@samcarter By the way, happy rubber duck day!

I would say: Open the pdf in inkscape and see, but inkscape does strange things sometimes.

@cis If you use luatex I assume it is determined by one of the font parameters

@mickep Oh, nono - I use nothing special, just pdflatex.

@CarLaTeX Are you in any way related to @Paulo Cereda (business, friends or blood)? Since you share this passion for the yellow tub duck?

@cis well, LaTeX is not more special than pdftex. But for pdftex I don't know.

@MaestroGlanz the one here who likes ducks most is @DavidCarlisle

@MaestroGlanz Of course I am github.com/TikZlings/Extravaganza2024

@MaestroGlanz it's a local meme.

@CarLaTeX Happy rubber duck day to you, too!

@cis \fontdimen8 of \textfont3 (defaut rule thickness) see tex.stackexchange.com/questions/88991/…

@mickep you've forgotten all your classic fontdimen? You are so modern:-)

@DavidCarlisle "For text fonts there are 7 parameters", continues to reference \fontdimen8 :P

@Skillmon but it's a math thing as no one would underline typeset text (surely:-)

@DavidCarlisle no, no one would ever do that. That's why there are no packages called soul, ulem or lua-ul :-)

\documentclass{article}

\usepackage{amsmath}

\begin{document}

$\nabla \times \mathbf{F}$

$\mathop{\nabla \times} \mathbf{F}$

\end{document}

Any physists around? is curl nabla times f with an infix times or is it the operation "nabla times" with mathop spacing for the pair, and either way, how do you vocalise this if reading an expression. Unusually for this chat, this is a real question (mathml/accessibility setup question)

@DavidCarlisle what's "curl nabla"?

Is that just nabla? If so the first one.

Nabla is just a vector containing partial differentiations so you use it just like a vector, not as an operator.

(but I'm just an engineer :P)

@Skillmon missing punctuation I meant is curl "nabla times F" or "(nabla times) applied to F" (and actually the main question is what sounds do you hear in your head when reading this (in a language that uses spaces between words)

@Skillmon should a screen reader say nabla times f, or should it say curl F (assuming the author has not supplied any accessibility text and the system is having to make something up)

Happy Rubber Ducky Day! @CarLaTeX @samcarter @PauloCereda

@DavidCarlisle Ah, I never really messed with that, though. But I've seen it.

@DavidCarlisle With that notation, I think I consider it as a "formal crossproduct" of \nabla and F.

@DavidCarlisle I suggest "nabla times F" rather than "curl F" for this.

@Skillmon @mickep thanks

@DavidCarlisle (I can understand an argument for "curl F" or "rot F" as well, though.)

@DavidCarlisle nabla times f

@UlrikeFischer Happy rubber duck day to you, too!

so tex-code person has now deleted their account....

@daleif probably gone to ask @samcarter at topanswers.xyz ...

@daleif His/her questions were not well-received, so I am not so surprised after all.

@Skillmon Yes, but...5 years ago only Paulo was promoting the duck thing (in each of his replies, at least one cs must have been named duck or something). I have to conclude, that he achieved his goal.

At least that's how I remember it.

@MaestroGlanz well you're writing to the maintainer and author of both duckuments and ducksay... :)

@mickep well, if you ask the same question 5 times and insist on bending the rules to your liking you shouldn't be too surprised

@Skillmon Oh, of course. And now we (at least I) are not surprised that he did not bother anymore.

@DavidCarlisle also you should consider that there is also "Nabla dot F", I'm not sure about the English language, but in German I think of it as "Nabla kreuz F" (so "Nabla cross F") to make the distinction between cross product and dot product very clear.

@Skillmon yes divergence (with dot product) is the same en.wikipedia.org/wiki/Curl_(mathematics)

@DavidCarlisle I know that that's the divergence, I was just trying to remind you of its existence so that you consider it when defining what the screen reader should say. Since I have only very limited knowledge about the English vocalisation of maths I thought this is the best way to achieve this :)

@mickep agreed. Though I'm a bit disappointed. I'd hoped they change their line of questioning to fit the format and maybe pose interesting questions.

@DavidCarlisle and my initial confusion about your "curl" is because I only think of Nabla cross F as the rot(ation).

@Skillmon yes Unicode has U+2A2F which is explicitly cross product, but in practice MathCat has to cope with a normal Times, and it might be harder to get a specific reading if normal times is used

@DavidCarlisle well, tagging should tag it correctly, there is little chance for ambiguity with \nabla\times ... not being the cross product.

@daleif Who's this and what happened?

@Skillmon Yeah, they seemed to be very angry that 'TeX' can be used with a variety of meanings - I'm still not totally sure what meaning they felt it should (exclusively) have

@Skillmon yes but that essentially was the question, tag as <mo>\nabla\times</mo><mi>F</mi> or `<mo>\nabla</mo><mo>\times</mo><mi>F</mi>``

@UlrikeFischer yaaaay! Happy Rubber Ducky day!

@daleif That dude? "What does it mean, if I say "TeX"?"

@MaestroGlanz Apparently so. TeX (the typesetting system) or TeX (the language)? Who knows?

@PauloCereda roast dinner to celebrate

@DavidCarlisle oh no

@daleif One of their question was basically similar to “how do I program a Web browser in Assembler?”

@samcarter probably a confused vim user.

@UlrikeFischer Happy rubber duck day to you as well! (I was just reading a paper from (I presume) a relative of the Bär: Ealeal Bear)

@DavidCarlisle how did they get out to ask a question?

what's happened to all the hyphen-.... packages?

@DavidCarlisle everybody now only writes in english?

[53/71, 00:38/01:18] update: hyphen-russian [34k] (58609 -> 73410) ... done
[54/71, 00:38/01:17] update: hyphen-sanskrit [5k] (58652 -> 73410) ... done
[55/71, 00:40/01:21] update: hyphen-serbian [25k] (58609 -> 73410) ... done
[56/71, 00:40/01:20] update: hyphen-slovak [11k] (58609 -> 73410) ... done
[57/71, 00:41/01:22] update: hyphen-slovenian [7k] (58652 -> 73410) ... done
[58/71, 00:41/01:22] update: hyphen-spanish [150k] (58652 -> 73410) ... done
[59/71, 00:42/01:21] update: hyphen-swedish [18k] (58652 -> 73410) ... done

@samcarter are there alternative languages available?

@DavidCarlisle can't think of any :)

@DavidCarlisle I just wonder why the download wasn't bundled (that's what I meant with "I had one or two" the other day)

@DavidCarlisle how do you tag \nabla\cdot F?

@Skillmon yes I had several yesterday so was a bit surprised to get a whol elot more today.

@DavidCarlisle I had the entire bunch yesterday, I guess.

@Skillmon well if the author supplies a mathml version for formula, we use that, otherwise luamml doesn't currently know that's any different from 2\times 3=6. The question is really from Neil wondering what MathCat (and similar systems) should do in the absence of any author tagging

@DavidCarlisle albatross got an update today, but we haven't done anything. :)

@PauloCereda birds already know all languages.

@mickep ooh

@PauloCereda so it still acts as a room heater?

@mickep Utbildade människor såväl som fåglar

@DavidCarlisle Oh, indeed!

@DavidCarlisle ooh

@DavidCarlisle Oh, you have access to very powerful tools on texlive.net

@egreg so an arch linux user...

@daleif ooh no!

@daleif I'm using Arch, btw.

@yo' 2 OverLeaf questions: (1) Compilation time limit is down to the individual user (aka does not matter if the owner of the project is a paying user. (2) What about the collab tools when the users on a project is a mix of free and paying users.

@Skillmon They either know what they are doing, or ask a lot of questions.

@daleif or (like myself) don't know what they're doing but keep quiet about it. (jokes aside: the Wiki is phenomenal)

@Skillmon it find it too easy to do something stupid with. Think we had a PhD student once who was using arch and needed to use cuda for this gfx card. He f'ed up his arch so bad...

@daleif yes, you can break your system pretty easily (just yesterday I had to rewrite my boot partition using a second laptop... I have to fix my update script)

@mickep I thought you were going to say this: texlive.net/showtags

@daleif 😂

@DavidCarlisle wasn't there also a version where one can compile code first?

@UlrikeFischer texlive.net/ngpdfb.html (I'll probably drop the b at the end and move that to texlive.net/ngpdf at some point)

@DavidCarlisle That does not go along with our non-tagging project.

@mickep You need powerful software (or @UlrikeFischer) to generate tagged pdf. If you insist on lesser systems it's not my fault.

@DavidCarlisle We want to protect the users from Adobe or others to collect the data...

@DavidCarlisle Oh, very advanced. (Hans sent me a show-off lua solution as well, but that one is too long to post at texlive.net)

@mickep not sure what you mean, but protecting users sounds like a good plan

@DavidCarlisle You can try this one

@mickep but without tagging, how do you get a screen reader to read it in an accessible way?

@DavidCarlisle That is not meant for the screen reader.

(Not that the screen readers seem to read math too well, last time I tried)

@mickep I'm confused, apart from demonstrating context, what is it showing?

@mickep that could also be said about some human readers too :)

@mickep we are working on that...

Not ducks, of course, we are very good at reading maths

@DavidCarlisle Did you try to copy and paste from it? :)

@PauloCereda Yes, and ducks.

@mickep result not too good, hyphenated words not re-joined:

We thrive in information--thick worlds because of our marvelous and everyday ca
pacity to select, edit, single out, structure, highlight, group, pair, merge, harmo
nize, synthesize, focus, organize, condense, reduce, boil down, choose, categorize,
catalog, classify, list, abstract, scan, look into, idealize, isolate, discriminate, dis
tinguish, screen, pigeonhole, pick over, sort, integrate, blend, inspect, filter, lump,
skip, smooth, chunk, average, approximate, cluster, aggregate, outline, summarize,

@DavidCarlisle :(

@DavidCarlisle Here, all characters are changed into ❌

Ah, maybe that was added after the TL2024...

@mickep I did wonder what that X was doing. I tried firefox and Chrome and got the same result in both

The latest non-tagging code is not in TL24.

@mickep so what does it do when not replacing everything by X? (I did always say that chickenize was the best thing about luatex)

@DavidCarlisle You have a secret non-tagging mail.

@mickep thanks

@DavidCarlisle If the directive is not known, I guess it does nothing.

@mickep that works (even if it is a secret)

@mickep no, I meant what is "nontagging" in this context, I thought you meant doing accessibility a different way, but did you just mean concentrating on other things at the moment?

@DavidCarlisle We are done with tagging (and nontagging). Different mode :)

(sorry, I guess your page is indeed useful!)

I just read a question about texmf. How do you read texmf in your head?

@MaestroGlanz tech em eff

@mickep and what am i to do with an emacs buffer full of red crosses? :-)

@DavidCarlisle emacs has never been more beautiful

@DavidCarlisle but that doesn't do validation? Or does it?

@UlrikeFischer no (it predates the schemas) it could do though. Also I should make the code pick up the error from show_pdf-tags if the file isn't tagged pdf (or not pdf at all) and show something more reasonable

@JosephWright \int_step_tokens:nn is dearly missing: tex.stackexchange.com/a/734942/117050

@Skillmon PR?

@DavidCarlisle On a bit more serious note, one thing we concluded was that tagged pdf is less natural than conversion to XML/html. For accessibility. And that export works quite well (but polishing is needed)

@mickep yes that tension is always there whether pdf is essentially a lost cause for accessibility and you should generate html. That is a common feeing with users with bad experiences with pdf. But we'll see whether good enough tagging can rescue the format for accessibility.

@mickep sure but with tagging you get an html conversion basically on-the-fly in the browser. Look at the texlive.net/ngpdf example: by clicking on the "as derived html" you get an html (or with the xml export an xml). That means you don't have to sent around say an invoice or booking confirmation as pdf and a additional html version for accessibility, you get both in one.

Interesting to read, about the facade of the UvA library: letterror.com/uva/index.html

@DavidCarlisle I wish you the very best luck!

@UlrikeFischer Yes, I understand that. And maybe that is a use for it, nice to have one file. I have a feeling though that tex -> html directly is easier than tex -> pdf and then pdf -> html. But maybe the result gets "good enough".

@mickep well you need it too. If the pdf format dies as it is not accessible and publisher have to turn to epub or something like that then who will be interested in all the fine-tuning you are doing for math and paragraph typesetting?

@UlrikeFischer Well, so far everybody I have met seems more interested in good output than in accessibility. But I understand I do not meet all people. And ConTeXt does a rather good conversion from tex to hmtl/xml.

@mickep well I loath it if I get some booking confirmation or an invoice as html. It is basically impossible to store or print in a reasonable way.

@mickep but that means that you mainly interested in a PDF output. You want PDF but do want to do something to keep the format alive. Beside this: how ernestly does context take html for accessibility? How many of the manuals exist in this format and are available for a user?

hm, there is a not missing in the second sentence ^^^

@UlrikeFischer None. There has been no demand so far. It is of course pdf that is the fun part. That is where we can do interesting and nice documents.

@DavidCarlisle Physics type here. "\nabla \times" is a notation documented by Gibbs to stand for "curl of". Curl is never correctly defined as a cross product. It is only a coincidence that it is calculated as a cross product in Cartesian coordinates, but this is most definitely not the definition of curl. The problem is that \nabla is an operator that takes different forms in different coordinate systems, and it's cross product incarnation only works in R^3 while curl is more general.

@DavidCarlisle Thus it it more correct to think of "\nabla \times" as the operator acting on a vector field. Similar arguments hold for "\nabla \cdot" and divergence. It's not defined as a dot product but it coincidentally works that way in Cartesian coordinates. "\nabla \cdot" is the operator symbolizing divergence. When reading, I almost always vocalize them as "curl" and "divergence."

@mickep yes but if it becomes illegal to distribute inaccessible documents then untagged pdf may become a problem. Of course it might be that distributing a parallel html makes things legal again but it's hard to know how laws develop in different places.

@LaTeXereXeTaL thanks

@DavidCarlisle It will be interesting to follow this.

can somebody tell me what the Windows command should be here?

I have a guess based on l3build-file-functions.lua and the OP's comments, but I am not at all sure about set since I wouldn't use export in this case, which l3build has as equivalent.

@cfr you could set it in a texmf.cnf file in a cross platform way rather than the environment but it is set on the commandline see superuser.com/questions/212150/…

@Skillmon You have a license called beer-ware which you insist that people include if they publish code involving something you wrote. I wasn't aware of this until recently, and have some of your code in a bigger project (which I intend to document appropriately).

I am concerned that there are possibly others with licenses like that who I've used the code of, without knowing of their license. What do you think would be a reasonable approach to respecting the authors of the code I used. I will mention that any time I used someone's code, I included a link to the tex.se thread.

I don't want to accidentally steal someone's property if I publish something. And rest assured, each time I use your code, it will include the license.

I now wonder if @DavidCarlisle has a tea-ware license.

I'm just scared that if I publish something one day and accidentally don't include a license (because I don't always read people's profiles) that it could come back to haunt me

@JosephWright github.com/latex3/latex3/pull/1658

@Jasper all TeX code I publish on TeX.SX is dual-licensed as beer-ware. You can choose whether you want to use my code under CC-BY-SA (which is the site's default -- and mandates you to use code under a compatible license to CC-BY-SA and that you add a comment to your code or documentation where you got the code/who wrote it), or as beer-ware.

@Jasper (not a lawyer) while users can choose to publish their content under additional licenses, they can't change the cc by sa license under which SE publishes posts by default. As long as you comply with that, you should be safe

@Jasper there's also a question where many of us who dual-licensed their code are listed, if that gives you peace of mind :)

@Jasper see here for a list of people who make any code posted to tex.stackexchange available under less restrictive conditions than the CC licence used by the site tex.meta.stackexchange.com/questions/1255/…

From ConTeXt list: tetris in pdf th0mas.nl/downloads/pdftris.pdf

(@UlrikeFischer, so the format seems ok...)

@mickep clearly context is more useful than latex

@DavidCarlisle It is not done in ConTeXt, it was just linked to on the mailing list :(

@mickep oh was that hand written? the pdf is set out with indentation like code not like machine generation

@DavidCarlisle You can read a bit about it here: github.com/ThomasRinsma/pdftris

@DavidCarlisle I guess, but is it really that hard to set it on the cli in Windows?

Thank you guys, I was worried about this, but now I feel reassured. I will abide by the site policy by default, and if I become aware of another license, then I will respect that in that instance.

@cfr I think the LPPL license would make the work seem more legitimate to people than the site's license :P. I dunno, it just sounds very professional

@cfr you can use set but windows users find it hard to find a commandline (you normally set the environment via a control panel menu:-)

@Jasper LPPL isn't my favourite licence, even though I'm one of the authors. Have you read it? It's somewhat longer than the MIT licence.

@DavidCarlisle WTFPL FTW

@Jasper Like @DavidCarlisle says, with a free hand MIT would be my plan

@DavidCarlisle I haven't read the license, the name is just very legitimate sounding. I'll have a look

@JosephWright What is a free hand MIT license? Does that just mean adding an MIT license?

@Jasper joseph meant if he had a free hand (ie, no existing constraints) he would choose MIT

Ah, thank you David. And thank you to everyone for your assistance in helping me navigate this

@Skillmon thanks

@DavidCarlisle Exactly

@Jasper LPPL is legitimate and I use it on latex related things as having the core latex distribution under a single licence is a good thing. But on anything else I tend to use MIT