TeX, LaTeX and Friends

12:47 AM

@samcarter yes, I plan to do this. I just submitted a small update to CTAN again yesterday with a bug fix and support for the dvips driver, so I was going to hold off for a couple more days. There are a couple of unanswered questions (here and here) which can now be answered.

@samcarter And I think it's worth adding an answer to this question which is I think the main question on this issue and leads to many other questions on using pgfplots as a workaround.

@samcarter And if there's other questions you think could benefit from a new answer or comment, I'm happy to add them (or you can of course).

David Purton

2:30 AM

@samcarter, aaaarrgghhhh. Missing percent sign in all the versions on CTAN. I've uploaded v1.1a which will delay things for another day. Stupid % signs.

boycott.se - yo'

7:58 AM

@DavidPurton expl3, expl3, expl3, expl3,

2

... :)

David Carlisle

8:19 AM

@boycott.se-yo' or @DavidPurton could have read any of egreg's answers. Oh no, silly idea: no one does that...

samcarter

8:53 AM

@DavidPurton I'm looking forward to your answers!

@DavidPurton Maybe you could also add answers to tex.stackexchange.com/questions/380371/bordercolor-in-baposter/… and tex.stackexchange.com/questions/313848/… ?

David Purton

12:41 PM

@DavidCarlisle I think I am missing the joke…

@boycott.se-yo' Yes, I'm doing that for new things. But pgf is not so friendly…

David Carlisle

@DavidPurton sorry! It was a joke at egreg's expense, not yours:-) There is a running assumption that egreg has got all his points by answering questions saying to put a % at end of lines.

David Purton

@DavidCarlisle Ha!

David Carlisle

@DavidPurton tug.org/TUGboat/tb36-2/tb113gregorio.pdf

Paulo Cereda

@DavidCarlisle hm

Also:

Recollections of a spurious space catcher

►

It was not a traitor, it was some duck. :)

David Purton

1:01 PM

@DavidCarlisle I see! At least it's nice to see I'm not the only one to miss % characters…

David Purton

1:19 PM

@samcarter the problem with the beamer questions is that because beamer makes such heavy use of shadings, colour mismatches are almost certain to occur as soon as you load xcolor with dvipsnames since then all named colours are in CMYK, but all existing colours in beamer are RGB.

@samcarter It's all too hard to know what kind of shading to produce unless you pass either rgb or cmyk to xcolor as well. But as soon as you do this, at least some colours will shift colour space and look different to what the user expects.

@samcarter In the end this is because xcolor does simple RGB to CMYK translations, but all our PDF viewers use colour profiles. If users don't really understand colours, profiles, RGB, and CMYK, then they will inevitably be confused by what they see.

@samcarter I can add answers to the beamer questions, but the solution that probably makes most sense is to always load xcolor with the rgb option for slides and not use pgf-cmykshadings at all. Colours like dvipsnames will look different, but at least you'll get consistent colours.

samcarter

@DavidPurton I think it is worth to put an answer. Even if the colours shift (gist.github.com/samcarter/fd0f4edbb614d2bc6e628e8c0ed7dd29) it solves the problems with the inconsistent shading. And if the new colour is not what the user expects, he/she/it can just pick another colour that looks like what they want in the colour space.

David Purton

@samcarter OK! I guess so long as users choose an xcolor model based on what the majority of their colours are actually defined as, you'll get the least unexpected result.

samcarter

1:57 PM

google.com/maps/@40.7087088,-74.0067921,3a,75y,249.17h,72.61t/…

2

^^ where our votes are stored :)

Paulo Cereda

3:22 PM

@cfr cwac! :)

Psmith

'ello, Psmith is now fully operational.

Paulo Cereda

/cat

Psmith

Paulo Cereda

ooh

/duck

Psmith

cis

3:24 PM

/cat

/duck

awwww

3:39 PM

/shutdown

Psmith

Initiating shutdown sequence. Cheerio!

Ulrike Fischer

@PauloCereda you got mail.

marmot

/Paulo

samcarter

4:25 PM

Ulrike Fischer

5:31 PM

@DavidCarlisle I think if we want more accessible pdf we will need ideas to get more or less reliably mathml from tex.

David Carlisle

@UlrikeFischer tell the authors to write in mathml?

Ulrike Fischer

8:38 PM

@DavidCarlisle then we would need to get tex from mathml. Is this easier? ;-).

David Carlisle

@UlrikeFischer much easier, yes.

Ulrike Fischer

@DavidCarlisle OK. So we tell the authors to write in latex, copy in mathjax, get mathml, insert in the document and then we translate back to latex ;-). I want an embedded mathjax library ;-)

David Carlisle

@UlrikeFischer in mathml, being xml, < is start of tag and & is start of entity reference and that's that, no fancy macros making space be ignored and : a letter and other arbitrary syntax changes at any point in the document that make parsing "interesting"

@UlrikeFischer well... converting mathjax from javascript to Lua would not be impossible....

@UlrikeFischer currently the best (math) latex to mathml convertors are mathjax, latexml and tex4ht, so that's javascript perl or tex/C/make/etc, so none are that easily embedded into a simple tex run, so you probably need to accept less good conversion or accept help of an external convertor, in practice.

Ulrike Fischer

@DavidCarlisle no. But would it be possible in a realistic time frame? And by whom.

@DavidCarlisle well an external convertor isn't so bad. people are used to call biber, so why shouldn't they be willing to call some "createmathml" tool too, as long as they don't have to copy&paste the mathml in der their document from some browser or so.

David Carlisle

@UlrikeFischer github.com/zekesonxx/pinecone and similar turn up via google, no idea if they only work with toy examples or if it would really make lua out of a million lines of mathjax....

@UlrikeFischer you can run mathjax on the commandline now (it was always possible, but recent versions have gone to a lot of effort to provide easy supported use with node.js (commandline javascript)

Ulrike Fischer

8:57 PM

@DavidCarlisle probably easier than fighting with a conversion software.

@DavidCarlisle perhaps I should mention on the context site that their math tagging doesn't work and look what happens ;-).

David Carlisle

@UlrikeFischer you could do a reasonable range of simple math just in tex (or lua) code, but of course the problem comes if you want to support arbitrary latex with any number of packages

Ulrike Fischer

9:18 PM

@DavidCarlisle imho a reasonable range is quite ok. E.g. whatever mathjax can handle. If an author really wants more, he must learn to configure the commands like they must be configured e.g. with tex4ht. Nothing will happen at all, if we accept only perfect solutions.

David Carlisle

@UlrikeFischer is "use luatex for tagging" an acceptable policy, or do things have to work in pdftex?

Joseph Wright

@DavidCarlisle You sound like me

David Carlisle

@JosephWright :-)

@JosephWright now you have assimilated @UlrikeFischer into the collective, it won't be long before she sounds the same.

Joseph Wright

@DavidCarlisle She's probably better at ConTeXt than me!! ;)

Ulrike Fischer

@DavidCarlisle imho currently "use luatex" is the policy anyway, unless someone solves the page break problem. And one can always setup things so that they work in pdflatex if people copy & paste the mathml manually ;-)

Joseph Wright

9:23 PM

@UlrikeFischer Ah, blame Frank ...

David Carlisle

@UlrikeFischer I suppose I should look what context does which would at least give a start Is there a simple "get xml from context" example i could run without reading a manual?

Ulrike Fischer

@DavidCarlisle this here creates a subfolder with various xml/html files:

\setupbackend[export=yes] % this is all to activate export!

\starttext
$a=frac{x}{z^2}$
\stoptext

@DavidCarlisle but I have to sleep now. I have a talk tomorrow about all this stuff ... I hope that I manage to be online at the meeting.

Joseph Wright

9:39 PM

@UlrikeFischer Have a lot of fun

David Carlisle

@UlrikeFischer thanks, goodnight

I was going to say Hans' MathML was odd, but then I see @UlrikeFischer neglected \ in \frac :-)

marmot

11:14 PM

@samcarter ^^^^ just as I thought, the other tikzlings have my crystal ball, they are not even shy about it....

samcarter

@marmot Are you sure, no marmots are involved? This is what I see:

marmot

@samcarter ^^^ this is what I got...

@samcarter BTW, why is the marmot in your picture surrounded by 2 snow men and a penguin? I fear for the health of this fellow, (s)he may catch a cold...

samcarter

11:31 PM

@marmot Please don't be worried about catching a cold, he/she/it knows from the crystal ball if it is safe to stand there.

@marmot How many crystal balls did you produce to get a marmot-less sequence this long?

marmot

@samcarter 100, I just copied your code... the probability is ((n_tiklings-1)/n_tikzlings)^N if the code is not biased against marmots. ;-)

samcarter

@marmot Nobody would dare to write code that is biased against marmots - this would cause way too much trouble with the equal tikzlings opportunities officer!

marmot

Now I am relieved. BTW, do you want to propose for https://tex.stackexchange.com/q/456653/121799 this code as a second possibility?
\documentclass[a4paper,10pt]{article}
\usepackage{tikz}
\usetikzlibrary{arrows,calc,shapes,decorations.pathreplacing}
\begin{document}
\begin{tikzpicture}
\node at (0,0) {$H: I \times I \rightarrow G$};
\node[label=below:$z_1$] (x1) at (6,0) {$\bullet$};
\node[label=above:$z_0$] (x0) at (9,4) {$\bullet$};
\node (G) at (9.5,2) {$\subset G$};
\draw (x1.center) to [out=5,in=-90]++(2.8,1.8) to[out=90,in=-95](x0.center);

samcarter

@marmot That said, I did not check how good the randomness of tikz random list items is.

marmot

@samcarter I think it is pretty good. It is not at all unusual to get such a sequence.

samcarter

11:42 PM

@marmot Why don't you add an answer yourself? I'd would certainly upvote it.

marmot

@samcarter Please forget my above code, yours is much better!

@samcarter Because there is a very good answer already.... ;-)

samcarter

@marmot Some time ago I had trouble with random numbers with xelatex, but I don't remember exactly what the problem was.

marmot

@samcarter Yes, and now tikz-feynman does not work any more, at least when one tries to use its automatic layout mechanism, and random numbers seem to play a role there, too.

samcarter

@marmot I like your solution to the rectangle question. Rotating $I \times I$ solves the problem with the width of the segments. Please write an answer!

marmot

@samcarter I'm on my way to basketball now (no, marmots are not too small for that ;-), please feel free to include whatever is useful from my above code.

samcarter

11:48 PM

@marmot Oh, that does not sound good!

marmot

@samcarter Basketball is good. Especially in Southern California, where you can play outdoors all year.

samcarter

@marmot OK, I'll steal the rotation of the text, but not rotating the whole scope - I'll leave this for you to answer. Have fun playing basketball!

manooooh

Hey guys!!

Did anyone notice that the columns of cases environment in MathJax vs. LaTeX? Them are different!!

In MathJax we have at least 3 columns with NO ERRORS... but in LaTeX if we put more than 3 columns it causes ERROR.

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends