TeX, LaTeX and Friends

@JosephWright also, needed to say, two of the top10 are SO-lang.

@yo' Yes, that's true

@yo' I've been looking as the mod team had a mail from the Powers: just a check we are all getting on OK

@JosephWright you mean the mod team of all sites I suppose?

@yo' No, I mean the three of us here :-)

@JosephWright ah ok, wow, you're so important <3

9:06 AM

@yo' Like I said, top ten by questions per day

@JosephWright IMHO a healthy site, just a bit misbehaving by SE standards :-)

@yo' Indeed

9:25 AM

An OT question: Anyone know any tool that can handle multi-file PDFs (does not know the correct term) other than Adobe Reader (which cannot handle it correctly on Linux, it seems).

In this case these PDFs are generated by our administration, containing the agenda for a meeting as the main document, and then several extra PDFs embedded into the PDF somehow.

@daleif I didn't even know this is possible!

Answer to my own question: pdftk file.pdf unpack_files output target_dir/ where the target dir probably have to exist in advance.

@yo' apparently they are attachments. PDFtk can attach them and unpack them

9:41 AM

@daleif ah

10:20 AM

@UlrikeFischer ooh :)

10:41 AM

Good maen

@ChristianHupfer 'ello!

@PauloCereda Da fehlt das(!) 'h' :D

@PauloCereda Eating Black Forest ~~cookies~~ bacon and garlic cheese ... delicious

@ChristianHupfer oooh das bakon!

11:01 AM

@JosephWright They always say the sites are community driven.

11:18 AM

I am eating a muffin, yay!

@PauloCereda aw :) I was eating a cheesecake yesterday :-)

@yo' ooh :)

@egreg Indeed

12:40 PM

What am I doing wrong here

\documentclass[a4paper]{memoir}
\usepackage{tikz}
\begin{document}
\begin{tikzpicture}
\draw (0,0) -- ++ (0,8);
\coordinate (B) at (0,6);
\begin{scope}[shift={(-0.35,0)}]
\draw (B) rectangle (0.7,0);
\end{scope}
\end{tikzpicture}
\end{document}

I was expecting the entire rectangle to be moved to the left, why is is only the right edge moving?

I'm missing something

@daleif I think that named coordinates don't shift in general

@yo' ahh, of course. I've been hit by that before.

some things are just so much easier in metapost

@daleif but I'm now not understanding another example. Ah no, everything is ok

@yo' Do you know how to solve this then: \draw (B) --++ (1,0) -- <projection to the x-axis> --++ ....

@daleif that's a bit of a problem

12:47 PM

I do not what to define a coordinate at the point generated by --++ (1,0)

@daleif I understand that

If I had a coordinate name shouldn't it be possible to use ((C) |- (1,0))? Thus I basically need the internal name for the current point.

how does the full path look like?

\documentclass[a4paper]{memoir}
\usepackage{tikz}
\usetikzlibrary{calc,intersections}
\begin{document}
\begin{tikzpicture}
\draw (0,0) -- ++ (0,8);
\coordinate (B) at (0,6);
\draw let \p1 = (B) in
	(B) -| (\x1+1cm,0) -- (4,4);
\end{tikzpicture}
\end{document}

\draw let \p1 = (B) in
(B) -| (\x1+1cm,0) --++ (-2,0)
|- (B)
;

would work, thanks. Have some images to replicate. That slice is going to move in three different images, so I'd like to make sure I had easy control

@daleif you're welcome. Btw, I think it's less confusing to write -- ++(-2,0) than --++ (-2,0), but it may be just a matter of taste.

12:59 PM

@yo' I agree

@daleif ;)

Anyone but me have loading problems in FaceBook? Each time I load it in my Firefox I get a iresponsive script error and the browser freezes until that error message appears.

@daleif works well to me, but FCB has probably hundreds of nodes to meet the traffic needs

@yo' It is an akami script. It seems to eat up memory (fan of my PC starts) and then FF kills the script.

And then FB sort of works until next time I reload

@daleif To many parentheses. \draw (0,0) -- (C |- 1,0); should work.

1:06 PM

2

:)

@UlrikeFischer But in this case I do not have a coordinate for C. Isn't there an internal name for the current point?

@daleif well:

\documentclass[a4paper]{memoir}
\usepackage{tikz}
\usetikzlibrary{calc,intersections}
\begin{document}
\begin{tikzpicture}
\draw (0,0) -- ++ (0,8);
\coordinate (B) at (0,6);
\draw let \p1 = (B) in
	(B) -- ++(1,0) coordinate (temp) -- (temp|-0,0) -- (4,4);
\end{tikzpicture}
\end{document}

Remember that you can use the same name temp again and again in different paths (and maybe even in the same path)

@yo' I said without having to define an extra coordinate.

@daleif well, it is almost a philosophical question whether you're really defining it if the definition is inside the draw command and you use it only once and only in the same draw command :-)

@PauloCereda s/facebook/stackexchange/

1:12 PM

@DavidCarlisle well, SE at least sells the content you create, not youself directly :-)

@DavidCarlisle That too. :) And s/pigs/ducks/g

@PauloCereda but google translate is of course a completely altruistic service for the good of the world.

@yo' how very strange. That does not work in my real life doc. I seem to have all relevant libs loaded. Hmmm

Ahh, error 40

@daleif and the second one (with the "defined" coordinate?)

@yo' haven't tested it. The dual |- is shorter ;-)

1:16 PM

@DavidCarlisle Of course. :)

@daleif but much less clear. The let construction is really obscure; you're basically defining a new coordinate called \p1, whose x and y coordinates can be accessed separately.

@DavidCarlisle and is really bad a translating Finnish Swedish (the fins have two languages)

@yo' it is not that bad, at least one can quickly guess what (\x1+1cm,0) means, whereas (A|-B) might be a bit obscure

@PauloCereda Apenas um geek paranóico iria sugerir o contrário

@daleif You get used to (A|-B) very quickly: you use the vertical coordinate from the one that is closer to | :) What I particularly dislike about \x1 is that you have to add 1cm and not 1 --- this is very confusing inside TikZ.

@yo' that is true, \x1+1cm is weird

1:23 PM

@daleif because this interface is indended for much more complicated stuff, it's really in the pgf layer rather than the tikz one

(brb, snack time and coffee maybe)

@yo', this might be simpler to understand then

\draw[orange] (B) -- ++(0.35,0) coordinate (temp) -- (temp |- 0,0)
-- ++(-0.7,0) |- (B);

@DavidCarlisle Wow! Perfect! 10/10!

@daleif I believe that this is the way to go, yes.

@daleif btw, is this what you want? \draw[orange] ($(B)+(0.35,0)$) rectangle ($(B|-0,0)+(-0.35,0)$);

Either way, use cycle in the end, not (B);

@yo' Hehe, that's even shorter. It is only going to be drawn, not filled. So cycle in not that necessary.

If this was metapost, then I could now save this path and reuse it in another image, shifted if needed.

1:55 PM

@PauloCereda naturally

2:16 PM

Just what I've always wanted: "access to site analytics"

\documentclass{article}

\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{algorithm}
\usepackage{algpseudocode}
\usepackage{multicol}
\algblockdefx[AdaptiveFunction]{AdaptFunction}{EndAdaptFunction}[1]{\textbf{adapt function} #1}{\textbf{end adapt function}}

\begin{document}

\begin{algorithm}[!htbp]
\caption{Adapt function $\mathcal{A}(p_1, p_2)$}
\begin{algorithmic}[0]
\AdaptFunction $\mathcal{A}(p_1, p_2)$
\State\textbf{variables:} $?x, ?y$
\State\textbf{generators:} $g^\star_1, g^\star_2$

Friends, is there a way to remove the space multicols adds between beginning and end of the block?

@PauloCereda \multicolsep=0pt

@UlrikeFischer Dankeschön! <3

It works like a charm!

2:42 PM

As I showed in the image in my example it seems to work with \hphantom -- unfortunately, the censorship seems to be on Germany '30 level. [1] [1]: i.stack.imgur.com/ixVfc.png — harry haller 3 mins ago

Oh my ...

2:59 PM

@ChristianHupfer not clear to me why the answer was deleted, seems a working answer (not the best markup, but why delete?).

@DavidCarlisle: I am not sure that this will work always. At the time of deletion there was no screen shot.

Schogetten!

Yaaaay!

@PauloCereda Oh! It's been years since I had those! Me wants! :)

@clemens It might take a while until it arrives. :)

@DavidCarlisle @ChristianHupfer I guess the real answer will be “use itemize” (perhaps in some modified enumitem way…)

@PauloCereda :D

3:07 PM

@clemens Yes, I assume this too. As soon as the label font etc. is changed I think \hphantom{...} doesn't provide the right spacing anylonger

3:19 PM

@ChristianHupfer If you are forcing line breaks anyway with \\ then it'll work well enough (as all inter-word space will be same width) But even if that wasn't the case I think it's an answer even if not thw best (there are certainly more incorrect answers to questions that have not been deleted)

@DavidCarlisle: I agree with the sentence in parentheses ;-)

@ChristianHupfer Sie sollten mit allen meinen Aussagen zustimmen.

@DavidCarlisle Essential information!

@DavidCarlisle I learned that there is less trafic on weekends -- would never have guessed it ;-).

@UlrikeFischer probably if I looked I'd find a disproportionate amount of traffic came from Italy

4:01 PM

@UlrikeFischer Site traffic or road traffic?

Torbjørn T.

4:48 PM

@JosephWright tex.stackexchange.com/help/privileges/site-analytics --> tex.stackexchange.com/site-analytics I guess.

@TorbjørnT.: A new milestone priviledge?

Torbjørn T.

@ChristianHupfer Seems like, for those with more than 25k rep.

@TorbjørnT. ah

@TorbjørnT. New one on me

@TorbjørnT. Never been a problem here: I've been able to see this since I was at (I think) 3k

@UlrikeFischer Stats are interesting, certainly: give you an idea of the reach of the site

@TorbjørnT. Just noticed by your link.

Torbjørn T.

@JosephWright Perks of being a mod. Us mortals got it today it seems.

Q: 'Missing \begin{document}' when using the \lowercase trick

5:00 PM

@TorbjørnT. Hmm, I get the notice of getting this privilege too: odd

@TorbjørnT. Ah, the two views aren't quite the same

@TorbjørnT. I get a lot more stats via the mod link :-)

2

Mostly for fun, I'm trying to make % comments manipulatable using the \lowercase trick. For some reason, I'm getting a strange (to me) error: ERROR: LaTeX Error: Missing \begin{document}. --- TeX said --- See the LaTeX manual or LaTeX Companion for explanation. Type H <return> for immediate...

Funny how things come up: I updated the behaviour in this area only yesterday

In the same way, @DavidCarlisle's been sorting a LuaTeX byte code allocator, then someone asks for one on the LuaTeX list :-)

@JosephWright :)

@clemens Well when I say 'yesterday', I mean that's when I sent the code to CTAN ;-)

@JosephWright sure: this is when the update happens for the rest of us :)

percusse

Does anyone remember what happened on 19 August of 2013?

Looks like Stackoverflow went on a field trip over here.

5:32 PM

@percusse I think that 464 downvotes in a single day is an absurd figure for our site. There would be 1135 in the days 16 to 24. Possibly a spam attack.

Math people, if S is a sequence, is it OK to say the first element is denoted by S_1 or should I go with a projection function \pi_1(S)?

@PauloCereda Just set out the conventions you use. This one is quite common.

@egreg Thanks. :) You mean the first one?

On the other hand, sequences should start at 0. ;-)

@PauloCereda Yes, S_1

@egreg Grazie. :)

<3

6:23 PM

@egreg Looks that way: big spike in some other data too

7:20 PM

@JosephWright in flags?

(good evening, btw)

@yo' Deleted posts

@yo' Hello

@JosephWright ah, ok

7:37 PM

@JosephWright: Is there something in expl3 with which I could easily test if a word contains an active char? I would like to reject things like \label{vollständig}.

@UlrikeFischer Not at present (and I think unlikely): I'd just \tl_to_str:n :-)

@JosephWright How could \tl_to_str:n work for a label-argument? The thing is then in the aux-file and still gives error when the aux-file is read the next time.

@UlrikeFischer I'm thinking 'future format' of course :-)

@UlrikeFischer it's probably easier just to make it work than test for it and reject it

@DavidCarlisle Well the labels are generated automatically from file names - and sometimes people forget not to use non-ascii-chars in file names. I simply want a bit meaningful error in such cases.

7:48 PM

@JosephWright Is there any str_to_base64 ? :D

@UlrikeFischer May be detokenize the tokens and search inside for one of the four “first bytes” of utf8 (I think there are four, so only four \tl_if_in:). Or may be redefine the way active chars work to get \protected and stop expansion? (I'm not sure if this last one would work)

@Manuel this is still problematic with lua/xelatex

@yo' but there isn't a problem there, the characters should just work

@DavidCarlisle but you're being a bit inconsistent, so it depends on what do you want...

@yo' There's \pdftex_if_engine:T { .. } for such cases.

7:54 PM

this works, patching the definitions of \label and \ref to add \detokenize automatically I leave to the reader:-)

\documentclass{article}
\usepackage[utf8]{inputenc}
\begin{document}
\section{aaa}

\label{\detokenize{\detokenize{vollständig}}}

and \ref{\detokenize{vollständig}}

\end{document}

@DavidCarlisle why twice \detokenize? So that one of them gets written in teh auxz?

@Manuel not always 4, can be 1,2,or 3

@yo' yes (or to put it another way, 1 doesn't work:-)

@DavidCarlisle I mean there are 4 starting points, the ones that are made active, and that have one, two or three arguments respectively.

@Manuel well four ranges, any byte with top two bits 11 is a start byte

@DavidCarlisle Okey, I'm not sure I understand.. but there are more than four to check, then, right?

7:58 PM

@Manuel 64

@DavidCarlisle but you could also test for \unicodeoctet@iii@orhowsitcalled after 1 expansion, for instance by locally letting the 4 important macros to the same thing. This way it's at least only 4 patches rather than 64

@yo' I was going to say that: brilliant bit of code adds those macros, I wonder who wrote it?

3

@DavidCarlisle Well, that's a bit more complicated :) But not too much.

@yo' Good :) In that case this may be works? Make those four \protected (locally), then do \edef, then do \detokenize and then check for the four whateveroctects@iii?

@DavidCarlisle is there an SE badge for having @David agree with me? :)

@yo' no

@Manuel yes but simpler just to detokenize it all and not worry about the encoding as done above

8:02 PM

@Manuel or wait for the L3 team to do that for you :) or simply create a new inputencoding that does the right thing or patch 64 active chars or ...

@yo' Funny you should say that ...

@yo' :)

@JosephWright sorry :-)

@Manuel if doing that way I wouldn't use \protected and checking just locally define (say) the \UTFviii@three@octets by \def\UTFviii@three@octets#1#2#3{\string#1\string#2\string#3}

@yo' I've been worrying quite a bit about 8-bit input for the last few weeks

8:04 PM

@JosephWright yeah I know you are concerned about this (seems the verb can't be used the other way around, right?)

@yo' Goes with the 'job'

@DavidCarlisle Speaking of which, when do you plan to set up a 'team meeting'?

@yo' 'concerned about'

@JosephWright Remember, "we're being paid for features, not for lines of code" :-)

@yo' Indeed

@JosephWright hm forgot all about that, suppose we should....

@JosephWright oh no!

8:06 PM

@yo' Bruno wins either way

@JosephWright ?

@JosephWright so what's a sociable time given the current geographic spread of the team....

@DavidCarlisle End of the month?

@DavidCarlisle Good question!

@yo' l3fp, l3regex, etc. long and full of features

@JosephWright ah, you refer to features, and I wondered about being paid :-)

@JosephWright well, he's incredible :)

@JosephWright something like that.

8:08 PM

@yo' something like that.

oh no, a déjà-vu ...

@Joseph Wright: May I address a TeX and Chemistry related question to you?

@Buttonwood Fire away

@Buttonwood I'd just noticed your 'home' site is chemistry

Right, both chemist and TeX interested.

@Buttonwood Sounds good to me

8:14 PM

Well as you maintain achemso, on one hand, and are chemist, on the other, I was wondering how do you organize chemistry references. How do you organize them -- as organic chemist, we tend to write reaction schemes -- but how to stitch them to Bibtex. Is there something like IsisBase with Tex/BibTex interface?

So that eventually one may draw a reaction that leads to a personally maintained BibTeX entry?

@Buttonwood Not that I know of

@Buttonwood To be honest, I've gone with the advice of a senior colleague. He concluded some time ago that modern (commercial) databases are better than you can hope to do by hand, so it's not worth trying to log 'everything interesting' yourself

Some time ago, there was a feature request in these lines on the project pages of zotero. A tool to collect references (as pdf) and to reference them; with export as BibTeX or BibLaTeX

@Buttonwood My BibTeX database is really just stuff I have references or can be really sure I'll need repeatedly (syntheses, etc.)

@Buttonwood Note I'm more at the organometallic side

Yeah, Pincer complexes

@Buttonwood In the past, yes, though I have a PhD student starting on a pincer project this month :-)

8:18 PM

So there is no "flash card"-something that would bring "organic reactions" and Bibtex together ...

@Buttonwood You could set something up, I guess, but it's not a 1-1 correspondence

One of my supervisors in the past had a ''book of ideas'' where he noticed reactions assumed to be interesting to keep, like the ''Methods of Organic Synthesis'', so I thought, with the advent projects like docear there would be something already done.

@Buttonwood I've not really found any of these 'new' reference things to be that helpful: a pile of printouts and Reaxys for me :-)

One reason to point to docear was their way that "normal BibTeX" (via JabRef) was brought together with mindmaps, that allowed both TeX-formulas, as well as pictures. I tried some time with reactions exported from ChemDraw, yet it was not (yet) as comfortable as ... dreamt.

@Buttonwood I've not come across docear before

8:27 PM

I thought there should be a way to express chemical \emph{reactions} in standardized way, like smiles or the old mdl mol-files, allowing to querry (later): "I vaguely recall a reaction of this type -- do I have something similar in my references?" and to put a (sub) structure in this interface.

@Buttonwood Like I say, the thing is that's already covered by SciFinder/Reaxys, really

Yet I agree, Beilstein & Cie does a good job there.

Changing the subject more to "TeX", I am on the tipping point to go to BibLaTeX, where again you maintain a package that, hopefully allowed to say so, is an offspring of BibTeX's achemso.

Is it right to assume, in the preparation of articles to be submitted to ACS, do you use more frequently the biblatex package?

@Buttonwood Nope :-)

@Buttonwood My involvement in biblatex is somewhat cursory: I provide 'low level TeX' stuff to the current team (after the original author vanished)

@Buttonwood achemso is strictly 'traditional' BibTeX: until recently, the ACS did not have the 'e-TeX extensions' on their servers, which biblatex requires

@Buttonwood I've not to date submitted anything to any journal in TeX format, though my PhD student and I are writing a review at the moment which will go in to the RSC in LaTeX rather than in Word

@Buttonwood Unless you are in a team who all use LaTeX you really do have to stick with Word

@DavidCarlisle I think this will work perfectly. I don't know if I will patch the ref-commands (I could, I have specific commands anyway) but my main aim was anyway to avoid error messages ...

Less non-Latin things can also be problematic. For example, \^w will do OK with OT1/T1 but ŵ is not known to inputenc :(. Similarly for ŷ Ŷ Ŵ :(. [These characters are not, of course, in the OT1/T1 encodings. They can, however, be composed from characters in those encodings which is why \^y will do OK.] — cfr Nov 21 '14 at 0:07

@JosephWright ^^ we could fix that....

8:40 PM

@DavidCarlisle s/we/I/g

@JosephWright s/we could fix that/I could fix that if we didn't complain/

@DavidCarlisle Would make @cfr happy

@DavidCarlisle Sounds good

@JosephWright having added Romanian comma accents seems only fair to support Welsh as well

@JosephWright I overestimated the momentum the big ACS-vessel would implement e-TeX. Ok, they are traditional / very conservative here. I agree (by personal experience, too) to stick to a tool all the collaborators intended are familiar with. Thank you.

@Buttonwood It's not really the ACS as such: they contract the server stuff out, and ti was going to cost to get e-TeX. That's now happened, but the thing with biblatex is that it's actively developed. For an end user that's great, but for a publisher it's not so exciting!

8:47 PM

"Not so exciting" as a synonym to "we want/have to prioritize our ressources on points more attractive/important to us". In the end, they do business. Maybe under an umbrella of 501(c)3.

But the later is speculation.

@Buttonwood Well actually there's a wrinkle here: I make the rules for achemso :-)

@Buttonwood My take was basically 'achemso for submissions is meant to just work as far as possible'

@Buttonwood While 'active' LaTeX users know about biblatex, there are a lot of people using LaTeX with old set ups, models, etc.: ask for example @egreg about some of the 'gems' he sees

Agreed: Better something stable, than fancy. If you have several co-authors, at one point you want/have to submit and don't want to have an additional hurdle.

@Buttonwood Exactly

@Buttonwood Also, a lot of achemso users are on the chem/phys boundary, and so use REVTeX too: it's a natbib-based BibTeX set up (as is achemso)

@JosephWright Looking at the unanswered queue last Saturday, I happened to see an example with \usepackage{newlfont}