« first day (4020 days earlier)      last day (1210 days later) » 

user202729
user202729
01:44
What's \file_path_include:n ? I see it mentioned exactly once in the LaTeX3 manual.
But no documentation entry.
 
1 hour later…
barbara beeton
barbara beeton
02:55
I just finished reading Louise Penny's mystery novel, The Madness of Crowds, and found this as the very last sentence of the acknowledgments. Remind you of anyone here? "All this to say, if you didn't like the book, it's their fault."
 
2 hours later…
VScode_fanboy
VScode_fanboy
04:37
@DavidCarlisle I know, when talking about server problem, I spoke on a wide range....
 
3 hours later…
David Carlisle
David Carlisle
07:49
@user202729 doc bug, see the ChangeLog the function was removed in ## [2019-01-01] (unless @JosephWright corrects me:-)
@user202729 l3deprecation file will show the equivalent new code is \seq_put_right:Nn \l_file_search_path_seq
Joseph Wright
Joseph Wright
@DavidCarlisle Fixed
@DavidCarlisle That's what I've just added ;)
Paulo Cereda
Paulo Cereda
08:16
quack
@barbarabeeton ooh
David Carlisle
David Carlisle
@PauloCereda breakfast
Paulo Cereda
Paulo Cereda
@DavidCarlisle oh no
For Mac people: Beware: Updating to macOS Monterey Is Paralyzing Some Older Macs Right Now gizmodo.com/…
Joseph Wright
Joseph Wright
08:32
@PauloCereda I've held off as my work still haven't caught up with Big Sur ;)
David Carlisle
David Carlisle
@PauloCereda they could install Windows 11 instead?
Paulo Cereda
Paulo Cereda
09:01
@JosephWright ooh
@DavidCarlisle oh no
David Carlisle
David Carlisle
@PauloCereda then they could install cygwin and have access to a broken arara
Paulo Cereda
Paulo Cereda
@DavidCarlisle oi
samcarter
samcarter
@DavidCarlisle :D looks very secure!
@PauloCereda I've also read articles that Monterey causes problems for many usb hubs/docking stations
Joseph Wright
Joseph Wright
09:19
@samcarter My docking station is USB-C so I'm probably safe
samcarter
samcarter
@JosephWright Let's hope you are (w)right :)
samcarter
samcarter
09:44
@DavidCarlisle They would also get all the expertise from stackoverflow: twitter.com/deanward81/status/1455109595844726787 twitter.com/tarynpivots/status/1455133975970676739 twitter.com/marcgravell/status/1455155985346174987 (seems as if they just hired the whole company)
David Carlisle
David Carlisle
@samcarter next up: buying the whole company?
Joseph Wright
Joseph Wright
@samcarter How did you spot that?
@DavidCarlisle In all seriousness, it would make sense
samcarter
samcarter
@JosephWright @NickCarver was so nice to summarise all these posts in his timeline twitter.com/Nick_Craver
Paulo Cereda
Paulo Cereda
10:00
@samcarter oh
samcarter
samcarter
@JosephWright ... if they would do it right, they could probably make a lot of useful combinations with github
Joseph Wright
Joseph Wright
@samcarter I can see the synergies for the main site, yes :)
@samcarter They would likely get into serious issues about askdifferent and askubuntu
samcarter
samcarter
@JosephWright :) I'm not sure they still are aware that there are other sites than SO
Joseph Wright
Joseph Wright
@samcarter Powers might not be, but you can bet Apple and Canonical's lawyers are (esp. as Canonical are or were paying for askubuntu)
Paulo Cereda
Paulo Cereda
@JosephWright oh no Canonical
@JosephWright I trust Canonical less than I trust Apple (and those two are bad)
samcarter
samcarter
10:04
@JosephWright yes, probably :)
David Carlisle
David Carlisle
@JosephWright possibly not since before long (if not already) the most used ubuntu installation likely to be wsl
Joseph Wright
Joseph Wright
10:46
@DavidCarlisle I was thinking from competition authorities
 
2 hours later…
Joseph Wright
Joseph Wright
12:33
github.com/josephwright/siunitx/issues/…
TL'13 issue ...
ihavenoidea
ihavenoidea
12:52
(Image from internet) Whoever made this, was it made "manually"? Or is there an automated way to draw 'infinite' trees?
user image
Joseph Wright
Joseph Wright
This is fun
@ihavenoidea Lots of ways: pick a tool
ihavenoidea
ihavenoidea
I'm currently using forest as it's the one I found it easier to use
Let me know if this should be an actual question, the other similar ones does not cover infinite trees like this one (to be the best of my knowledge, at least)
user202729
user202729
Use Python or something to generate TeX file
... is a solution
samcarter
samcarter
@JosephWright Tomorrow TL'13 Never Dies
Joseph Wright
Joseph Wright
@samcarter I suspect it's TL'13 initial release, so I'm going to be testing using TL'12 :)
samcarter
samcarter
13:28
@PauloCereda You've got some funny chaps in your neck of the woods bbc.com/news/av/world-latin-america-59136233
2
Paulo Cereda
Paulo Cereda
13:52
@samcarter ooh
 
1 hour later…
David Carlisle
David Carlisle
15:20
@JosephWright send them a decades worth of uktug coffee mats dvds
@samcarter looks dangerous, everyone is driving on the wrong side of the road!
samcarter
samcarter
@DavidCarlisle coffee mats? I though they are supposed to prevent the table from wobbling
David Carlisle
David Carlisle
@samcarter sure if you stand the table leg in the mug it has that effect
samcarter
samcarter
@DavidCarlisle ok, if the resident table expert says this :)
Joseph Wright
Joseph Wright
15:45
@DavidCarlisle I solved it, but the SAGE setup is weird: they've got TL'13 DVD expl3 but siunitx from the TL'12 era
David Carlisle
David Carlisle
15:58
@JosephWright yes saw the comments in the issue
Joseph Wright
Joseph Wright
16:16
@DavidCarlisle It's a weird one
Paulo Cereda
Paulo Cereda
16:41
Fedora 35 is out!
Joseph Wright
Joseph Wright
16:56
@PauloCereda yay?
Paulo Cereda
Paulo Cereda
@JosephWright Yay! :)
@JosephWright updating my workstation as we speak :)
Joseph Wright
Joseph Wright
I have a maths question ...
I've been asked to work out how to script some data analysis up. The basics are easy (read CSVs, plot, extract info, plot the result of that, do regression). However, the info bit is not.
We have some data where there's a 'flatish' bit that we need the midpoint for - how does one sort out a 'flatish' bit of a curve from 'the curvey bit' other than by eye?
I guess it's more a curve-fitting question really
Phelype Oleinik
Phelype Oleinik
17:11
@JosephWright Without much thought, I'd compute the derivative of that curve, then define a threshold of values for the derivative that are "close enough for it to be flat". That threshold depends then on how flat 'flatish' is
Joseph Wright
Joseph Wright
@PhelypeOleinik Good plan: I was thinking it's got to be about the derivative
samcarter
samcarter
@PhelypeOleinik You type too fast! :)
Joseph Wright
Joseph Wright
@PhelypeOleinik That I can probably manage: just have to fiddle with the threshold to match the by-eye approach
@PhelypeOleinik Thanks
@PhelypeOleinik Now I need to decide what to code it in: I'm obviously tempted by expl3, but this might be better in Lua/Python/Perl/something else
Phelype Oleinik
Phelype Oleinik
@JosephWright Yeah, I'd guess it's not much escape from some visual tuning
@JosephWright Glad it helped. Though secretly I stole @samcarter's idea ;-)
@JosephWright Yeah, I'd probably use TeX for that if I were in the mood to feel some pain :)
@samcarter If you were going to suggest that, I'm happy to know my idea wasn't that bad :)
samcarter
samcarter
@PhelypeOleinik I was typing basically the same comment, just with a prefix of "if the data is t not too wiggly" :)
Phelype Oleinik
Phelype Oleinik
17:19
@samcarter Ah, yes, a wiggly dataset would make things more interesting :)
Joseph Wright
Joseph Wright
@samcarter It's a straight-ish line with bend at the end, basically - so all I need to do is find the 'bendy' part
samcarter
samcarter
Different approach: fit both a constant and a slope to intervals of the data, compare test statistics if the addition of the additional parameter significantly improves the results. (determining the sizes of the intervals might be a pain though)
Joseph Wright
Joseph Wright
@PhelypeOleinik I've been asked based on my pgfplots approach to data plotting, so expl3 makes sense. Probably can arrange to do it on Overleaf
samcarter
samcarter
@JosephWright fit a broken line to it and let the fit detrmine the position of the break?
Joseph Wright
Joseph Wright
@samcarter This I'm not familiar with: some kind of Origin wizardry?
user202729
user202729
17:21
After all I feel expl3 is a reasonable programming language though, although a little assembly-like
@samcarter Isn't that polygonal approximation?
Joseph Wright
Joseph Wright
@samcarter Hmm
Phelype Oleinik
Phelype Oleinik
@JosephWright @samcarter In another note: I'd love to see the expressions 'flatish', 'the curvey bit' and 'straight-ish' written in a paper :)
user202729
user202729
Like, you can't use \tl_map_function if the computation itself is not expandable, But have to compute each element then \tl_put_right it
Joseph Wright
Joseph Wright
@samcarter I guess the derivative approach is like the by-eye one
@PhelypeOleinik Basically I'm trying to automate what we do anyway - the approach is not novel, it's just it's boring, and the group want to make it easier (we have the data in CSV from our hardware)
Rmano
Rmano
@JosephWright I know a collegue of mine here used algorithms designed to fit piecewise-linear function to data. I am unable to find the paper now, but I can ask
Joseph Wright
Joseph Wright
17:23
@user202729 Huh? \tl_map_function: is expandable, \tl_map_inline: isn't
user202729
user202729
No I mean, if you can't manage to implement the function itself as expandable then you can't use it
Joseph Wright
Joseph Wright
@Rmano It's OK: It think @PhelypeOleinik/@samcarter's idea will work nicely
@user202729 Oh, right: well yes
@user202729 If you are building up a token list, see tl_build approaches for better performance (we still need a better name for that area)
samcarter
samcarter
@JosephWright I'm not using origin, doing all my fitting with root. I would make piecewise definition of the function, e.g. f(x) = mx+t if x<z, h if x>z and then let the fit algorithm determine the parameters.
@PhelypeOleinik yes, that would be interesting to read
user202729
user202729
@JosephWright Probably is fine. Java has StringBuilder as well
Joseph Wright
Joseph Wright
@samcarter Ah, right
@user202729 I mean 'FMi doesn't like the name' ;)
@samcarter I'm not familiar with root, and we are talking here small data sets and trying to speed up some tedious-but-manageable manual work
user202729
user202729
17:28
~~~send data and we figure out~~~ (confidential?)
samcarter
samcarter
@JosephWright :) I had to use this chance to suggest some unnecessarily complicate way
user202729
user202729
Unfortunately by my experience (although I don't have much experience) for any "real life" "loosely defined" things you need either a lot of trial error (classical approach) or a lot of labeled data (AI stuff)
Joseph Wright
Joseph Wright
@samcarter I had a quick look at ROOT: it's probably overkill ;)
@user202729 Data isn't my research group's
Rmano
Rmano
@JosephWright found it: orbi.uliege.be/bitstream/2268/195856/1/paperLHM_IMPU98_ref.pdf
samcarter
samcarter
@JosephWright probably -- but you get all the power of c++ :)
Joseph Wright
Joseph Wright
17:35
@samcarter Ah, so I can blow my foot off
David Carlisle
David Carlisle
@samcarter wouldn't you rather have fortran
samcarter
samcarter
@JosephWright :) I didn't think a chemist needs help in blowing stuff up
@DavidCarlisle Yes, most of the software I use has some fortran core, but usually there is a wrapper of c around it :)
Joseph Wright
Joseph Wright
Right, lots of ideas: I will go home, eat and think :)
@DavidCarlisle Er, perhaps @samcarter would, but I am a simple chemist
(Currently debugging a DFT project and setting up another one, but still)
samcarter
samcarter
Funny story: twitter.com/foone/status/1229641258370355200?lang=en
2
David Carlisle
David Carlisle
17:53
@samcarter such wrapping is my day job...
samcarter
samcarter
@DavidCarlisle /me needs to investigate if I can blame @DavidCarlisle for all my non-tex problems :)
GeekestGeek
GeekestGeek
The developer of cleveref has not responded to e-mails to toby at dr-qubit dot org since 2021-10-22. His last cleveref version is from 2018. Any idea on whether he still has the chance to develop/debug cleveref and is interested in this? Any other active e-mail address to reach him?
David Carlisle
David Carlisle
@GeekestGeek you'd need to wait months rather than weeks before assuming it was dropped, I note it says you need to verify your email the first time, did the verification work OK?
GeekestGeek
GeekestGeek
@DavidCarlisle Thx! I think so, but I'll retry because I don't see it in my sentbox.
David Carlisle
David Carlisle
@GeekestGeek did you try toby-cleveref @ rather than toby @ that's the email address given at dr-qubit.org/cleveref.html
GeekestGeek
GeekestGeek
18:04
I quote (replacing @ and . with words)
<toby-cleveref at dr-qubit dot org>: host mail.protonmail.ch[185.70.42.128] said: 550
5.1.1 <toby-cleveref at dr-qubit dot org>: Recipient address rejected: Address
does not exist (in reply to RCPT TO command)
David Carlisle
David Carlisle
@GeekestGeek hmmm
GeekestGeek
GeekestGeek
Exactly. That's why it seems to me like the developer is not available any longer.
@DavidCarlisle Exactly. That's why it seems to me like the developer is not available any longer. A few minutes ago I re-asked him about tex.stackexchange.com/questions/619875 (to have a more permanent workaround than the suggestions of you, egreg, and Javier). I received no verification e-mail so far.
@DavidCarlisle By the way, in your opinion, is the current TeX Live stable so far for local (non system-wide) Linux installation? Or are there any nasty bugs around?
@DavidCarlisle And I've just received an automated reply from Toby now. This is a new thing.
In short, 1 year or more. :-(((((((((((((((((((
Ulrike Fischer
Ulrike Fischer
@GeekestGeek I got an answer in july after I contacted him through the university (which he didn't like, but as he never answered ...). He wrote "I do still maintain cleveref", but gave no date when he will address the bugs, but only "when I find some spare time".
GeekestGeek
GeekestGeek
@UlrikeFischer Thanks!
Ulrike Fischer
Ulrike Fischer
18:20
@GeekestGeek not sure what for ;-). Basically it means it could be tomorrow or in five years or never ... Perhaps you should try with a real letter.
GeekestGeek
GeekestGeek
@UlrikeFischer Thanks for answering. The automated reply I got and the answer to you together would mean “never” in practice for the folks around me. So, a real letter is a good hint! However, I don't see a real landline address anywhere. Is the university address ucl.ac.uk/quantum/people/dr-toby-cubitt correct (though no office is mentioned)?
Ulrike Fischer
Ulrike Fischer
@GeekestGeek that should work.
Paulo Cereda
Paulo Cereda
@samcarter Hilarious!
David Carlisle
David Carlisle
18:36
@GeekestGeek yes as far as I know (any texlive is really) especially if you are in a situation where you can do tlmgr updates. It's a harder choice for systems like overleaf (or arxiv etc) that take a snapshot of texlive on a specific day and are then more or less stuck with it for a year
GeekestGeek
GeekestGeek
@DavidCarlisle Thanks! My system is Debian stable. I have a choice between taking the version from Debian testing and the current Texlive from TUG.
David Carlisle
David Carlisle
@GeekestGeek most people here are using the version from tug (or compiling from source) but we are tex insiders, the linux-shipped texlives are perfectly fine for the vast majority of uses.
5
Paulo Cereda
Paulo Cereda 
All generalisations are dangerous, including this one.

[paulo@ferrara ~] $ uname -a
Linux ferrara 5.14.14-300.fc35.x86_64 #1 SMP Wed Oct 20 16:14:50 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
@PabloGonzálezL ^^ yay
David Carlisle
David Carlisle
$ uname -a
CYGWIN_NT-10.0 DESKTOP-3D30UL2 3.2.0(0.340/5/3) 2021-03-29 08:42 x86_64 Cygwin
Paulo Cereda
Paulo Cereda
@DavidCarlisle ooh I am a TeX insider too?
David Carlisle
David Carlisle
18:44
@PauloCereda tex insiders need dinner
Paulo Cereda
Paulo Cereda
@DavidCarlisle ooh NT = Not Thy fault
@DavidCarlisle oh no
Joseph Wright
Joseph Wright
@DavidCarlisle, @PauloCereda Looking forward to the first of our important votes this week?
Alan Munn
Alan Munn
Anyone have any experience solving a "your client doesn't have permission" 403 error from Google Scholar when trying to download BibTeX?
Paulo Cereda
Paulo Cereda
@JosephWright Yes sir!
GeekestGeek
GeekestGeek
@DavidCarlisle Thx!
Skillmon
Skillmon
19:12
@JosephWright a friend of mine did something along those lines as well (fitting multiple linear segments). We did that for a simple gesture recognition. He found the algorithm to do that in some book. I can ask him if you want (and probably send you C code if he allows it and still has it)
Joseph Wright
Joseph Wright
@Skillmon I think the 'simple' plan will work well enough, but it can't hurt
samcarter
samcarter
19:38
@PauloCereda Wouldn't Linux ferrari 5.14.14-300.fc35.x86_6 be much faster? :)
Paulo Cereda
Paulo Cereda
@samcarter ooh
@samcarter but that defeats my naming scheme for machines :)
TeXnician
TeXnician
@samcarter Maybe @PauloCereda's Fedora would then even be able to catch up to 5.15 which has been out for ages (speaking as an Arch testing user) :D
samcarter
samcarter
@PauloCereda There seems to be village of the same name near Serino, would that do for your naming scheme?
@TeXnician :)
@TeXnician anything cool in 5.15 we can look forward to?
TeXnician
TeXnician
20:01
@samcarter If you are forced to interact with Windows yes: there is now a kernel-integrated NTFS driver which implements the full spec (in contrast to the fuse-3g ntfs driver).
samcarter
samcarter
@TeXnician Argh, now that I changed the format my last remaining external ntfs hard drive (mac didn't like it) :)
TeXnician
TeXnician
@samcarter Understandable. And Mac won't like it in the future. But Linux is improving. For me it matters as I support some laptops with dual-boot and there better NTFS support (especially concerning crash recovery) is immensely valuable.
Hmm … only one mod is elected for now: opavote.com/results/5149632783974400?
samcarter
samcarter
@TeXnician yes, mac got a lot worse. On the previous machine I was using fuse-3g without problems, but on big sure it is a) painful to install and b) would randomly crash the whole system from time to time.
TeXnician
TeXnician
@samcarter Oh, well, fuse is not exactly a perfect fit for the Mac architecture :D
samcarter
samcarter
@TeXnician There was only one open position, wasn't there?
Joseph Wright
Joseph Wright
20:12
Election result: opavote.com/results/5149632783974400 - @werner is duly elected to serve
@TeXnician Er, yes
TeXnician
TeXnician
@samcarter Ah, yes, sorry, mixed that part up.
David Carlisle
David Carlisle
Congratulations @werner
3
samcarter
samcarter
@TeXnician I think there were more positions in the SO election the week before
Joseph Wright
Joseph Wright
0
Q: 2021 Community Moderator Election Results

VannyModerator election #2 on TeX - LaTeX has come to a close, the votes have been tallied, and the new moderator is: They will be joining the existing crew shortly — please thank them for volunteering and share your assistance and advice with them as they learn the ropes! Also, please join me in tha...

discussion featured election
TeXnician
TeXnician
@samcarter Right, too many elections :D
CarLaTeX
CarLaTeX
20:28
Winner is Werner!!! (Tongue twister)
 
2 hours later…
Joseph Wright
Joseph Wright
22:10
@Werner Been pointed out in the mod chat that you are now the third-highest rep moderator on the entire network
Also, people all over the place know about @egreg - when he gets to a million rep, I think there will be a big party
Paulo Cereda
Paulo Cereda
22:44
@TeXnician you are mean :) 5.15 is in testing (for Fedora 36 and rawhide), so it might reach the F35 channel any day now. :)
@samcarter I made an oath to never use NTFS ever again. It's either EXT4, XFS or BTFS (the first two for external drives).
I keep a FAT USB stick around just to share stuff with my dad's machine. :) Not even ExFAT. :)
@samcarter I had proprietary NTFS and EXT4 drivers (from Paragon) in my old Mac. Even their support was very, very deficient...

« first day (4020 days earlier)      last day (1210 days later) »