@CarLaTeX @samcarter another duck ;-)

@marmot Yay!

@CarLaTeX OMG, when do you sleep?

@CarLaTeX Did Italy move to the east? I see also @egreg seems to be awake. (Please avoid getting too close to the Chinese border. ;-)

@marmot I just wake up... to go to the toilet, lol! And I'm just going to sleep again. I think @egreg wants to reach its rep cap before breakfast.

@CarLaTeX maybe brunch ... ;-)

@UlrikeFischer, @samcarter: Netflix is about to unveil a reboot of Carmen Sandiego. I think we need all the accessories for a Carmen duck. :)

@DavidCarlisle @JosephWright I just got answer from ctan, that the description field came through (but naturally without the newlines/paragraph). Did you try if explicit <p>...</p> in the string works?

@JosephWright Could you please clean up the comments under my answer, starting downwards from this one? tex.stackexchange.com/questions/468466/…

@HenriMenke Done

@UlrikeFischer Haven't tried that

@JosephWright Thanks

@CarLaTeX the Bär is not a mascot! He is a very valuable member of my team - even if he is a bit mad and currently refuse to accept that christmas is over and to put down his cap.

@CarLaTeX as a proof that we are not biased against other animals you could mention the new video.

@JosephWright I will do it next time (I want to upload a new luaotfload anyway).

@UlrikeFischer Would be good: if we can do a simple search-and-replace in the script it should be easy

@UlrikeFischer yes

@UlrikeFischer Please edit my answer. I'm not at home at the moment and I can't do it now, thank you!

@DavidCarlisle as I said, I didn't read it all. But I think in VIM it is easier to read code than in Emacs, as browsing through is easier, you don't have to hit like 5 key sequences to jump one line down.

@marmot regarding the question with the many comments we deleted: He doesn't need more vectors, the 3 he shows are the one transformed. But calculating the variance of such a small data set seems odd. You basically get the PCA with the eigenvectors and eigenvalues of the covariance matrix of your data set.

@AndréC As @HaraldHanche-Olsen says, please flag this stuff if you have concerns. That builds up a history that the moderators (including SO staff) can see.

@JosephWright And do not flag stuff in chat, otherwise a zillion of people will come here out of nowhere? :)

@PauloCereda Well in this case it would be appropriate, I think: the concern is a chat message which does need to be logged

@JosephWright ah understood. :)

@PauloCereda You haven't voted for my latest answer :(

@DavidCarlisle link?

@PauloCereda latest question on meta, you'll find it:-)

@DavidCarlisle ah

@DavidCarlisle you are... oh

@DavidCarlisle I voted it.

@Skillmon :-)

@JosephWright How can a description of events be inappropriate? Just want to understand the logic behind this.

@marmot The point was a general one: if anyone feels there are issues with other users, the appropriate thing to do is flag stuff.

@JosephWright Fine. Then I want to flag "Can you stop these recurring attacks against me?" because I did not attack this person. How can I do that?

@marmot You can flag chat messages: I'm going to clean some of them up

@marmot Flag any and every single comment or message that's bad. Sometimes, I see a single flag about reasonable complaints about a lot of stuff that's said to have happened before, but I simply cannot read through hundreds of comments of others to evaluate which ones of all are bad.

@JosephWright @StefanKottwitz Thanks!

@UlrikeFischer Maybe the Bär has some Russian relatives and wants to celebrate again on 7 January?

@PauloCereda Which accessories do you have in mind?

@marmot duck inflation :) Soon you'll have the only non-duck avatar

@samcarter Nope ;-) This is not a virus causing The Walking Dead Ducks here ;-) Undead Ducks...

@samcarter Yes, I guess @DavidCarlisle will flip very soon. ;-)

@ChristianHupfer Oh, undead ducks! There is already a mummy duck... maybe a zombie duck could join

@samcarter My comment was not a suggestion for adding new ducks ;-)

@marmot mmm, his recent success in avatar-related contests may make him wait longer

@ChristianHupfer :) suggesting new ducks is not necessary, they appear by themselves

@marmot with enough honey liquor, inventing blue seems easy

@samcarter Looks like that the pandemia is worse than I've thought ;-)

@ChristianHupfer and there is no suitable medicine

@samcarter Yes, sure. But now the honey liquor is gone :-(

@marmot oh no!

@samcarter Yes, and still almost a month to go before Groundhog Day. (When the groundhogs are with the humans, it is usually the opportunity to borrow some of their honey liquor.)

@samcarter ....

@ChristianHupfer I'm afraid you have a bad copy of the real painting

@samcarter Hm, I wonder about the Munch paintings I saw last weekend -- there was no owl on it. Because he did not know anything about TikZ...

@marmot While Phil distracts all the humans the other groundhogs get the honey liquor?

@samcarter No, we get the honey liquor from the groundhogs. ;-)

@ChristianHupfer The Munch paintings with owls are much more expensive -- not many museums can afford them

@marmot First the groundhogs have to get it from the humans, don't they?

@samcarter I suggest to glue some owls on any Munch painting you see in all museums around the world then, in order to raise the value of the paintings ;-) And then proceed to Van Gogh, Renoir, Rembrandt, Monet...

@ChristianHupfer Adding owls to Monet paintings is ridiculous, they are obviously missing frogs on the water lillies!

My texlive updated lipsum today, now I get an "! LaTeX3 Error: File 'lipsum.ltd' not found." with a simple document

Is that me (bad update) or can anyone recreate this?

@samcarter Well, that's their problem. Bears, humans, doesn't matter where they get it from, as long as it is good. The one we get has the name "Bärenfang", so most likely it is made by bears. ;-)

@MartinScharrer already reported on texlive list (it's a packaging error the file got put in the doc tree)

@DavidCarlisle Ah, thanks

Does someone know what unicode-math does with a .fontspec file? The log says the file is read but the UprightFont key is ignored and hence it tries to load FontName.ext instead of FontName-Regular.ext…

@TeXnician sorry but this rather unclear. unicode-math is about math fonts, UprightFont is a textfont option. Better make an example.

@UlrikeFischer Okay, so using Fira Math I want to use a .fontspec file to load this as math font (the reason for that is some path trickery). With the following files (and Fira Math located in a fonts/fira-math subdirectory) I get an error message about FiraMath.otf not found although I want it to load FiraMath-Regular.otf.

%content of FiraMath.fontspec
%\defaultfontfeatures[Fira Math]{
%  Path           = ./fonts/fira-math/,
%  Extension      = .otf ,
%  UprightFont    = *-Regular,
%  BoldFont       = *-Bold,
%  FontFace       = {ub}{n}{*-Heavy},
%  FontFace       = {eb}{n}{*-ExtraBold},
%  FontFace       = {sb}{n}{*-SemiBold},
%  FontFace       = {m}{n}{*-Medium},
%  FontFace       = {sl}{n}{*-Light},
%  FontFace       = {l}{n}{*-ExtraLight},
%  FontFace       = {el}{n}{*-UltraLight},
%  FontFace       = {ul}{n}{*-Thin},

(./FiraMath.fontspec)
luaotfload | db : Reload initiated (formats: otf,ttf,ttc); reason: "File not found: ./fonts/fira-math/FiraMath.otf.".

! Package fontspec Error: The font "FiraMath" cannot be found.

For immediate help type H <return>.
 ...

l.20 \begin
         {document}
?

Has something changed recently with hyperref? My links no longer show up coloured, in documents that as recently as last August did have coloured links.

@TeXnician I'm not quite sure what this should do. E.g. what do you expect the * to expand too? There is imho no FiraMath-Heavy.otf, and UprightFont doesn't really make sense with math fonts. There was a similar question some time ago (where I didn't understand the sense either): tex.stackexchange.com/questions/437855/…

@AlanMunn well there was an update, and if it broke something @DavidCarlisle will probably blame me ;-). Can you make an example?

@UlrikeFischer There is a FiraMath-Heavy (it's the current 0.4-beta). I would expect * to expand to FiraMath as it would do for regular fonts and indeed it does (if one removes the font files from the given path it complains about FiraMath-Regular.otf not found). My question is only whether there is a way to load a math font with a .fontspec file. As it seems the fontspec file will be loaded but unicode-math will look for the wrong font, so the answer seems to be "it is not possible".

And yes, I have seen that question as well but since it did not receive an answer I just wanted to ask whether somebody would know if it was possible (intended use) at all…

@TeXnician well your font call doesn't make sense in math. You wouldn't use UprightFont or FontFace together with \setmathfont. So what do you expect unicode-math to do with it?

@UlrikeFischer Well, I would have expected unicode-math to load the regular font and the other fonts like the bold font for something like range={bfup}. Maybe I did not get the concept of math fonts right and have to read up a bit.

@TeXnician And where you would heavy and extralight fit in your idea?

@UlrikeFischer It turns out I had draft mode as a documentclass option because I was trying to track down some overfull boxes. I hadn't ever noticed that it turns off the link colours.

@UlrikeFischer Probably nowhere. But as I am no expert in math fonts in TeX I just tried to apply some knowledge about text fonts. Obviously, that does not work. Hence, I guess that .fontspec files are simply an unsupported input for unicode-math because their keys do not make any sense for math.

@AlanMunn it's @UlrikeFischer's fault.

@UlrikeFischer can I blame you for xr as well?

@DavidCarlisle too late, you already took it: chat.stackexchange.com/transcript/message/48367256#48367256 ;-)

@marmot Just the machine doing its nightly chores. Today I was at a basketball tournament.

@egreg Did you win? (and does the machine also write answers? ;-)

@marmot I've not been a player for a while. ;-) I coach young referees.

All the 10-year-old boys after the final.

@egreg replace that net thing by stumps and you could coach cricket

@egreg I see. Yes, if ducks can play, also marmots can. And there is a duck on the field. (We are playing here twice a week, and the oldest player retired just before he got 80...)

@marmot Much younger than me, according to @DavidCarlisle

@egreg You see. The guy who retired is a retired professor of computer science...

@marmot He retired very young. ;-)

@egreg I guess he retired in his mid 70ies... (in the US it is not compulsory to retire, Nobel laureate Weinberg is famous for his statement : "I'll retire one day after my death." ...)

@samcarter Actually, just a red clothing. :)

Got this from the American Amazon:

@AlanMunn ^^ :)

A penny earned is a penny saved, so they say!

DAS VAMPIR

@egreg Is that you in the duck suit? :)

@PauloCereda Good news then: just for you I added everything necessary tshirt=red :)

@PauloCereda The reaction by CNN is immediate: Amazon under threat.

@UlrikeFischer Corrected

@UlrikeFischer Added link to the video

@CarLaTeX good ;-) (I was away during the day too).

@UlrikeFischer Don't worry :)

@GodMustBeCrazy Why've you posted there here?

@DavidCarlisle do you know about some incompability between bm and mtpro2 complete which could give such an error:

Runaway argument?
\s@@b j
! Forbidden control sequence found while scanning use of \next.
<inserted text>
                \par
l.2 $\bm {S^{i,t+1}_j}

@JosephWright For getting quicker response (maybe).

@GodMustBeCrazy Er, you know this is the TeX chat room?

@JosephWright: Yes. But others sometimes also do the same thing, i.e., posting irrelevant contents. :-)

@egreg ^^^^ Your avatar has actually a nice reflection:

@egreg I think you could easily beat @DavidCarlisle in this competition

@AlanMunn That's @PauloCereda, of course!

@Skillmon Yes. I was hoping it is only the projection part in the Wikipedia article. And this really only amounts to invert a matrix and use it as the basis vectors. As I said, you'd need then to have further vectors that get subjected to these transformations.

@PauloCereda I am wondering if your article "Formatting posts" could be used to answer this question. That is, could you perhaps provide the user a link?

@marmot Prof. van Duck is not @PauloCereda

@CarLaTeX Oh, I see. Scusi.

@marmot Don't worry, it's easy to confuse the two: they are both ducks and have (almost) the same first name :)

@UlrikeFischer ooh my \outer safety net got used.

@DavidCarlisle Oh goody

@DavidCarlisle ??

@CarLaTeX Yup, that's what made me think it's him. (Nice article! Have you ever figured out how to post a code directly after a list on SE, I mean without any ordinary text inbetween?)

@UlrikeFischer bm inserts an end token that's a macro holding an\outer token so that if it gets a runaway condition it stops somewhere near the end of the original argument rather than eating the rest of the document (as happened in some pre-release tests)

@marmot Yes, just put <> after the list, it's written also in the article

@UlrikeFischer although looking again it looks like the par rather than the outer end bm token, that's a bit odd. example input?

@marmot indent it twice, or use a ---- horizontal rule

@CarLaTeX Grazie! (couldn't find it...) (BTW, on a Mac the shortcuts do not come with ctrl but with the command key.

@DavidCarlisle Thanks! (Indenting twice obviously is not achieved by pressing the {} button twice, right? So you do it by hand?)

@marmot Sorry, I only have Windows...

@marmot yes (I think I remember doing that once) but haven't tried recently

@CarLaTeX Yes. I guess Mac users will figure that they have to use another key instead of ctrl.)

@marmot Prof. van Duck will put an addendum in the next article

@DavidCarlisle the equation is simply as in the message: $\bm {S^{i,t+1}_j}$. The whole document uses some complicated class and mtpro2 complete - which I don't have so I can't test the document myself. It is a guess that the redefinition of _ in mtpro2 is the problem. If you never saw the error there is imho not much you can do - I will have to ask to get the fonts for a test and more debugging (or they change the code of the equation ...).

@marmot We're used to Windoze hegemony. :)

@DavidCarlisle OK, it is as efficient as my "And here comes the code:" ;-)

@UlrikeFischer I should be able to test even without the fonts (just lie and make it load something else)

@AlanMunn Well, those who use the key shortcuts are usually not windows users, who want to do everything with mouse clicks... (BTW, has your university also switched to canvas? If not: Lucky you!)

@marmot @CarLaTeX But in the code editor on the site at least in Firefox, control-K works as well as Command-K to format code.

@CarLaTeX Grazie!

@marmot ah, ok. But I understand the question differently (meaning to calculate the whole PCA on some data set).

@AlanMunn Indeed there is only ctrl here: tex.stackexchange.com/help/formatting

@CarLaTeX @AlanMunn is right, ctrl also works.

@marmot We use D2L. It's not great, but not awful either. But I find all this adaptive site code so awful. It uses way too much real estate, blank space etc. Theoretically this is good for accessibility reasons, but it would be nice to be able to control the view.

@marmot Phew!

@CarLaTeX For me control-k kills a line, so many times when I want to delete something I don't get the right effect. :)

@AlanMunn We switched to Canvas. Everything is by clicking. Did you see Will Robertson's talk in Rio. He seems to have "hacked" it but our IT does not allow me to use these tricks. It's a mess.

@AlanMunn Ops

@Skillmon Yes. But there is no data set, right?

@marmot Yes, I saw his talk. It was very interesting. I guess D2L also works only by clicking too. But I've just given up on worrying about that. All of these systems take a lowest common denominator approach to the UI, and having seen some of my colleagues in action I understand why.

@AlanMunn Well, I do not propose to take the click click click approach away but it would help a lot if one was not forced to use it because it is the only possible way.

@marmot But I use the absolute bare minimum for my courses. I post problem sets and handouts and use the submission system for homework, but not much else.

@marmot or a very small one...

@marmot Sure, I agree, but that adds a level of complexity to the system and also to the support for the system. If you've ever tried walking someone through simple command line stuff who doesn't know anything, you'll know what I mean. And the fact is that once you open an 'expert' interface option, you can't just tell non-experts to go away when they try to use it.

@marmot Just saying that I understand the rationale of their decisions.

@marmot That's weird for sure. Our grades aren't uploaded through D2L (although I think they can be). We have a separate instructors' system that lets you download CVS of the classlist and then upload it back with grades filled in.

@Skillmon It ultimately boils down to the statement that LaTeX is not a computer algebra system, so I agree with all you're saying. Just tried to do something potentially useful, namely inverting a general transformation in 3d, which may not have been done. I am certainly not an expert on machine learning, but the little bit I know just reiterates that your statement that one may not use LaTeX for that.

@AlanMunn We had another system before, EEE, which is also bad, but not as terrible as canvas. The old system was at least a bit intuitive. The new one is really really terrible.

@UlrikeFischer I got mtpro2.sty from ctan changed all n fonts to cmsy10 and all b onesto cmbsy10 but left macros untouched and this then runs without error

\documentclass{article}
\usepackage[complete]{mtpro2}
\usepackage{bm}
\begin{document}

$\bm {S^{i,t+1}_j}$

\end{document}

@marmot what you posted is very interesting, though :) I didn't know this was possible with (relatively) little effort.

@DavidCarlisle yes, I just changed enough in the file and with book alone it compiles, so it is something in the rest of the code. I will try to minimize it ...

@Skillmon I partly reinvented the wheel as the matrix inversion has already been done before, but OK now you can do it with TikZ. I am always hoping that this code will become part of a package but it does not seem to happen soon, have not seen the user in a while. It does much more sophisticated things.

@JosephWright what's changed with current encoding?

@DavidCarlisle Huh?

@marmot sometimes I wonder whether you TikZ wizards have any spare time left not spend on TikZ.

@DavidCarlisle I mean that passing accented chars through \usepackage is a risk I wouldn't take: everything gets expanded

@Skillmon Yes, of course. (If you know a tiny bit of TikZ syntax and elementary geometry, you can answer all these questions very quickly. I think that strictly speaking these are not LaTeX questions.)

@JosephWright ah I misunderstood the question, I thought he meant it had been working and a kernel change had broken it, but you mean it never worked

@DavidCarlisle I don't think it ever worked

@marmot I know a tiny bit of TikZ syntax, maybe my bit is still too small.

@JosephWright it's almost as if the people who programmed that feature never used accents...

@DavidCarlisle :)

@Skillmon Everything is relative. Small meaning relative to over 1000 pages of pgfmanual. You need only 20% of these to be able to understand most things.

@DavidCarlisle They just had accents.

@AlanMunn Yes, a strong British accent, I suppose.

@marmot Actually a mixture of British, American, Dutch and German I think.

@AlanMunn I was referring to those responsible for the bugs. ;-) (Just kidding.)

@marmot just German then.

@DavidCarlisle When in doubt, blame Frank.

@DavidCarlisle Germans don't have an accent, do they?

@AlanMunn we didn't have @JosephWright and @UlrikeFischer in those days.

@marmot just a weird beta thing pretending to be ss

@DavidCarlisle Und ¨.

@DavidCarlisle Comments removed, sorry I forgot the answer was yours, not @Skillmon's

@DavidCarlisle Well, it's not my fault that they started modifying the beautiful Bavarian language.

@CarLaTeX thanks

@marmot aren't umlauts accented characters?

@Skillmon Who uses umlaute?

@marmot Bavarians do

@Skillmon That's what people think. (München is only call München by foreigners.)

@marmot how do you say Schüssel or Schlüssel?

@Skillmon These are oldfashioned objects. we us retina scanners, of course.

@Skillmon Arguing with marmots is hard work.

@marmot how do you mix ingredients in a retina scanner?

@Skillmon I don't.

@marmot :)

@marmot qed?

@marmot How do you be polite without 'möchten'?

@Skillmon doesn't have any umlaute....

@AlanMunn Bavarian and polite? Come on!

@AlanMunn BTW, möchtest is really magst.

@marmot So you've abandoned Sie? At least that's in keeping with the lack of politeness.

@marmot but Bavarians don't say "magst", they say "m<oa>gst" which should be some umlaut of a...

@AlanMunn That's indeed true. Locals don't have "Sie"

@marmot Really? Even in formal contexts?

@Skillmon moag doesn't have an umlaut.

@marmot Funny to know there're jokes between North and South also in Germany

@AlanMunn There is eahna and so on for that, but "Sie" is really not used.

@AlanMunn nowadays you don't hear "Sie" very often. Everyone tries to act like they are your friend or something. Feels really odd to me (and I consider myself as young)

@CarLaTeX Bavaria and Germany, please!

@marmot what about Franken?

@marmot Like Padania and Italy :)

@Skillmon Well ...

@CarLaTeX Maybe ...

@CarLaTeX and give us back Süd Tirol!

@Skillmon Interesting. Not very surprising I think, but I hadn't realized that it was disappearing.

@Skillmon @marmot doesn't want it back because it has an umlaut.

@Skillmon I think South Tyrolean would be very happy to go back to you :)

@AlanMunn Actually it is a bit like the opposite. If people think you're local, they do not use "Sie". It is almost like an insult if they do.

@AlanMunn I wonder how @marmot refers to the direction diametral of north...

@Skillmon Unten

@marmot maybe in Bavaria, but not in the rest of Germany (especially in restaurants, shops, etc.)

@Skillmon Certainly that's my experience as a foreigner.

@Skillmon No, it does happen in particular in shops and in restaurants.

@marmot that you're called "Sie"? Not in my experience (being a Prussian (not really, but speaking in Bavarian terms) in Franconia).

@marmot so slow with align* :-)

@DavidCarlisle because these guys are distracting me...

@Skillmon In remote places people do not use "Sie" if they think you're local. (Of course, if you go to lower Bavaria and they think you're local, they speak as usual and you do not understand a single word.)

@marmot chat.stackexchange.com/…

@DavidCarlisle OK, the user is all yours. ;-)

@marmot I manage to understand Franconian after 6 years of living here, maybe I could learn to understand those Bavarians as well, but I don't think I want to :)

@Skillmon I never understood them... (I think it is really a different dialect/language.)

@marmot it is a different dialect, but since I never learned a dialect as a child, it is the only German dialect I know well enough to make others believe I'm Franconian myself. Maybe I could learn Bavarian as well, but it would need some time and me living there...

@Skillmon You need to be fast, it's dying out. (I am not very good at it either, but I can understand people.)

@DavidCarlisle See? ;-)

@DavidCarlisle I pulled it down to this. The error message changes depending on if tcolorbox is loaded or not. With it I get the already mentioned message, without it ! Undefined control sequence._->\futurelet \next

\documentclass{book}
 \RequirePackage
   [
    complete,
    subscriptcorrection,
   ]{mtpro2}
\usepackage{bm}

%\RequirePackage{tcolorbox}


\begin{document}
$\bm {S^{i,t+1}_j}$
\end{document}

@DavidCarlisle One has to select the posts where one is slower, can be an advantage. ;-)

@marmot @Skillmon Don't worry, Milanese don't understand the Bergamo dialect (50 Km distant)

@marmot I am leaving that OP to others...

@DavidCarlisle See how I scooped you. I did this much earlier than you. ;-)

@UlrikeFischer yep I get same now. bm makes @ifnextchar safe-ish but that sems to use \futurelet directly, which is a pain:-)

@DavidCarlisle well I guess I will have to tell them to use \disablesubscriptcorrection for this part of the document.

@UlrikeFischer give me a sec..

@UlrikeFischer try with these two lines added:

$ diff -c bm.sty~ bm.sty
*** bm.sty~     2019-01-05 23:07:56.639099200 +0000
--- bm.sty      2019-01-05 23:10:41.412371000 +0000
***************
*** 41,46 ****
--- 41,47 ----
  \NeedsTeXFormat{LaTeX2e}
  \ProvidesPackage{bm}
            [2017/01/16 v1.2c Bold Symbol Support (DPC/FMi)]
+ {\catcode`\_=13 \gdef\zzzunderscore{\let_\relax}}
  \def\bm#1#2{%
    \let\bm@pmb\install@mathalphabet
    \let\bm@pmb@\getanddefine@fonts
***************
*** 166,171 ****
--- 167,173 ----
      \let\mskip\mkern

@Skillmon thanks

@DavidCarlisle what for?

@Skillmon comment removal

@DavidCarlisle no problem, you're welcome.

@DavidCarlisle looks ok ("look" is the wrong word, as I exchanged all fonts the output is rather horrible ;-), but I get no error). But shouldn't it be \bm@zzzunderscore? Or at least \zz@zzzunderscore ;-).

@UlrikeFischer yes or actually I could locally make _ active while defining \def\bm@genera (it already set up to have active '0 then I wouldn't need a macro at all just inline \let_\relax ..

@UlrikeFischer vvv

@DavidCarlisle that's fine too.

@UlrikeFischer If you'd told me a few hours earlier I could have slipped it in to today's tools release:-)

@DavidCarlisle sorry ;-). I got the question only today and pulling down a 3000 line style to a minimal example takes a bit time.

@UlrikeFischer at least they didn't ask you to clone an entire tool source repository and run the documentation make script:-)

@DavidCarlisle ;-).

@JosephWright is there something wrong with expl3 in texlive? tex.stackexchange.com/questions/468763/…

@UlrikeFischer I've no idea why it's not updated ...

@JosephWright if it isn't in tomorrow 's update I can ask on the list.