@samcarter Heh thanks! (Are you saying I flood meta? I don't) :P

Sorry heh I went to sleep, you messaged me at midnight! (for me at least)

@samcarter And what's it's advantages over using word?

@DialFrost I chose LaTeX over World because once I tried to add a math formula in a very big Word document and all the document went crazy - no way to amend it. I had to restore a previous version and write the formula as text. LaTeX is incoparably better in the way it handles formulas, indexes, bibliography. But there is also a problem, if you have to collaborate with people who don't know LaTeX, it's better if you use Word.

@DialFrost For one thing, it's a programmable system. I think maybe Word has some kind of macros, but I doubt they are comparable. In TeX one can alter every aspect, because it's a programming language.

@CarLaTeX Well, one could try to help them to see the light. Of course, not everyone wants to see the light.

@FaheemMitha Unfortunately, it is not easy. I have colleagues who have problems unzipping a .zip file...

And similar comments apply to LaTeX, because LaTeX is just TeX macros.

Allow me to ask the TeX gurus in here informally: What is the status of PDF/A and PDF/X in LaTeX (in the form of the pdfx package)? Is it considered mature and stable at this point? More importantly, has it reached a status where this is something that I “ought to” use in my everyday LaTeX work, including notes, articles, and teaching material? That is, should I always run a code snippet for either PDF/A or PDF/X, depending on whether the document is for printing or digital distribution?

@Gaussler the pdfx package is not compatible with the new pdfmanagement. There have to set the standard with \DocumentMetadata options (only pdf/A currently, I didn't get requests for pdf/X support yet). pdfx is not compatible with hyperxmp. I'm also not very happy how pdfx loads hyperref and how it patches or overwrites other packages.

Beside this: LaTeX can only enforce some parts of a standard, for others you are responsible, which means you need to use a validator, that is extra work, so I would only do it if you need it.

@Gaussler no idea if he is still involved. Imho currently only Ross is working on it. It wouldn't be very difficult to adapt it, it mainly has to replace primitives like \pdfcatalog with the management commands, but as pdfx has no driver files but intermixes the engine codes it is difficult to do it from the outside. And yes you can use \DocumentMetadata{pdfstandard=A-3b}, see texdoc ltdocinit. For the xmp-meta data you should then use hyperxmp.

@CarLaTeX So, what do those colleagues do, professionally?

@FaheemMitha ordinary clercks

Strange: I have two issue reports about rounding to uncertainty which both came up late yesterday: github.com/josephwright/siunitx/issues/616 and tex.stackexchange.com/questions/653180/…

@JosephWright With one more you could say that you have five of them.

@barbarabeeton Do you know if it, for each unicode symbol, is written somewhere why it was added, and what it is supposed to denote? Take compart.com/en/unicode/U+29B3 for example (there are many many more!). "Empty Set with Right Arrow Above". What did it denote? Why was it added (shouldn't it be up to the program to be able to add arrows on top of symbols?)?

^ Same question to @DavidCarlisle, or anyone else who happens to know.

@DialFrost let's say I appreciate very much that the volume of ipad questions on meta slowed down in the last couple of weeks :)

heh

Funny :P There was only like 1 post on that

@mickep U+29B3 was part of the Unicode 3.2 STIX submission about which @barbarabeeton knows more than anyone. but in that case it comes also from SGML ISO AMSO Entity set as &raemptyv with (full) description circle, slash, right arrow above The SGML entity definitions are famously vague, literally that is all you have, no mapping to a font, no picture of a glyph, no usage description other than it being in AMSO implies it is an "operator"

@DavidCarlisle Thanks, can I find that information you just wrote somewhere? I probably better search for "SGML ISO AMSO" or so...

@mickep I got all that from unicode.xml which is, as I say, "everything I know about Unicode"

@DavidCarlisle Hehe, ok :)

@DavidCarlisle ooh-nicode

@PauloCereda U+Breakfast

@DavidCarlisle 6f68206e6f

@DavidCarlisle 🍳

@UlrikeFischer pan fried duck?

@DavidCarlisle very young ducks

@mickep w3.org/2003/entities was part of a proposal to update the SGML sets but it has some background info (annoyingly the ISO documents are not freely available)

@DavidCarlisle Thanks, I will have a look. (This whole unicode math mess is a bit frustrating.)

@UlrikeFischer oh no

@DavidCarlisle oh no

@PauloCereda oh no

@mickep welcome to my world

@DavidCarlisle we could have unicodexit :)

@PauloCereda who needs accents anyway?

@DavidCarlisle :)

@DavidCarlisle Oh, so this is what you struggle with?

@PauloCereda Best suggestion, so far. :)

@mickep ooh

U029B5 CIRCLE WITH HORIZONTAL BAR ohbar math other \ElsevierGlyph{E260} Sm infix

I wonder how much Elsevier payed to get that one in...

(looks a bit like \ominus, with a too large minus...)

@DialFrost If you care about typographical output, tex.stackexchange.com/a/110140/250119. Otherwise, probably not much benefit.

There are such things as antiword (Python package to parse Microsoft Word file) and VBA, as well as Python HTML libraries.

Although I don't really use them myself.

@mickep as one of the primary supporters of stix, probably rather a lot, given how long it took:-)

@DavidCarlisle Oh, yes, many big players there...

At some point in the past, it appears there were plans for a new output routine called xor, complete with grid typesetting and other things, see e.g. this question and the answer by @JosephWright. What happened to that project? It appears that Github has not received any updates on xor for six years or so

@Gaussler there is new O.R. salvaging bits of xor in the works (see footmisc for example)

@Gaussler Frank has ideas: at present we have other priorities (tagging PDFs is needed to 'stand still' much more) - email him about 'Alice' (it's not public code)

@JosephWright So in what decade can I expect to be able to do grid typesetting in LaTeX?

@Gaussler A bit unfair. Grid typesetting seems to be very difficult to get right.

@mickep I wasn’t trying to be unfair, sorry if it sounded that way ;-). I am simply looking greatly forward to it :-)

@Gaussler I see. I know you played with it in context, but also there it is not always working as expected, as you know. I could only imagine how difficult it would be to have that running in a stable way with all the various documentclasses and packages around.

@mickep That’s why implementing a new output routine would probably be equivalent to making new document classes. I know there are no plans for an actual LaTeX version 3, but one way to almost get there, and simultaneously do something about the package hell, would be to have classes with names like l3article, l3book, l3slides etc. These would then take care of the stuff that was previously handled by hyperref, microtype, fontenc, fontspec, etc.

Of course, such classes should be for new documents and should make no attempt to be backwards compatible with the old classes

LaTeX will never be like ConTeXt, but it could become a lot closer to it if we wanted it to be

@Gaussler I know too little about LaTeX and programming to have an opinion about that.

@Gaussler grid typesetting is really so far from what tex was designed for it's always going to be tricky. While xor had some things that helped grid setting wasn't really the main focus. context has an easier rout as it is luatex-only, doing grid setting with pdftex or xetex it's always going to be a matter of how much you can avoid the underlying typesetting algoritms pushing you off grid.

@Gaussler So basically you get no help from the engine at all. TeX is designed throughout for an elegant box and glue model which is fundamentally incompatible with grid typesetting.

@DavidCarlisle There should be an obvious solution to that problem: Implement grid typesetting as a LuaLaTeX-only feature

@Gaussler Not really possible, it is not a discrete feaure: every math display, every theorem, every list, every section heading, ... needs to avoid adding glue, so really you need a luatex-only format which may not be a bad idea, but it's not really latex.

@DavidCarlisle Can't you just add a callback that changes how the gluing primitives work?

@Gaussler add_to_callback_ magic_wand yes, perhaps. Can probably do something (witness the many packages that try) but the chance you can make some single global change that makes all the logic adding glue around every vertical display construct pick a meaningful grid position with no changes required to the logic in individual environments seems optomistic, so we are back to current situation that things work when they work and don't work when they don't

@mickep -- If you have the patience, you can get some information from the STIX page at AMS. The information provided by the STIPub organizations consisted of tables or lists of symbols that they "needed" (had used) in their publishing activities; there was almost never any meaning indicated, and later exploration was required to find in-context use, in print, as requested by the UTC. I'll follow up in email.

@barbarabeeton Thank you! Is this in-context use documented? If not this seems just a bit too arbitrary. (I think for example about the requests we sometimes see here on the site for some strange combination of known and used symbols, often solved with the help of the stackengine. That could by the logic of UTC make them candidates for unicode...)

@Gaussler -- Grid typesetting and math content are not easily compatible. For example, multi-line section headings at a type size larger than the regular text; display math, especially multi-line blocks; footnotes in a smaller type size; there are more, and authors can come up with many that I've never seen before. It can be done, but is a real challenge.

@mickep -- No, not documented, but in-context examples, preferably showing the meaning of the symbol as defined by the author, is required by the UTC if the symbol is not well known. Well, not absolutely required, but it certainly makes acceptance easier. Searches could be difficult, but having an archive of the sources was often helpful; otherwise, a request to likely knowledgeable mathematicians, followed by a library search, was involved.

@barbarabeeton It would be intereresting to see such examples for ⦴ (U+29B4), available probably only in stix, in unicode as \emptysetoarrl. I mean, why not just an arrow on top of the emptyset? (I know that you are not responsible, and I am not after blaming you or anyone else. But you might be one of the few who know how the reasoning has been going.) We do not have a slot for a right-arrow put on top of u (\vec{u}), something that must be so much more common.

@UlrikeFischer another public git adoption:-)

@mickep -- As @DavidCarlisle has already pointed out, that symbol was already present in ISO 9573-13, most of which was not present in Unicode when the STIX project started. As it already had ISO recognition. 9573 was adopted as the starting set; no special effort was made to further vet or justify those symbols.

When I retired from AMS, I left the STIX paper files in reasonably well-organized condition. Whether they still exist, or have been tossed in the recycle bin is unknown to me.

@yo' did you see tex.stackexchange.com/q/653014/1090

@barbarabeeton Ah, yes, that is right. So it goes back to iso...

@barbarabeeton Oh, I really hope they kept the files. What a waste to throw that away.

@mickep the one time I (with @barbarabeeton's help) tracked down one of the 9573 editors to ask the origin and intended meaning of one group of characters, the answer was more or less "can't remember any discussion, you have probably thought about it more, define a suitable meaning" The SGML math entity sets were done with no real discussion or testing and are just a list of names with no font or encoding information.

@mickep Following generations struggled to make sense of them and XML, not having SDATA entities required a mapping to Unicode and preferably to character data not (say) mathml over-arrow markup so they are what they are

@DavidCarlisle So you could then define 29C3 as "duck as I draw it", if you'd like...

@mickep yes

@mickep -- I've just checked in the most recent table posted from the AMS STIX page I cited earlier. U+29B4 appeared only in ISO 9573-13; it was not attested from any other source. As for the paper files I accumulated, I can't predict their retention or disposal. I was taken off the project before its completion, and some loose ends were never tidied up.

@mickep whch probably makes more sense than defining FORKING to be a symbol with fork crossed out and NON FORKING to be a fork symbol

@DavidCarlisle Oh, haha! I could not see that one coming! Hilarious!

@DavidCarlisle I just saw it. Where do you want to create it?

@barbarabeeton OK, thanks for having another look!

@UlrikeFischer it's there already I just need to flip it to public, did you not get a ping from github yesterday? github.com/LaTeX-Package-Repositories/ucs

@DavidCarlisle -- I actually remember that one. I think it may be an example where there was in-context documentation. I remember it not making sense, but if the original author defined it that way, ...

@DavidCarlisle no I didn't see it, notification settings were wrong.

@mickep like TeX in a way, they value stability over common sense so it is more or less impossible to get any bugs fixed

@DavidCarlisle I understand that. For math in the long run, it would perhaps be best to start over from scratch. Not that anyone would care about a new try... :(

@DavidCarlisle and @mickep, if you want an example of "stability" over accuracy, take a look at U+2118, Weierstrass p. When I first started looking at this stuff, it was in Unicode 2.0 as "power set" or something like that, and positioned so that it looked like upper case. Urk.

@barbarabeeton Interesting!

@mickep -- There's also the angle bracket foulup. Originally, I was forced, er, strongly urged to assign the CJK fixed-width IDs, since the shapes basically matched. Only after Murray Sargent came up with an example showing what would happen to the Hamiltonian equation using that strict definition (see Unicode TR 25) did the UTC give in to the concept that math is a language and start to accept math-specific items.

@barbarabeeton Oh, I saw those the other day. Another strange story. :(

@mickep -- The STIX/Unicode effort is a whole volume of strange stories. Not quite Stephen King level, but ...

@barbarabeeton I hope for a good ending.

@mickep it's opera: everyone dies

@DavidCarlisle -- Golly, I would have preferred Gilbert & Sullivan.

@DavidCarlisle -- I've just learned from a browser search that the important objects in lawn bowls are not totally spherical, but elliptical in one axis. Elliptical billiard balls, anyone?

@DavidCarlisle If you write manus, please let me go with a stab of 2B3D or strangled with a 29DC.

@DavidCarlisle From the MikTeXado?

@AlanMunn yes

@barbarabeeton yes you have lawn bowls with flat surface but weighted non spherical bowls, or crown green bowls which have spherical bowls but a small hill in the green.

@barbarabeeton Yes, that's part of what allows them to curve so much when they are bowled (although obviously that's not the only reason).

@AlanMunn actually had planned to "adapt" the lyrics but that verse seemed good as it was so no miktex needed:-)

@DavidCarlisle Ha, so the 'filling up the blanks' is original. Nice.

@DavidCarlisle What are the chrome clones you mentioned?

@AlanMunn ctan.org/pkg/carlito?lang=en is calibri I'm trying remember the other names

@AlanMunn en.wikipedia.org/wiki/Croscore_fonts#Crosextra_fonts

@DavidCarlisle Thanks. I found answers of yours pointing to them. :)

@DavidCarlisle and @barbarabeeton: Funny exercise: Go to search.arxiv.org. Search for bumpeq (that is 224F, but also from ams binary relations). You will get one (!) hit. Go to that article and find out how it was "used". (I have done this with various symbols, and many of them are simply not used in math or physics papers, it seems.)

@mickep -- Another place to check is the Comprehensive Symbols List, which I just did. \bumpeq (and \Bumpeq) is in the amsfonts collection, which means it was needed for AMS publications ... sometime. Those aren't necessarily om the arXiv, and possibly precede the use of TeX. That's a different search though, and although I have some older reference material, it's i a box, currently inaccessible.

@barbarabeeton yes, thanks, it was in that list I found it. I would have expected it to be in some arxiv manuscripts if it was really used. But maybe not. (I have no clue how it is usually used.)

@UlrikeFischer can you run make and/or perl/ I just deleted dozens of data files from git as they are genratd but I guess on windows that may be tricky? just typing make should be enough but if not, if this perl works, I can write the make file as l3build, I don't really want to translate the perl to Lua...

wget unicode.org/Public/UNIDATA/UnicodeData.txt
mkdir data
perl -w makeunidef.pl --nocomments --targetdir=data config/*.ucf

@DavidCarlisle I can run make there, but it fails then with ! I can't find file `ucs.dtx'.

I think because it sets texinputs

@UlrikeFischer ah OK I'll have a look, I'm adding some l3build tests anyway so I may get rid of the make file but if perl is OK that makes it easier

@DavidCarlisle it says

TEXINPUTS=.:data:utils: pdflatex -draftmode ucs.dtx
This is pdfTeX, Version 3.141592653-2.6-1.40.24 (TeX Live 2022) (preloaded format=pdflatex)
 restricted \write18 enabled.
entering extended mode
! I can't find file `ucs.dtx'.

That is in a bash.

@UlrikeFischer ah well you want ; not : in TEXINPUTS I think

@DavidCarlisle we need a .gitignore file ...

@UlrikeFischer local_pdflatex=TEXINPUTS=".;data;utils;" pdflatex works for me (in GNUmakefile)

@DavidCarlisle yes works for me too. But it deletes all log-files.

@UlrikeFischer

Running checks on
  sx-1 (1/2)
  unichar-1 (2/2)

  All checks passed

@DavidCarlisle good test coverage ;-)

so l3build says it's good and we can blame @JosephWright for any errors

@UlrikeFischer you have to copy utf8x.def to build/local first at present as l3build only testing not building, but I'll get to that

@DavidCarlisle but one still need to run make first?

@UlrikeFischer yes I'll fix that tomorrow..

@DavidCarlisle it is not a problem.

@UlrikeFischer currently if you use \usepackage[utf8x]{inputenc} you get standard [utf8] plus a (new) \unichar , if you load ucs before that then instead you get existing ucs/utf8x (and nothing works:-) i think ucs.sty needs to revert at least MakeUpperCase, but not today.

@DavidCarlisle it should also revert the utf expansion of hyperref. Btw: I looked up the docu a few days ago, and found that it has the interesting "feature" to provide a number of chars as little pspictures.

@UlrikeFischer yes, but not as well implemented as this:

@DavidCarlisle ;-)