TeX, LaTeX and Friends

Paulo Cereda

12:00 AM

@egreg Oh my!

@cmhughes: Chris, things are getting serious! See the TeX gods in action. :)

egreg

@PauloCereda Also yours builds a loooong list of nested \int_eval:n calls

cmhughes

@PauloCereda I've got my popcorn ready :)

Paulo Cereda

@egreg Hey I'm a humble TCS guy. :)

I have no idea of what I did, apart from the syntactic nesting. :)

David Carlisle

@egreg yes I wondered about that but the non tail recursive version allows expressions as wide as the page, and the code looks (and is) shorter.

egreg

@PauloCereda Here's a version which is free from the defect:

\documentclass{article}
\usepackage{xparse}

\ExplSyntaxOn

\cs_new:Npn \paulo_epicrecursion:nn #1 #2
  {
   #1^{
     \int_compare:nTF { #2 > 1 }
      { \paulo_epicrecursion:nx { #1 } { \int_eval:n { #2 - 1 } } }
      { #1 }
   }
 }

\NewDocumentCommand{ \powertower } { m m }
 {
  \paulo_epicrecursion:nn { #1 } { #2 }
 }
\cs_generate_variant:Nn \paulo_epicrecursion:nn { nx }

\ExplSyntaxOff

\begin{document}

$\powertower{x}{10}$

\end{document}

Paulo Cereda

12:04 AM

@egreg Would you like to post an answer? I don't feel comfortable in adding your code to my answer. :)

egreg

@PauloCereda If you add \tl_show:n { #2 } just before #1^ in the body of the definition, you can see what's passed as #2

Paulo Cereda

@egreg Oh. Why one line has :nn and the other has :nx?

egreg

The trick is to fully expand the second argument.

David Carlisle

@PauloCereda "London bridge " not "The London bridge " (just because), you'd have to get our resident linguists to explain why it's London Bridge with no "the" named after "The Tower of London" which has "The".

egreg

@PauloCereda Notice the generated variant. The first call doesn't need the full expansion, but the recursive ones do.

Paulo Cereda

12:07 AM

@egreg Ah I see. Will you post an answer or should I update mine? :)

@DavidCarlisle oopsie, my bad. :)

egreg

@DavidCarlisle The same as "Oxford University" and "the University of Oxford", I guess.

@PauloCereda It's just optimization.

David Carlisle

@egreg yes but I don't understand English:-)

egreg

@PauloCereda Actually the inversion of arguments would be better (fully expanded arguments should come first, according to the guidelines).

Paulo Cereda

@egreg Oh.

egreg

@PauloCereda I had my solution. ;-)

Paulo Cereda

12:22 AM

@egreg Very nice! :) I was quite curious how to deal with the group thingy. :)

@egreg: in other news, beware of David, apparently he's practicing his Italian: tex.stackexchange.com/questions/144446/…

egreg

@PauloCereda ^\bgroup x\egroup is legal.

Paulo Cereda

@egreg I had no idea. :)

egreg

@PauloCereda You never studied how ' for ^{\prime} and '' for ^{\prime\prime} works. :P

Paulo Cereda

@egreg Nope. :P

@egreg: by the way, tomorrow I'll be in São Paulo. :) I'll miss you guys. :)

egreg

@PauloCereda I added a picture of \powertower{x}{254} which is the maximum without changing the internal parameters.

Paulo Cereda

12:35 AM

@egreg OH MY.

egreg

@PauloCereda Actually it's 253, but don't tell anybody.

Paulo Cereda

@egreg :)

egreg

@PauloCereda With lualatex it's possible to do \powertower{x}{499}; at 500 the semantic nest size is overflown.

Paulo Cereda

@egreg With the same code?

egreg

@PauloCereda Yes.

Paulo Cereda

12:41 AM

@egreg Wow.

egreg

@PauloCereda Have a nice Pokémon tomorrow.

Paulo Cereda

@egreg It's the only time I have to play. :) Not during classes though. :)

I'll try to go online for some minutes tomorrow. :)

David Carlisle

@egreg, Not all solutions have the same limitation as your feeble attempt, picture mode strikes again:

\documentclass{article}


\begin{document}


\tiny
\hspace*{-2in}\begin{picture}(0,500)
\put(0,0){$x$}
\multiput(4,5)(4,3){500}{$\scriptscriptstyle x$}
\end{picture}
\end{document}

egreg

@DavidCarlisle :)

Paulo Cereda

@egreg power tower wars. :)

egreg

12:49 AM

@PauloCereda The elegance of the \prg_replicate:nn solution is undisputable.

Paulo Cereda

@egreg It's very clever. :) One for opening the sequence, the other for closing it. :)

Marienplatz

7:27 AM

How to make the following code work?

\documentclass[preview,border=12pt]{standalone}
\usepackage{pgffor}
\begin{document}
$\foreach \x in {1,...,3}{x^\{}x\foreach \x in {1,...,3}{\}}$
\end{document}

Marienplatz

7:41 AM

@PauloCereda: One of your superheros is jailed (click).

Marco Daniel

@Marienplatz The code works. What do you want to achieve?

Marienplatz

@MarcoDaniel The math must be rendered as $x^{x^{x^{x}}}$

Marco Daniel

8:19 AM

@Marienplatz The following can be simplified but it shows the concept.

\documentclass[preview,border=12pt,varwidth]{standalone}
\usepackage{pgffor}
\usepackage{xparse}
\newcommand\XOR{\mathbin{\char`\^}}
\begin{document}
$\foreach \x in {1,...,3}{x\XOR\{}x\foreach \x in {1,...,3}{{\}}}$

\ExplSyntaxOn
\tl_set:Nn \g_tmpa_tl {}

\foreach \x in {1,...,3}
 {
   \tl_gput_right:Nn \g_tmpa_tl { x ^\bgroup }
 }
   \tl_gput_right:Nn \g_tmpa_tl { x }
\foreach \x in {1,...,3}
 {
   \tl_gput_right:Nn \g_tmpa_tl { \egroup }
 }
$\tl_use:N \g_tmpa_tl $
\ExplSyntaxOff
\end{document}

Marienplatz

@MarcoDaniel OK. Thank you for the nice code!

David Carlisle

@MarcoDaniel still suffers from the same problem a egreg's version unlike my picture mode version above:-)

Marco Daniel

@DavidCarlisle There is a similar issue?

egreg

@MarcoDaniel When doing $x^{x^{x^{x^x}}}$, before TeX sees the first { it has opened four groups (one for the math formula); there's a limit of 255 group depth.

Marco Daniel

8:38 AM

@egreg Indeed.

But who needs 255 groups ;-)

@JosephWright: The video of Arthur Reutenauer has no slides :-( vimeo.com/79040764

Joseph Wright

9:05 AM

@MarcoDaniel No, there were none to add

@MarcoDaniel He did a live demo and I can't really reproduce those

Marco Daniel

@JosephWright Well, this is really difficult.

Thanks

Joseph Wright

@MarcoDaniel I have a similar issue with Simon Dales talk, but he had a load of 'ready to go' TeX files I'll use as slides

@MarcoDaniel I'll look another year to get a splitter for the video feed, then I'll be able to grab there. More complex for the recording, though.

Marco Daniel

@JosephWright You need an award for your effort.

Joseph Wright

@MarcoDaniel :-)

@MarcoDaniel I hope the recordings are useful

I'm about halfway through the fifth recording, which does have slides but lots of them!

Marco Daniel

@JosephWright Yeah next video ;-)

David Carlisle

9:28 AM

@MarcoDaniel my picture mode version can do 1000 or more :-)

@JosephWright a live demo in vim :(

@PauloCereda ^^^^

Marco Daniel

11:26 AM

@egreg: Do you have the same issue?

1

Q: TexLive installation unable to find system fonts in OSX Mavericks

It seems my TeX system on OSX Mavericks is no longer able to find my system fonts. My tex documents no longer compile, and when I do: $ xetex opentype-info.tex I get: This is XeTeX, Version 3.1415926-2.5-0.9999.3 (TeX Live 2013) restricted \write18 enabled. entering extended mode (/usr/loca...

lockstep

Anyone prepared a showcase "question no. 50,000"? @egreg? ;-)

4

egreg

12:16 PM

@MarcoDaniel Yes, if the font is not installed in the system fonts. See my comment.

David Carlisle

1:19 PM

most well deserved badge ever?

@lockstep no

Canageek

5:53 PM

Huh, no one has answered my silly, spur of the moment question

8

Q: How do I put a die roll at the top of each page?

I was thinking about a trick I heard of in old RPG books and wondered how you would do it in LaTeX. So it used to be hard to find RPG dice in stores (That is, dice with 4, 6, 8, 10, 12 and 20 sides. Note that all of these but the d10 are platonic solids), so that one book put a series of dice at...

page-numbering random-numbers fun

Also! Whoohooo! It only took a massive bribe, but I finally got an answer to my question on Boardgames.SX

7

Q: What edition of Empire Builder do I have?

My family has an old edition of Empire Builder by Mayfair games, and the box I have isn't listed on Boardgame Geek's edition list. I found an image of the box on that site. It is pre-fourth edition, as it doesn't have Mexico in it, and very little of Canada. The rulebook lists a copyright of 19...

crayon-rail-games

Sean Allred

6:25 PM

>50,000… crazy

David Carlisle

8:42 PM

@egreg's trying to steal my tick before I've even got it now:-)

egreg

@DavidCarlisle I'm surprised you didn't propose a simple \ialign.

But, of course, U+0301 is better.

Torbjørn T.

8:59 PM

I wonder if tex.stackexchange.com/questions/144647/… (or very similar) has been asked before, does anyone know of a duplicate?

David Carlisle

@egreg well I was trying just to correct the coding starting from where it was:-)

rm -rf

Does anyone have suggestions for a serif font other than CM that goes well with the default sans-serif font?

egreg

9:18 PM

@DavidCarlisle That code was hopeless.

Joseph Wright

Fifth video being uploaded to Vimeo, and I've got a post on the UK-TUG website including video links and photos: uk.tug.org/2013/11/12/…

David Carlisle

9:32 PM

@egreg ah so cruel

hhh

$$f(x_1,\cdots, x_j,\cdots,x_n)_{x_j=h}=\{ h \alpha_{J^{-k}} \prod x_j | J\in\mathcal J, j\in J, j\neq h \ | J\in\mathcal J, j\not \in J, h=1\}$$

How would you rewrite this nicely?

tohecz

9:49 PM

\[
  f(x_1,\cdots, x_j,\cdots,x_n)_{x_j=h}=
  \bigl\{{\textstyle h \alpha_{J^{-k}} \prod x_j \bigm| J\in\mathcal J, j\in J, j\neq h \bigm| J\in\mathcal J, j\not \in J, h=1 \bigr\}
\]

@hhh just I'm not sure what are two vertical bars supposed to mean in a set declaration

hhh

This better?

I think it is.

@tohecz I meant the above thing with the vertical bars.

$$f(x_1,\cdots, h,\cdots,x_n)=
\begin{cases}
h \alpha_{J^{-k}} \prod x_j & J\in\mathcal J, \; j\in J, \; j\neq h \\
\alpha_{J} \prod x_j & J\in\mathcal J, \; j\not \in J
\end{cases}$$

I think this is now enough readable, better than the first one with a long long line...

tohecz

@hhh oh you mean it this way, then the first one is completely non-sense.

egreg

10:36 PM

@hhh Don't use $$ in LaTeX. Never ever.

hhh

10:49 PM

?

What is the symbol for set union without the set-intersection?

David Carlisle

11:00 PM

@egreg do you know a canonical answer for [ in array/tabular? search failed me for once:-)

egreg

@DavidCarlisle There should be one. ;-)

13

Q: "[" as first character in table row

Whenever I try to compile a table that has a "[" as first character in a row it gives me the error: Missing number, treated as zero. Illegal unit of measure Example: \begin{table}[htb] \centering \begin{tabular}{r|r} [some text ] & ... \\ [some more text ] & ... \e...

tables

David Carlisle

@egreg thanks

egreg

@DavidCarlisle However the answers there are bad

David Carlisle

@egreg ah:-)

@egreg how did you find that I couldn't search for [ I think?

egreg

@DavidCarlisle bracket array eqnarray

David Carlisle

11:06 PM

@egreg oh words!

@egreg oh it'll do it is clearly a duplicate question, we could always answer the old question if need be

egreg

@DavidCarlisle Done.

David Carlisle

@egreg voted (I suppose you'll get a necromancer shortly:-)

egreg

@DavidCarlisle Using braces works in most cases, but can have adverse effects in array; on the other hand, \relax has no problem whatsoever. Or can you think of one?

David Carlisle

@egreg short of cheating with redefinitions of \relax seems about right

hhh

11:29 PM

@egreg Can you explain in one sentence in layman words, why not $$x_1$$? I haven't yet understood "why not" hence using all time and my document gets compiled nicely with them...

David Carlisle

169

Q: Why is \[ ... \] preferable to $$ ... $$?

I've heard that you should use \[ ... \] for displayed equations instead of $$ ... $$, but why is that? I'd assumed that it's so that you can more easily tell which are starting and which are ending delimiters, but if I always use a syntax-highlighting text editor, I can see that easily based o...

math-mode equations syntax

hhh

...too long to read, in one sentence?

David Carlisle

@hhh add [fleqn] and nothing happens, and something would happen if you used supported syntax

hhh

?

What is [fleqn]?

David Carlisle

@hhh try reading the linked answers, why ask us to type them in again?

hhh

11:32 PM

Because I tried to read it once once once but never got it.

David Carlisle

@hhh \documentclass[fleqn]{article} compare $$a=b$$ and \[a=b\]

hhh

@DavidCarlisle So what? https://www.writelatex.com/501552dtzczd

$$a=b$$ makes things nicely to the center.

[a=b] not doing it.

David Carlisle

@hhh perhaps you don't use fleqn but the underlying issue is that unlike any command name which latex can redefine as required $$ is fairly unique in TeX as there is basically no way to avoid using that sequence, there is no primitive control sequence that starts display math. So $$ is not supported in latex because it can't be redefined to fit optionally specified behaviour, but conversely it sort of half works, simply because it can not be undefined

@hhh incorrectly to the center as you have specified flush left equations

hhh

Still, I cannot understand the demoralizing the use of $$..$$. I have used for a long time due to its simplicity.

egreg

@hhh Do as you like. But if two LaTeX users with 25+ year experience tell you not to, probably they have a reason.

David Carlisle

11:39 PM

@hhh It's wrong. It is not mentioned anywhere in any official latex documentation.

hhh

If I need one day change my behaviour, it is just a simple SED command looping over the files. Now I adhere to one syntax until I understand "why not" -- I haven't yet understood the reason -- perhaps one day I understand why to use [fleqn], why to use it anyway?

If $$..$$ was really bad, why it is not banned in the distribution, why does it get compiled? I cannot understand...

Good night!

David Carlisle

@hhh Latex is designed so that teh same document markup can be restyled without) having to run an editor over the source, a journal style may require flush left rather than centred equations, if the document is correctly marked up then just specifying the class makes things work, if there is incorrect document markup the latex style mechanism breaks.

@hhh as I said above it only works because it is virtually impossible to stop it doing what it does, you can undefine a command with \let\foo\undefined but you can not redefine $$

@hhh It's not bad it just isn't LaTeX.

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends