.

@PauloCereda Yes, I've (up till now) only developed on Linux and just run the jar files on Windows. I'm pretty satisfied with the results nowadays.

@PauloCereda One should only avoid heavy graphics components on ARM architecture. That doesn't work out well (believe me, my Pi had to digest many applications).

@TeXnician Interesting. Is the Oracle FX thingy or the OpenJDK one?

@TeXnician Agreed!

@PauloCereda OpenJDK. I've abandoned the Oracle SDK since the release of OpenJDK8. And I've never looked back.

@TeXnician Great to hear, now I am really interested. I am using OpenJDK since version 6.

@PauloCereda As even Debian stable contains OpenJDK 8 nowadays (which in turn opens up for a good openjfx) it should be possible to run this even on "not so recent" machines.

@PauloCereda Well, I'm using OpenJDK since version 6 too, but OpenJDK 8 brought the most things which one would call "killer features" to switch.

@TeXnician Yes, agreed. :)

@TeXnician: do you have, by chance, any pointers about FX?

I am really a newbie.

@PauloCereda Yes, but later, have to take some time off now.

Just got reminded (again) why one should use \lVert…\rVert and not \|…\|: $\|\exp(A)\|$ produces an unwanted thin space before the exponential. (Sorry if mumbling to myself on the chat bothers ya.)

@TeXnician Thanks. :)

@HaraldHanche-Olsen Oh, don't excuse yourself. I did not know that, thank you.

@mickep :) The reason is that \| makes an ordinary math atom, whereas \lVert make an “open” math atom (like a left parenthesis). Operators like \expget a space in front of them when there is an ordinary item in front, but not if there is mathopen there.

@HaraldHanche-Olsen Thanks for the explanation. I simply have to relearn. In fact, right now I mostly write in ConTeXt, and then I guess (hope) one could use the fences system.

@HaraldHanche-Olsen Would it be possible to redefine the \| and \| to work "as expected" also in the situation you describe above?

@samcarter Check the "Background" explanation in this question tex.stackexchange.com/q/427208/2388

@UlrikeFischer I like that @marmot refers to page 1060 of the tikzduck manual in a comment.

@mickep where?

@UlrikeFischer Comment to the question: "In the example of the manual on page 1060 on the bottom it is explained how to add parameters. And I think you are missing a number of \pgfpathclose commands after the squares." The emphasize is mine.

(Oh my, my sense of "humor" is so bad)

@UlrikeFischer Oh, I'm wondering what chequerboard-coloured item the ducks will get! Thanks for the ping to this question!

@mickep Probably not in any robust manner. Instead, I'd suggest loading mathtools and then say \DeclarePairedDelimiter\norm\lVert\rVert, so you can now type \norm{\exp(A)} instead. This approach brings multiple benefits.

@HaraldHanche-Olsen Oh, so the mathtools package has support for such things. Excellent!

@UlrikeFischer These thanks belong to you: tex.stackexchange.com/questions/427285/…

@PauloCereda You should have got mail.

@PauloCereda If you are also ready for some further technical ideas for ArTeXmis just let me know.

Yes, I am ahead of my time, and already know what will be written on page 1060 of the future version of the tikzducks manual. I can already tell you guys that this will be part of the introduction then. ;-)

@marmot You probably meant Introduc(k)tion :)

@samcarter Yes, you're right. I guess you'll be very busy typing all the pages, but I can't wait seeing all the additional ducks. ;-)

@marmot As the original philosophy of the tikzducks once had been to build everything as a combination of basic geometric shapes (that was before all these complicate paths for hair styles etc. were added), maybe a brute force approach could do: generate many random combination and keep all which resemble something....

@samcarter I guess using combinatorics you can draw an arbitrarily large number of ducks. You have just to be careful not to exceed the number of atoms in the universe, because then it might be hard to print the manual out. ;-)

@marmot \prg_replicate:nn { 2↑↑↑↑2 } { \duck }

@egreg No, they would all look the same, right? I'm more thinking of @UlrikeFischer's chess board, it is known that the number of possible constellations of chess figures on a board is huge, and now these chess ducks could have additional features like glasses and so on. (Really looking forward to seeing this manual. It will use up 99% of the future TeXLive installation ;-)

@marmot How can you know? :-)

@egreg Hmmh, only certain Englishmen would make replicate not make replicate things, and you are not English. ;-)

@santimirandarp M$_\mathrm{w}$ or M\textsubscript{w}. But why do you want the M not in math mode? I'd do $M_\mathrm{w}$.