out of curiosity, I asked the institutional accessibility tool what was wrong with my pdf. for some reason, it gives it a score of 6%. I have no idea what gets the 6%. obviously most of it is tagging and one bit is language, but it also doesn't like any number which is a hyperlink. non-linked numbers are fine, but if they are linked, it does not recognise them. is that expected? (I didn't do anything whatsoever to make it accessible, so I'm not sure what I got 6% for.)

@cfr What tool are you using?

@cfr We have Ally at work and I can get good scores provided I have PDF 1.7 rollmapping

@JosephWright Ally. I'm not surprised it gives me a terrible score. I just don't understand why it doesn't like numbers if they are links, but is perfectly OK with them if they are not in links.

@cfr That is odd - but perhaps I misunderstand - does it give more info?

@cfr I don't have that many links in the docs I upload that get a score - currently my slides don't work, only pout-and-out notes or worksheets, etc.

As I need to work on xbeamer

@cfr I can't test you tool, but links must be tagged and I would expect it to complain about untagged links.

@DavidCarlisle :*

@DavidCarlisle By then you will be able to do \framed[width=3lh-1sd]{I am framed}. Here lh and sd are lineheight and strutdepth, two new handy units, but the nicer new things is that you can do the calculations directly in width=...

@mickep that's also possible in l3keys or expkv-def's edimen-type (the calculations, not the extra units).

@Skillmon Very good!

@JosephWright they are missing the option "a mix of everything"

@UlrikeFischer :)

@JosephWright for your upcoming talk:

@UlrikeFischer :)

@JosephWright the one before was for you too ...

@UlrikeFischer that one's great!

@samcarter I saw your comment regarding the improvement of tabularray's speed, and it looks very promising, but it's at the expense of l3regex. I don't know what that is, but it's apparently powerful, and so I'm wondering if there's any downsides/limitations to constructing tabularray's code without using it?

@Atex regular expressions are a general language used for string matching. The code was using a giant powered sledgehammer with its own fusion power plant to crack a nut, so probably no bad effects will happen if it uses a manual nutcracker instead.

@Atex It isn't at the expense of anything.

@UlrikeFischer agreeing to @yo' that one's really great!

@Skillmon Possibly non-native speakers are more amused? :D

@DavidCarlisle This makes better sense! I was also wondering why the author was experimenting with replacing it if it was so powerful. I guess I misunderstood what "powerful" meant in this context. This definition clears it up and explains why the author is willing to replace it

@samcarter this is a huge win for tabularray, then! I wonder how its new runtime compares with the regular tabular environment?

@Atex @JasperHabicht left a comment under a recent question that some examples were still about 4 times slower than longtable but that's basically quite reasonable since it does rather a lot more. You would not have loaded tabulararry into tex on a 640K memory on an IBM PC, which was a constraint longtable had

@DavidCarlisle I see, but longtables are also more difficult to deal with, and not used as commonly, right? How would the regular tblr environment compare to tabular using the updated statistics? By my calculations, it's approximately 71% faster going from 0.7 to 0.2. Using this comparison, it'd mean that the test for "tabularray-small" should render at approximately 8 seconds instead of the

original 27 (approximately) seconds under "real", but I'm not sure I can just translate the data like that to the comparison. There's a lot of details I don't know about how both tests were performed, after all

@DavidCarlisle Never mind, I found Jasper's comment, and it seems to be in line with my theoretical calculation - Amazing to see this tremendous improvement in a practical example! While it still isn't nearly as fast as a regular tabular, it certainly does decrease the gap significantly. That leaves only 1 other major problem with tabularray that I've heard of: pdf tagging.

@JosephWright not really. it just draws red boxes around them if I ask for details.

@Atex Just for your information: I used l3benchmark and put tic directly after \documentclass and toc after end document for my tests.

@Atex Tagging is a planned addition as far as I know, but it might require coordination of different people. Currently, the code seems to roughly support it, as far as I understand. But I am not a tagging expert

@Atex not sure why you say longtable is less common it's part of the core latex release and had 30 years head start on table, so I would be very surprised if it's not used far more at the present time. It is also of course utterly brilliant code by a well respected author

@UlrikeFischer hmm, yes, but it complains that it does not recognise the characters specifically. obviously it complains about the lack of tagging. (and the lack of language in the metadata.) but this is supposed to be a different issue. It does not complain about links to the glossary, bibliography or URLs, for example. only things like page numbers.

@Atex even with getting rid of regex it's still slow (but not as terminally bad as with regex).

@JasperHabicht this look really promising. It seems like all of you people working on the tabularray package have been focussing on addressing the package's weakpoints this time around rather than only adding new features/optimizing - Huge respect! It seems like tabularray is finally on the way to become the "one-package-to-rule-them-all" within tables, kind of!

@Atex apart from recently @JasperHabicht none of us are "all you people working on tabulararray" it's all mostly one person's work.

@DavidCarlisle oh wow, I didn't know longtable preceded table! Perhaps I don't use it commonly because I don't deal with that large datasets as a 2nd year student yet, but it'll probably come, if we choose to translate the data of the python code to a LaTeX table in our reports

@Skillmon Agreed, but the gap is not as significant as before (only approximately 4x slower than regular tabular now as opposed to around 12x previously!), which is still a huge win, and perhaps the trade-off in speed is now worth it with the increased utility as well as the cleaner code layout!

@DavidCarlisle True, and I only provided some basic code for a TikZ library. To really contribute to the package in an efficient way, one has to understand the code and the logic first which takes a lot of time ... thus, I only "contribute" with ideas currently =D

@Atex no obviously it didn't precede table (it is older that latex2e though) I was comparing with tabularray when i said it had a 30 year time difference

@DavidCarlisle Oh, I just didn't remember the list of contributors and wouldn't risk excluding anyone's effort, and thus I thought I'd just address everyone involved. They'd know who they are, I hope;) But yeah, I've got a huge amount of respect for the author. He practically revolutionized the layout of tables, which I'm appreciative of. Before discovering this package, I hated writing tables, because correcting them was a nightmare (cluttered code listed in cells)

@Atex shrug, most of the code people add to tables (lines, coloured panels etc) make them look worse, i can't say I have ever needed more than a&b&c\\ in a real document and tablr doesn't make that any simpler.

@JasperHabicht Humble as always - Am looking forward to the tikz implementation, too! That'll (hopefully) eliminate the need for nicematrix as well, and thus you can have everything you need within 1 package :D Your contribution, even if they're "only" ideas, are appreciated, friend!

@Atex Currently the issue is how to implement the TikZ stuff with longtblrs ... which means that everything needs to be split up somehow. But it seems Jianrui is quicker with implementing new code than I am with coming up with ideas currently ... So, it feels that the tikz library is almost ready

@DavidCarlisle oooooh, that makes more sense! In that case, it wouldn't surprise me if tabularray still needs a significant amount of optimizations in order to truly be able to replace longtable/regular table in effectiveness. The package is practically a baby in comparison;)

This guy is really amazing.

@DavidCarlisle True, I suppose if all you need is the booktabs layout for your tables, you'll be just fine without tabularray and keep simple and manageable table code. I guess you keep to clean/minimalistic style tables, which is also a very efficient approach!

@Atex but surely you aren't saying the standard tabular packages such as array, longtable, dcolumn, delarray, colortbl, tabularx, tabulary, hhline are deficient in any way or confusing? You surely wouldn't want to offend their author?

\fpeval{1+1} is standard at current LaTeX. But \FPeval{\result}{1+1} only with package, right?

@cis yes but you don't need a package for that, just \edef\result{\fpeval{1+1}}

@JasperHabicht Absolutely. I've been keeping track of the development of #29, and it seemed to stay still for a while, but when I checked earlier, I was shocked to see such rapid process at such a short time! It's truly amazing to have such talented members in the LaTeX community, uplifting the potential for members of the LaTeX noob community such as myself!

@DavidCarlisle checked

@DavidCarlisle In no way did I intend to disrespect any author of any of the other packages. As a matter of fact, I haven't even used any of the packages you just listed except for array. When I say I wish for tabularray to be able to replace the need for any other table packages, it means that I wish for it to have all of the possible functionalities one could think of for using a table so you don't need to load any additional packages for that particular purpose. It doesn't mean...

@Atex you might want to check the author of all those packages:-)

... that I think any of the other packages are deficient/confusing. As a matter of fact, they may be more efficient/better/robust at implementing whatever feature they're meant for. Meanwhile, the tabularray could have a mediocre (or perhaps even bad) implementation of that functionality, because the tabularray only included that function with the intention to cover that area. "A jack of all trades but master of none", so to speak, but I'd rather have all of the features collected...

... in 1 package rather than spread out. Then, if tabularray wants to truly compete with the specialized table packages in their respective areas, it can always have its code optimized/improved. That may take years or decades, especially in the case of the longtable example, hence why the bad joke about the baby example (I apologize)

@DavidCarlisle \toprule, \midrule, \bottomrule?

@DavidCarlisle WHAAAT, you've made that many packages pertaining only to tables!? That's also an incredible amount of work deserving of respect! Wouldn't it be easier to maintain these packages if they were all merged into 1, however? I could imagine it becomes quite difficult to manage/update all of the packages you've created, especially since the packages you listed only pertained to your work within tables. How many packages have you authored or co-authored in total?

@Atex don't forget indentfirst 😊

@Atex ctan.org/author/carlisle and ctan.org/author/latex

@DavidCarlisle Oh, I did not know longtable was done with your namesake.

@UlrikeFischer omg, he's been involved with 200+ packages by my count! That's a truly commendable effort - Thank you for your service, @DavidCarlisle! It seems like you've been working on various aspects of LaTeX for years. Probably also laid the foundation for many other packages where authors may have borrowed functionalities of some of the packages you've been involved with or been inspired by them:)

@Rmano He's authored so many packages, even for small, specific purposes, that I don't think it's possible for me to remember them all even if I knew them all 😅

@Atex ...not only packages... latex-project.org/about/team

@Atex He (but also a lot of other people around here) also helps people who try to do their first steps in creating packages, classes and other projects, passing his (their) knowledge on to others. I think this is one of the main reasons why the community here is so nice! There is so much support here which I really appreciate (and try to give back by answering on the main site).

@UlrikeFischer not to mention secret things like ho-tex (although that's secretly you these days:-)

@Rmano Wow, that's incredible! @DavidCarlisle, you've really dedicated your heart and soul to improving and making LaTeX available and easily accessible to us newbies! I also recognized some of the other names on the site. Also huge respect to them. These are the people who work in silence in order to maintain the good LaTeX experience for us!

@Atex oh I don't know a few hundred, pus Ulrike and I maintain all the packages that Heiko Oberdiek wrote (ho-tex support) amd amsmath then of course there is the latex format to which i have had write access since we took over from leslie lamport in 1992 or so, so i have has a while to write a few packages

@Atex Beside all great work he does, he is also a bit fun to tease. As you hopefully will notice eventually. :)

@mickep that is a very bad habit, you should not lead people astray

@JasperHabicht Indeed, it really makes you comfortable knowing that there are some people you can refer back to when encountering some struggles solving a problem :) Who knows how many of the newer LaTeX experts, such as Jianrui, for example, has come to be, "standing on the shoulders of giants", thanks to the likes of sir @DavidCarlisle. It's really nice to see this loop it creates, because then you pass on that knowledge to others!

@DavidCarlisle Like now. Sir?! Isn't it Duke?

@DavidCarlisle did I just spot... 2 typos!? I thought that was illegal when one has as much typographical experience as you - Can't afford such misspelling when producing new code for us to benefit from;) On a serious note, you've done (and are still doing) a lot of work for LaTeX and its community. How do you manage all of that, especially when you also work in another company?

@mickep I think I'm already starting to pick up on it a bit;)

@Atex But it is with warmth. I even suggested that the next version of ConTeXt should be named after him, mkdc.

@DavidCarlisle Sometimes it's just quite easy to pick up on bad habits - Just like when you had to correct me when I used to indicate a new paragraph with \\ 😉

@mickep of course, a little friendly tease won't hurt :-)

@Atex you must be mistaken. I see @barbarabeeton has just shown up she has known me for years and will confirm that I never make typos.

@Atex -- @DavidCarlisle has in some cases apparently redefined what teh correct spelling is. On the other hand, indicating a new paragraph with \\ isn't good practice; there are good reasons.

@barbarabeeton Since David is teh main authority on English spelling, you should also fix your dress colours before going to teh theatre.

@barbarabeeton On teh otehr hand, a forthcoming Executive Order will declare a felony using British spelling.

@barbarabeeton Across the world, of course.

@barbarabeeton ah yes, I had nearly forgotten that teh english dictionary is supposed to be updated according to teh spelling of @DavidCarlisle. Don't worry about teh \\ issue - I was schooled thoroughly in teh comments of this answer according to my rather bad practices.

@Atex you are not the first soul to be corrupted in this way

@egreg -- Oh, dear. Too many already. Not sure whether I should invest in Sharpie stock, or hope that they go out of business. As for British spelling, all I insist on is, if somebody submits something to TUGboat, the usage be consistent. (And am willing and able to ensure compliance.)

@barbarabeeton don't you think it's a brilliant idea to consistently use US spelling for command names but UK spelling for documentation, like the color package?

@DavidCarlisle :)

@DavidCarlisle Ah, I see that it's almost natural that people go down this route when getting acquainted with chat. Bullying you (in a caring manner, of course) seems to be the norm, not the exception;)

@DavidCarlisle -- (I've been away, trying to get the title of my deceased husband's car transferred to me, and then registered. Talk about complicated!) I'm willing to follow Don's preferences on spelling, although we disagree on some matters of punctuation. So commands, indeed, will continue to follow U.S. practice. "Ordinary" text, on the other hand, follows the practice of the correspondent, as long as it's English. I would question "Farbe", for example.

@Atex -- Some individuals deserve what they get on account of they dish it out. Doesn't mean we don't appreciate them for other qualities they may have. And, hey, I can nag DC about amsmath and MathML as well as teh idiosyncratic spelling. He should be glad I don't mention his accent (probably because he'd then laugh at mine.)

@barbarabeeton I guess the old saying applies: You reap what you sow

What? I stole a tick from @egreg? This must be some parallel reality... 😯

@Rmano excellent.

@Atex that and mentioning daily meals like for instance breakfast to see @PauloCereda's reaction.