12:28 AM
@HaraldHanche-Olsen -- I remember one (right-handed) professor who wrote on the blackboard with chalk in his right hand, and used an eraser in his left hand to wipe it immediately. (And there was another professor who ate chalk, or at least got it smeared all over his face by the time the hour was up. He was really a mess the day his wife was delivering a baby.)
@egreg -- aren't the commas pointing the wrong way in that example?

@barbarabeeton Heard of a physics prof at Hopkins who did the erase with the left hand thing, so you had to clarify issues on the spot. So when he said, "It's obvious that...", a student interrupts with, "excuse me, professor, but is that obvious?" The professor, taken aback, looks at the student, then at the blackboard. He stands in concentration for 45 seconds. Then, turning to the students, replies, "Yes, it's obvious," and continues without blinking an eye.

1:10 AM
@StevenB.Segletes Yes, sometimes one has to insist a little bit. ;-)
@StevenB.Segletes Physicist and Nobel laureate Frank Wilczek gave a conference talk, and at the end someone asked a question "It is not obvious to me that ...", and Wilczek replied "Well, it might not be obvious to you." ;-)

@marmot Yes, there's the rub, isn't it? I guess we've proved that "obvious" is not an inertial quantity, but indeed is very frame dependent.

@StevenB.Segletes I always thought of it as as synonym of "I really don't want to, and perhaps even do not know to, explain this." The Greek thought that if body is twice as heavy it will fall with twice the velocity, and it was obvious to them. It was obviously correct until Galileo dared to make an experiment. ;-)

2 hours later…
3:25 AM
Conversion from TeX/LaTeX to HTML is traditionally quite difficult to do well. Is this something that LuaTeX would help significantly with?

@DavidCarlisle In the meantime, I figured out how to do it (for a given token list, replacing an arbitrary number of matched catcode=12 left/right braces with their catcode 1 and 2 equivalents) using some parsing and looping with the listofitems package. But if there is a simple approach, I would still love to know that, too.

3 hours later…
6:12 AM
@barbarabeeton I decided that YH knows better. ;-)

6:44 AM
@StevenB.Segletes if you are adding tokens that you know you'll replace then it might be easier to use something normal like \zbegin....\zend rather than catcode 12 {...} but no difference really. On \scantokens it doesn't expand its argument, so that isn't an issue but the question is if you have full control over the tokens, it will make {} back to normal but if your tokens were defined by say \edef\tmp{\abc\string_xyz} to get a catcode 12 _ then ...

1 hour later…
8:00 AM
One legendary professor here (long since dead) would hold two pieces of chalk in his right hand: One in the normal position, and one between the ring finger and the pinky. The latter was his differentiation chalk: With a quick flick of the wrist, he could turn f into f'.
@barbarabeeton (I seem to have forgotten the back pointer on my previous message.)
@barbarabeeton And there was this poor TA who obsessed over whether or not his fly was closed. So he would check it frequently, while turned towards the blackboard. But at the end of the lecture, there would be chalk dust all over his crotch.
@StevenB.Segletes I suspect that story is told in every university. I certainly heard it in Oslo when I studied there. In that version, the professor left the classroom, then came back ten minutes later with his hair in disarray and said “yes, it is obvious”.

1 hour later…
9:27 AM
@barbarabeeton You're welcome! It's really amazing! My mental image of a pipe organ was much more like the 19th century one. I was susprised when they emphasized "the largest pipe organ in Chicago". I'll definitely remember your advice when I get the chance to visit one :-)

9:47 AM
Are there any L3 variants of \fontcharwd etc.?

10:03 AM
@DavidCarlisle could you reopen tex.stackexchange.com/q/495031/2388?

@UlrikeFischer done (different issue? I thought OP confirmed it was the same)
@UlrikeFischer it got a bit late but I checked in the oberdiek tests with very long log lines last night

@DavidCarlisle it is a problem in one of the tex4ht files, it calls convert and need to changed to "magick convert" on windows, see the comments.
@DavidCarlisle just starting the tests, holding my breath .... ;-)

@UlrikeFischer sure but wasn't that the same issue in the old question? (oh except it's tex4ht rather than standalone where the change is needed I suppose)

10:18 AM
@DavidCarlisle yes it is more or less the same issue, but the place to do the change is different.

10:34 AM
@Skillmon \tex_fontcharwd:D?
@Skillmon Seriously now: searching for fontcharwd in the sources brings up only one (two actually) occurrences in l3names, where it's renamed to \tex_fontcharwd:D, so I guess "no".

@PhelypeOleinik :D ones don't count

@DavidCarlisle But they look so happy :D

@PhelypeOleinik :D

@PhelypeOleinik so I noticed still better than most of @PauloCereda's jokes

10:37 AM
@DavidCarlisle oh

10:47 AM
Is there a lua function to print a human-readable representation of a node list?

@DavidCarlisle I have two failures in the noxetex folder: embedfile-test3 and that has to do with line endings, I need to remember what I did last time to force git not to change to windows line endings. The other is in pdfrender-test3. There your tpf contains a /Producer setting I don't have: /Producer(pdfTeX-1.40.20)

@UlrikeFischer must admit I was a bit unsure why I was getting diffs in the tpf tests, will look tonight, probably can't do much now. Thanks for checking, still only 2 differences means things are getting better
@HaraldHanche-Olsen put it in a box and \showbox ?

@DavidCarlisle the line ending is something I should change. I'm not quite sure about the producer setting. It could have to do something with the newer expl3 files. Can you sent me the log?

@UlrikeFischer we have a call now?

@DavidCarlisle That's the plan

11:00 AM
@PhelypeOleinik question was more whether there is something at the code level, the renamed primitive was clear :) But thanks for looking (I should grep through the sources myself, I guess)

@DavidCarlisle I want to see what the data structure looks like in lua (so I can figure out how to manipulate it there), not the TeX level.

@Skillmon Perhaps something in l3trial or l3experimental; I looked in the kernel only.
@Skillmon The only occurrence outside of l3names: github.com/latex3/latex3/blob/…
I currently have MiKTeX and I'm trying to use ConTeXt. Is there an easy way out or should I give up and install TeXLive?

@PhelypeOleinik I would suggest that you use the standalone instead. (Sorry, I don't have miktex, so I know nothing about ConTeXt in miktex.)

@PhelypeOleinik Use the standalone

@mickep Apparently MiKTeX doesn't support ConTeXt for some reason (neither [eu]ptex, I found out when trying to build expl3). Thanks for the link!
@JosephWright Thanks you too :-)

11:12 AM
@PhelypeOleinik If you want to go with the "really new", then you could try lmtx. It works pretty well here.

@mickep I was just about to ask if that would install MkIV or LTMX. Thanks again :-)

@PhelypeOleinik Welcome to my world :)

@JosephWright Oh my... I asked just for the record, but I already had the TeXLive page open. I didn't really enjoy my experience with MiKTeX...
@JosephWright You have [eu]ptex installed separately like ConTeXt?

@PhelypeOleinik No, I have TL (lots of TL), plus MiKTeX, plus ConTeXt stand-alone

oh, black belt in distributions...

11:24 AM
@JosephWright Oh, wow O_O
@JosephWright I'm in the lots of TL team :-)
Windows news: "MiKTeX Utility stopped working". :(

12:06 PM
@mickep The tempered ones are especially nice to work with.

@HaraldHanche-Olsen ultranice, indeed.

1 hour later…
1:10 PM
@egreg -- Hmmm. I think I'll just ask him directly.

1:40 PM
@DavidCarlisle I initially did the \zbegin...\zend approach and it worked well...unless groups were nested... \zbegin...\zbegin...\zend..zend then, not so good. The approach I ended up with was a recursion that seeks the innermost cat12 faux-group, replace it with a real group, rinse and repeat.

4 hours later…
5:37 PM
@egreg In English BS has a very clear meaning.

@marmot ;-)

6:26 PM
final publisher pdf
you win some and you lose some, I guess

@EmilioPisanty One shouldn't write such complicated equations. ;-)

@marmot yeah, I know, they found this really taxing, apparently
the alignment of the equals sign on eqs (12) and (13) was apparently beyond them
I'm not particularly impressed with the typesetting of the maths, for such a high-brow publisher
particularly given that they print in colour by default but they're still charging us for colour figures

@AlanMunn One can use arabxetex with xelatex. I'll try to add an alternative. In same time i think that the solution given work for arabtex (arabi in fact) she just need to give it a try.

6:41 PM
@touhami That's good. I know she's still around since she dropped into chat last night (US time). And thanks again for helping out.

6:53 PM
So, to ask my earlier question again, would LuaTeX help with LaTeX to HTML conversion? Does anyone know?

@FaheemMitha yes I think context has html output doesn't it?
24

ConTeXt does not directly output XHTML, it outputs XML. However the current browsers (at least Opera, Firefox and Chromium) are able to display XML correctly. The XML can be styled using CSS. When you want real XHTML, you have to transform the XML to XHTML using external tools. ConTeXt standalon...

@AlanMunn never mind.
Don't mention it :-)

7:12 PM
@DavidCarlisle It is more effective than traditional examples like tex4ht? Which I used, but it never worked very well for me. Lots of bugs.

@EmilioPisanty You got your nice tildes! :-D

@egreg some of them, anyways
to my eyes the ones in the displayed equation are waaay too florid
and there's a mismatch between the inline and displayed tildes

@EmilioPisanty That formula is typeset horribly! The hat over G is misplaced

@FaheemMitha I have never used context so by extension i have never used its html/epub generation:-)

I'm not classing this as the best typesetting I've ever seen

7:17 PM
@EmilioPisanty I agree. The space between (for example) subscripts and following letters in formulas look far from optimal.

@egreg what do you make of eqs 12 and 13 in the Methods?

@FaheemMitha however you can't really compare as tex4ht won't understand context files, and you can't give latex filesto context

@EmilioPisanty Never heard about align, I guess. And too much blank space between them.

@egreg yeah. I specifically requested that they align them during proof corrections, but it's not important enough to make a big fuss about it.

@DavidCarlisle Just meaning in a general sense, in terms of the quality of the generated HTML.

7:20 PM
@EmilioPisanty Not to mention the uneven spacing

@egreg ugh. yes.
also, Methods is ragged-right while the main text is justified
no idea why

@FaheemMitha no idea, I have only seen a couple of examples, but quality of generated html usually depends greatly on quality of the tex input so again hard to compare as you can not use the same input for the two systems, in the way you can compare tex4ht/latexml/lwarp etc

@AlanMunn a solution with arabi, arabtex is added.

7:38 PM
@touhami Hopefully she responds to the answer.

7:51 PM
@EmilioPisanty That could be the house style. Equation 21D shows that they can do multiline equations without excessive vertical spacing. Perhaps the problem with 12 and 13 did arise because both lines were numbered? (But why? It does not make a lot of sense to give separate numbers to the two equalities in equations of the form A=B=C.)

@HaraldHanche-Olsen numbering was subdivided in my original
ragged-right methods definitely is house style, not that that makes it better

@EmilioPisanty Well, you couldn't have known that it would cause this vertical spacing issue (if that is what caused it). But you always refer to the parts together, as in (12A) and (12B), so I still think it makes little sense. Even so, they should have been able to deal with it more gracefully.
@egreg But (19) is aligned. Except it almost looks like eqalign? They didn't manage it in (10), but there they would have to use aligned
The inconsistency in alignment between (21D) and (22B) is something to behold.
(Apologies to readers who can't get behind the paywall … you're not missing much, except the science is hopefully good! Not that I am qualified to judge.)
@EmilioPisanty Maybe you need to give up physics and turn to mathematics instead. For the time being, at least, the typesetting is noticeably better.

8:07 PM

@HaraldHanche-Olsen oh, hell no. I do real stuff =P

@EmilioPisanty I couldn't get it to work. Maybe because of all the tracking protection in my default browser setup. Instead of investigating, I just turned on the VPN to get inside my work network.

@EmilioPisanty Real Analysis is something real.

@EmilioPisanty So it seems.
@mickep Until you encounter fat Cantor sets; then it seems rather surreal.

@HaraldHanche-Olsen Haha!

8:36 PM
@HaraldHanche-Olsen Very inconsistent.

@EmilioPisanty My worst experience was when the journal took a LaTeX generated figure of mine, scanned it while tilting and rotating it, and included the scan in the publication. You should consider yourself lucky. (These were just some proceedings but still...)

@marmot yeah, the pit of awful publisher stunts is deep, I know

@marmot There was a journal in linguistics that would cut and paste stuff as images instead of actually typesetting them.
The nice part of this story though is that the entire editorial board and the editor resigned over Elsevier's position on open access and started a new open access journal which is now one of the major journals in the field, while the other one is publishing crap, since nobody will review for it.

8:51 PM
I just got a few hours a message from a very well known journal saying:

The references have been fixed.

... after I kindly reminded them to run bibtex before sending the stuff to the referees... ;-)
@AlanMunn @EmilioPisanty ^^^^

9:10 PM
@JosephWright I get a global/local error here when compiling with xelatex:
\documentclass{book}
\usepackage[enable-debug,check-declarations]{expl3}
\begin{document}
blub
\ExplSyntaxOn
\__pdf_backend_object_new:nn {xxxx}{dict}
\ExplSyntaxOff

\end{document}

@UlrikeFischer Looking

\int_const:cn
{ g__pdf_backend_object_ \tl_to_str:n {#1} _int } this should be c__
@JosephWright ^^

9:48 PM
@JosephWright works ;-). Do you think it would be possible (and if yes worth it) to extend the tests to dvips/ps2pdf?

@UlrikeFischer That is tricky: much of the PDF stuff we only test in the sense the token lists are unchanged

@JosephWright I'm mostly interested to check if the pdf is still the same. I could e.g. add a batch file to do latex+dvips +ps2pdf -dCompressStreams=false -dCompressPages=false, but I don't know if the system could use it as "engine".

@UlrikeFischer We can make an appropriate Lua function ...

10:12 PM
@JosephWright it could be useful, dvips is so special, without any testfiles it is difficult to handle.