Quack! :)

Quack quack quack!

@PauloCereda Lost almost one hour experimenting the video-conference system. I'll have to lecture in two rooms simultaneously: too many students.

@egreg Goodness

Is this even possible? tex.stackexchange.com/questions/332720/…

You could do a tooltip of course, but I'm not sure if that is what is asked for.

@JosephWright Tell it to me!

@JosephWright Of course it's because my Linear Algebra course is too attractive. ;-)

@egreg Proudly made with TeX? :)

@PauloCereda Sure!

@egreg :)

@egreg How many in the class?

@JosephWright More than 270; but our biggest room fits 210/220

@egreg Big, though here our biggest room is I think 400 (not an issue for me: chemistry year 1 is about 90, down to 50-60 by the time I do year 3)

@egreg largest here is 176, the largest class there is for almost 300.

@egreg Holy duck! I've never seem such a big audience!

So my uspace package is submitted to CTAN for review! Yay!

I am no a roll! In few decades I will be like you, guys! :D

@wilx ​­​­​­​­

@PauloCereda egreg's algebra class:

@TorbjørnT. isn't it just what beamer does all the time?

@DavidCarlisle ooh

@DavidCarlisle I first read the question as something like "expand this hidden paragraph and push everything that comes after it further down the page/onto the next page etc.", which would be trickier than the suggestion in one of the linked questions I suppose.

@TorbjørnT. oh push to next page would be tricky but within a page you could just use (or steal bits of) beamer I would guess.

@DavidCarlisle Probably.

@DavidCarlisle Yes?

@wilx hint: you could typeset my comment with your package:-)

@DavidCarlisle Oh! Cool. :D

@DavidCarlisle I usually wear that fancy hat, as you know.

@egreg one you missed:

@DavidCarlisle Damn! I should check SE when lecturing

@egreg it's difficult to get a signal in the middle of the stadium, probably.

@DavidCarlisle Our wireless network is very powerful

@Canageek You can add thick or ultrathick to any \addplot options to make the lines thicker.

@AlanMunn I discovered that @cfr doesn't know about the Constantinopolitanischerdudelsackspfeifenmachersgesellschafft

@AlanMunn @Canageek or line width=100pt for that sake

It's May 1st, 2016, but May 4, 2016 and not May 4th, 2016, right? (I mean, when used as an equivalent of May 22--24, 2016.)

@egreg :) How did this come up in the conversation?

@yo' Personally I never use the 'st' or 'th' for dates at all.

@AlanMunn Use the fork, Luke. :)

@PauloCereda I suspect there's a Star Wars joke I'm missing here. :)

@AlanMunn ok :)

@AlanMunn I was too quiet, I decided to break the ice with some random quote.

@egreg Twain had lots of funny things to say about German. The best one is "I'd rather decline a German beer than decline a German adjective".

@AlanMunn You know Red Dwarf, right? :)

@PauloCereda Yes, but I don't have much patience for it.

@AlanMunn I read some. The description of the "verb-at-end" is hilarious.

@AlanMunn I was going to quote this:

Lister: What is death like?
Rimmer: Death? It's like being on holiday with a bunch of Germans.

:)

@Hans-PeterE.Kristiansen Because we don't close immediately. The OP should have time to respond, whih we give. We don't use "On hold" like other sites on the network.

@JosephWright Has something changed in biblatex such that \parencite inside a \section{} shouldn't work? I'm working on an old document that has that (and used to compile) but no longer does.

@JosephWright False alarm. It's the damn underscores and gb4e \automath. I was just fooled by where the error showed up.

@DavidCarlisle Admit it. You are secretly gathering with LAPACK guys and make fun of the world with drinks in your hands right?

The documentation is really like xii.tex. You have to read six commands at the same time

@percusse lapack doc? this version is better:nag.co.uk/numeric/fl/nagdoc_fl26/html/f07/f07intro.html

@percusse And the names are a tour de force of readablity... :) The person who came up with them must have been Czech.

@AlanMunn lapack names are a bit cryptic, like DGESV that's why we also have the nag names such as F07AAF

@DavidCarlisle Progress! Is F07AAF really the nag name for DGESV?

@AlanMunn yes

@DavidCarlisle Did you use Google Translate for the conversion?

@AlanMunn we do have "long names" (although not so useful in the lapack chapters) but they are not so useful and harder to maintain, there re 1800 routines in the library and thinking of 1800 "natural" names and then thinking of new names if you introduce a newer version of the same functionality (we never re-use names, even going back 40 years) then using a compressed naming scheme has many advantages:-)

@DavidCarlisle True, especially if you have a consistent naming scheme.

@PauloCereda That's why I prefer being on holiday with a bunch of Brazils! Then, what about being at work with, hm :-)

@StefanKottwitz :)

@PauloCereda German's chat messages are spell-checked and fully TeX compilable.

@PauloCereda German's maths: Es sei €f(x)=y^2€.

@DavidCarlisle Petr Olšák will be proud of me

You can use opmac (which I recommend anyway if you insist in using plain TeX)

@egreg I assumed it was wipet then scrolled to the end and found it was you:-)

@egreg how many macros are you allowed to input and still call it plain tex, if I input latex.ltx and amstext.sty would that be allowed?

@PauloCereda ABNT says R$y=x^2R$.

@DavidCarlisle latex.ltx is too large for Petr, I suppose.

@StefanKottwitz ooh :)

@egreg I really didn't try to debug the code in the question:-)

@DavidCarlisle Neither did I. But I'm not at all surprised that \endtesting pops out.

@egreg surprising anything comes out

@DavidCarlisle maybe that's the problem :)

?gesv then ?gesvx and now ?gesvxx is not helping

It's a hopeless cause but I think 40 years is enough to have some changes but who am I to judge?

I'm not suggesting that suddenly Pythonic names to appear BUT! some documentation would be nice.

@percusse Also I have a thesis to write. :)

Currently, it's Klingon with an obscure dialect :)

By the way, I'm always looking at the NAG docs ;)

LAPACK is hopeless netlib.org/lapack/explore-html/d7/d3b/…

@PauloCereda Can you call fortran functions from arara ?

@percusse I think I can make it work. :)

@PauloCereda Oh please don't :) Then Bruno will implement Fortran in TeX

I've been adding some solvers to official SciPy lately. And I need ?gesvx at some point. They told me to use f2py and oh my, that is some cryptic stuff happening there

@David: who's the author of the article class? :)

@egreg A question on font size for presentations?

@Johannes_B You are back? :-D

@StefanKottwitz i have internet again.

@StefanKottwitz After a few days without any internet and helping a friend rebuilding the garden, i decided to watch some silly youtube videos and take care of the LaTeX business tomorrow.

@Johannes_B The whole TeX world was holding the breath! (The non-commercial sites :-) )

@StefanKottwitz :-)

@StefanKottwitz Oh, a new site discovered.

@StefanKottwitz Is that you?

@JosephWright The site was intended for aggregating blog feeds and more, I used it as a quick site review for now.

@PauloCereda in its present form, Johannes mainly (although not so different from Leslie's original of course)

@DavidCarlisle How can I give attribution?

@percusse serious question, do you use pdf or html version?

@PauloCereda ? texdoc article then copy whatever it says as author:-)

@DavidCarlisle I was guessing the answer to that was going to be "me"

@DavidCarlisle ooh you are clever. :)

@David: thanks (you will understand why I asked this in my UK-TUG talk). :)

Damn

thick makes the data line thicker, not the line between the label and the number

The line that points at the data

and I don't even know what those are called to look them up in hte manual

@Canageek The legend?

@Canageek Can you add an example though, works fine when I tried here. Could also be a viewer problem, did you try zooming a bit?

@TorbjørnT. No, um

Draws a number and a line. The line is SUPER thin.

If I add thick it makes the line between data points thicker.

I might be able to whip up a MWE

@Canageek Ah, I though you meant plot lines. My bad.

@TorbjørnT. No worries, my fault.

Any idea how ot make those thciker? Or what they are claled to look it up in the manual?

@Canageek \begin{tikzpicture}[every pin edge/.append style={ultra thick}]

@TorbjørnT. Testing!

Ok, everyone. @TorbjørnT. wins Tex.SX chat. It is over, they have won. @DavidCarlisle comes in second, since he found my problem yesterday.

@Canageek Or remove the append, the default color is grey, if you redefine the style altogether it will be black.

@Canageek disqualified for using tikz, so I win.

@DavidCarlisle Why is that a dq?

@TorbjørnT. This looks pretty much exactly how I want so I'm not messing with it.

@Canageek you are supposed to use picture mode tex.stackexchange.com/questions/114783/…

@DavidCarlisle Can I point that at a CSV file? ;)

@Canageek you can point, I'm not sure it'll follow

@DavidCarlisle So, pgfplots is a front end for pgf, which is a front end for TikZ?

@.@

@Canageek No, TikZ is the frontend to pgf. pgfplots is a plotting package based on pgf, I'm not sure if frontend is the right word though.

@TorbjørnT. So the order is pgfplots > TikZ > pgf?

@TorbjørnT. front end: thing that sits between me and scary, hard to use code and acts as an interface?

@Canageek Well yes, but TIkZ also has plotting features, both very basic and the datavisualization library. pgfplots is a separate project with a narrower scope than TikZ.

@egreg: could you help me? I wanted to have the \author definition, but texdef is being naughty. :(

@DavidCarlisle HTML 100%. And not even using the ToC. Directly from google via LAPACK name + "NAG"

@PauloCereda What document class?

@egreg article

@PauloCereda Here it is

\author:
macro:#1->\gdef \@author {#1}

@egreg Thank you! <3

Turtles all the way down

Don't ask what \@author is

@PauloCereda Did you install a new version of Perl?

@egreg It's Fedora's default, it seems.

@PauloCereda The fix, if you're bold enough, is simple: there are two braces that should become \{; Perl even tells you which ones.

@PauloCereda Homebrew is telling me that libass needs updating. I'm a bit worried about the people who chose the name.

@egreg :D

github.com/libass/libass#information-about-the-ass-format

@egreg: I cloned texdef and ran it. :)

@egreg ooh ass information. :)

@TorbjørnT. Limited front end then? Subset?

@PauloCereda It remembers me when I advised the people at the university's computing service not to use the domain name stud.univr.it for the students' email addresses.

@egreg LOL

@PauloCereda They went for studenti

@percusse the pdf cost us a lot of effort, and browser rendering is a lot better than it was so considering not making them, but hard to know whet people use in practice

@DavidCarlisle I don't think anybody will download a catalogue of PDFs and browse through the names and open 14 of them

But I have 17 tabs as we speak open LAPACK/NAG pages

@percusse no but people who buy the library get the html and pdf, and it's hard to know what they use:-0

Most probably html for reference material and PDF for archive.

@Canageek If you really need to give a name to it, I suppose limited frontend might fit.

@DavidCarlisle To be honest I don't know if anybody ever needed the last bits about Program Text and Result parts. For me, it never clarifies anything.

It's a self-proclaimed prophecy :)

@percusse yes but putting the (wo)man hours we do into something that's only archived and not read is hard to justify.

@percusse oh our suspicion is that many people mainly read that (that is they just take the example program and modify, and skip the actual documentation text)

@DavidCarlisle Oh I think you can safely ditch the PDF part for what its worth. But I don't know your customer portfolio.

@DavidCarlisle That's really interesting.

@percusse that's my suspicion as well

@percusse if you install the library and run the e04ucf optimisation example and get 17 and a bit as the answer, it confirms the installation worked, which is another common use (you wouldn't necessarily need 1800 example programs just for that use, I admit)

@DavidCarlisle Oh I never used the NAG library unfortunately. But I'm using it as the unofficial LAPACK docs :)

I don't have the license

@percusse so I guessed but that's Ok. Incidentally if you are working on python interfaces for lapack you could drop me an email, I'd be interested:-)

@DavidCarlisle Thank you, very nice of you. Actually, that's easy :) Here is the SciPy signature file github.com/scipy/scipy/blob/master/scipy/linalg/flapack.pyf.src that we type in for every LAPACK routine for exposing in SciPy library.

The signature file is for something called f2py --> sysbio.ioc.ee/projects/f2py2e/usersguide

It is really arcane but that's what they are stuck with.

It emulates a F90-ish syntax but really driving me mad :)

@percusse when you say you type them in, you mean you do those by hand? but anyway well yes but we have some thoughts on generating more natural interfaces than a generic generator like f2py can do, but not here:-)

@DavidCarlisle Yes really typing by hand :D

@DavidCarlisle If I get a bit better I really would like to drop you an email with meaningful sentences, thanks

Especially the lwork definitions are making me crazy

@percusse behind our doc is an xml description of every routine and we generate for example the entire matlab toolbox interface just from the fortran library doc, we have some thoughts of doing something similar for python (our xml encodes enough to generate workspace allocations for all of lapack)

@DavidCarlisle That would be amazing.

@percusse it's why nag employ an xml geek:-)

For example, to see the scope of work to add one additional ?gesvd functionality this is how much you have to fiddle github.com/scipy/scipy/pull/5994/files

@DavidCarlisle ahaha makes perfect sense now

@percusse I'm sure we could generate most that from existing xml data

@DavidCarlisle Probably so. I can't judge. I'm very new at this backend. I actually needed all this to avoid SLICOT :D Found myself committing Riccati Solvers to SciPy eheh