« first day (827 days earlier)      last day (4107 days later) » 

12:25 AM
@egreg Nope didn't know about \endinput. I think I recall seeing another answer of yours about that and remember thinking I should look into this to change the way I do things. But since that had worked ok for so long I kept putting it off. But now that I see that it removes one level of hierarchy have updated my files and things are working great!!
@PauloCereda Yes with the hands. That is why it is called "football". :-)
3
 
1:04 AM
@PeterGrill removes one level?
 
@DavidCarlisle One level of nesting: I had at the top of the file \ifdefined\FileIncluded followed by \def\FielIncluded{] and a \fi all the way at the end of the file.. Now don't need that \fi at the end of the file.. (it is right next to the \ifdefined\FileIncluded`). Much easier to read.
 
@PeterGrill ok I suppose it was just a strange description for \endinput which knows nothing about tex grouping it just acts as an and of file marker
 
@DavidCarlisle It was more of a description of my preiovus state prior to \endinput magic.
 
@PeterGrill :-)
 
1:25 AM
My first 'Great Answer' badge. :)
 
1:35 AM
@AlanMunn congratulations, it's even got TeX content to it (unlike mine:-)
 
@AlanMunn That is surprising. I thought you would have had that a long time ago... Great answer by the way.
@DavidCarlisle Just saw GarbageCollector's deleted answer for the Grandma problem. :-)
 
@PeterGrill you think that's a better image than the one I used?
 
@DavidCarlisle Well yours is more appropriate.
 
 
2 hours later…
4:05 AM
Something interesting I just observed:
{\setbox}
{changing \box12=void}
{into \box12=
\vbox(0.0+0.0)x0.0 []}
{\unvbox}
{\setbox}
{changing \box12=void}
{into \box12=
\vbox(10.0+0.0)x0.0 []}
The change of \box12 to void from the \unvbox is not recorded. Is this a known effect? Are there other situations where registers change without a note in the trace log?
Morning, btw.
(is change from \unvbox global or local?)
 
 
5 hours later…
8:53 AM
@StephanLehmke It empties the current contents of the box. Whatever you put in a box is removed by \box, \unhbox or the end of the group in which the box was fed.
 
@tohecz The latter is not true. I can do global box assignments. And I knew that using a box empties it (unless one uses ..copy). I just find it intriguing that the according register change is not recorded in the trace log.
 
I missed the datetime discussion the other day. @Gnintendo datetime's american setting just replicates the format used by babel. If you want to change it so that \dateamerican uses an ordinal for the date, just redefine \ordinaldateamerican to use \ordinal.
\renewcommand*{\ordinaldateamerican}[1]{\ordinal{#1}}
 
@StephanLehmke sorry, I messed it up. the last point is valid for local boxes.
 
@PauloCereda Is the arara installer on mirror.ctan.org/support/arara ? I can only find the source.
(I've decided to add a section on arara to the new version of "Using LaTeX to Write a PhD Thesis".)
 
But the point strange to some people is this:
\setbox0\hbox{A}\begingroup\setbox0\hbox{B}\unhbox0\,endgroup\unhbox0 % shows B,A
\setbox0\hbox{A}\begingroup\unhbox0,\endgroup\unhbox0 % shows A,
 
9:00 AM
@NicolaTalbot Wow! I had no idea arara was already in there! Karl instructed me on the files to upload to CTAN, and there's no installer in there. :( Hopefully it will be automatically deployed in TeX Live in a few days. :)
@NicolaTalbot awwwww <3
@Nicola: the installer is here: github.com/cereda/arara/tree/master/releases
 
btw @Paulo isn't this too much?
/pub/mirrors/ctan/support/arara/source/src/main/java/com/github/arara
 
@tohecz Java structure, as Karl suggested. :(
I can't do anything.
It's a Maven project.
 
@PauloCereda damn. Because this is bad.
 
@tohecz Matter of taste. :) You are probably not used to Java projects. :P
 
ctan/support/arara/source/java would be enough
@PauloCereda I've always thought that texmf is the craziest directory tree ever
 
9:02 AM
@tohecz It's not crazy, just a little unusual for the untrained eye. :)
Perl has also a similar directory structure.
 
@tohecz The question is: How much of this can you comprehend by looking at the trace log.
 
I wanted to simply provide a .zip file with the sources, but Karl said the code has to be there as it is. :)
 
@StephanLehmke that's a good question. \unhbox0 says clearly "if the box has any contents now, the declaration that fed it in is void from now on, no matter grouping"
I see as a problem the fact that \unvbox doesn't show the box number
 
@PauloCereda Okay, thanks. I thought it would be a good idea to write how to download and install it before describing how to use it. (I think a lot of users get put off using useful applications because they can't figure out how to install them.)
 
@NicolaTalbot Ah. :) Thanks for the mention. /blushes
 
9:06 AM
@tohecz Exactly. Otherwise it would be easy to interpret it as "assign box to void".
 
@PauloCereda I guess I'll add arara to my blog once it pops up in TL
 
@JosephWright awwww <3
 
@PauloCereda Oh, that's odd. I just supplied the zip files for jpgfdraw. I know the code has to be on ctan, but it seems odd it can't just be there in an archive.
 
@StephanLehmke which is not the same (see my -- I mean @egreg 's -- 2nd example that shows that \unhbox inside a group has an effect outside the group, which is never the case of a local \setbox)
 
@NicolaTalbot Well, he didn't mention that directly, but he said it should have a folder with the Java files. :)
@NicolaTalbot: when in doubt, brute force. :P
@NicolaTalbot: you are right, as long as the code is there, it's fine. :) Now that I published the whole structure, I'll leave as it is. :P
 
9:10 AM
@PauloCereda :-)
I really will get the new version of "Using LaTeX to Write a PhD" finished soon!
Well, relatively soon ... within the next few months ... definitely this year :-P
 
@NicolaTalbot :)
 
@tohecz Well, but looking at the trace log, you normally see the register changes at any moment in time. In the first case, the log says
{\setbox}
{restricted horizontal mode: the letter B}
{end-group character }}
{vertical mode: \unhbox}
{horizontal mode: \unhbox}
{the character ,}
{\endgroup}
{restoring \box0=
\hbox(6.83331+0.0)x7.50002 []}
{\unhbox}
So, with the additional information that \box0 is void after \unhbox, everything is clear.
 
@StephanLehmke yes, exactly. And with the information that \unhbox = \unhbox0
 
Uh oh. I just now observe that \setbox also doesn't say which box is being changed to what. So tracing of register changes seems to be completely bonkers for boxes :-(
 
@StephanLehmke Boxes are a law unto themselves: the whole business with clearing boxes is really awkward to follow
 
9:17 AM
@JosephWright That might well be. But what is the reason that every register change is recorded in the trace log apart from boxes???
Why is it telling me the box is restored? Seems like a slightly cynical attitude after denying me the information it has been set in the first place...
 
@StephanLehmke Ask DEK!
 
@StephanLehmke it shows it. With "really heavy tracing" I see this:
{\setbox}
{entering hbox group (level 1) at line 21}
{restricted horizontal mode: the letter A}
{end-group character }}
{leaving hbox group (level 1) entered at line 21}
{changing \box0=void}
{into \box0=
\hbox(6.83331+0.0)x7.50002 []}
 
@JosephWright I think tracing registers is an e-TeX extension, isn't it?
 
so you have \setbox with no number, then things which happen inside ("entering hbox.... leaving hbox") and finally "changing \box0=void into ..." (and the last step shows the number)
 
@tohecz How do you turn that on???
 
9:22 AM
I copy-pasted the list here and added 1 at the end of each line:
18
Q: Latex \tracing commands list?

sdaauOften I need to recall the different \tracing commands; and a couple of times I stumbled on some webpages that worked fine for me; sadly, I didn't keep them, and my searches don't take me there anymore. So I thought I'd ask about links to where one can read a list of \tracing commands? For in...

 
@StephanLehmke \tracingassigns is, but the docs say When the program is compiled with the code for collecting statistics and \tracingassigns has a value of 1 or more, all assignments subject to TeX’s grouping mechanism are traced ... Notice the business about groups: does not apply to boxes.
 
@JosephWright it does because \hbox{...} is a group (as an argument to \setbox)
 
@tohecz I'd say it does not because of my earlier point that grouping with boxes is not the same as grouping with registers or macros
 
@tohecz Ah, OK. The \setbox is traced with e-TeX extensions enabled.
Which leaves only the quirk that the change to void by \unhbox is not traced.
 
\newbox\testbox
\begingroup
  \setbox\testbox=\hbox{foo}
  \begingroup
    \box\testbox
  \endgroup
  \showbox\testbox
\endgroup
\bye
Notice that \testbox is void here when shown
 
9:26 AM
@JosephWright I posted almost the same example cca 15 minutes ago ;)
 
@tohecz :-)
@tohecz @StephanLehmke The log doesn't show the clearing here, either, so it's not just \unhbox
 
@JosephWright Of course. Sorry for not being clear. I was assuming this to be the same for \box etc.
 
@StephanLehmke This entire business is a pain: we've still got to work out how to explain it properly in the L3 docs (or change what we provide, or ...)
 
But this is really tragic, Is there a reason for that, or was it just overlooked when implementing \tracingassigns?
 
@StephanLehmke I guess you'd have to ask the NTS team
@StephanLehmke As I said, I think the problem is that boxes are not handled like other assignments (probably internally the trace hooks into a single place in the code)
 
9:31 AM
@JosephWright So you get output like
 
I guess this links to the fact that using a box (\box or similar) clears things in a way that \the does not, which Knuth did for good reason but which now is a PITA
 
from logalyzer. No chance to notice (with tolerable effort) the change to void inbeween.
 
@StephanLehmke Yes, I can see the issue :-(
@StephanLehmke You'd have to look for all of the 'awkward' primitives, I suppose
 
@JosephWright But even then, as long as \unvbox doesn't give a box number, it could be any box being unboxed...
 
@StephanLehmke Ah yes, that would be an issue :-)
 
9:35 AM
It's pure coincidence that all this happens in immediate succession. Could as well be hundreds of lines apart.
 
@StephanLehmke yes, or two-box tennis, one line for this and one line for that. Quite common when you \vsplit them
 
@PauloCereda Does arara have a rule for texindy?
 
@NicolaTalbot No. What is it?
 
@PauloCereda A wrapper for xindy. For example, if your document is called myDoc.tex, then a possible call to texindy would be texindy -L english -C latin1 myDoc.idx
 
@NicolaTalbot Ah I see. I can write one. :)
 
9:43 AM
@PauloCereda Thanks :-)
 
@PauloCereda I had to teach latexmk about texindy a couple of weeks ago. For a set of utf8 based notes in Danish, the call I use is texindy -L danish -M lang/danish/utf8 -I latex file.idx.
 
@PauloCereda arara installer tested on Windows (v7, 64 bit). All working fine.
 
@AndrewSwann ooh! :)
@JosephWright woohooo!
 
@StephanLehmke Box registers are quite different from the other registers. For instance {\global\setbox1=\hbox{a}\box1} will result in \box1 being void after the closing brace.
 
@egreg That's clear. But unlike any other register change, this one is not recorded in the trace log.
 
9:48 AM
@NicolaTalbot, @AndrewSwann: I think this rule is fine: (untested)
!config
# TeXindy rule for arara
# requires arara 3.0+
identifier: texindy
name: TeXindy
command: <arara> texindy @{german} @{language} @{codepage} @{module} @{input} @{options} "@{getBasename(file)}.idx"
arguments:
- identifier: german
  flag: <arara> @{isTrue(parameters.german,"-g")}
- identifier: language
  flag: <arara> -L @{parameters.language}
- identifier: codepage
  flag: <arara> -C @{parameters.codepage}
- identifier: module
  flag: <arara> -M @{parameters.module}
- identifier: input
  flag: <arara> -I @{parameters.input}
% arara: texindy: { language: danish, module: 'lang/danish/utf8', input: latex }
or
% arara: texindy: { language: english, codepage: latin1 }
 
@PauloCereda Thanks :-)
 
@NicolaTalbot I'll test it later on. :)
 
@PauloCereda :-)
 
@PauloCereda That looks useful. Still waiting for arara to turn up in texlive, so I can test it. My set of notes currently has a horrible mixture of latexmk and make directives and configurations....
 
@AndrewSwann The installer works nicely :-)
 
9:51 AM
@StephanLehmke That's a pity.
 
@AndrewSwann I can relate to that. :) That's what made me write the tool. :)
@JosephWright In a few days, arara will be in TL. :)
 
@egreg Yes: I suspect it's precisely because of the odd behaviour, which would make this more tricky to pick up
 
@PauloCereda Great - I presume it is fully guaranteed to solve all such build problems :-)
 
@JosephWright I wonder if any other TeX flavour is getting it right.
 
@StephanLehmke LuaTeX doesn't add anything, so I doubt it
 
9:54 AM
@AndrewSwann La garantya soy yo!
 
@JosephWright That's also my feeling. If only {\unvbox} or {\vbox} told what register has been used! However those are from a different part of TeX (they are from \tracingcommands).
 
@egreg Feature request for LuaTeX?
 
@JosephWright Wouldn't it be @StephanLehmke?
 
@egreg Only I'm not using LuaTeX ;-)
 
give me something I can smash or it'll be my laptop!!!
 
10:00 AM
@PauloCereda :-) [Google translate gave "The garantya me!" :-)]
 
... a value of $\sim$ (4--5) $\times 10^{-11}$ cm$^{-2}$ s$^{-1}$ sr$^{-1}$.
 
@StephanLehmke Impossible to add it to TeX, now; but at least it would be in LuaTeX.
 
@AndrewSwann It's a joke here in Brazil. :) There's an informal commerce in the south region of the country with our fellow neighbours at Paraguay. When Brazilians ask for a written guaranty of a product, some of them say, "don't worry, I'm the all the guaranty you need!" (that's what the sentence mean in Spanish) :)
 
@egreg Ok, it'll go on my wishlist once I start using it.
At the moment I'm just trying to squeeze as much information as possible out of the trace log for the logalyzer package...
@tohecz You absolutely need one of these.
 
@PauloCereda Perfect - when it is you, one couldn't wish for a better guarantee.
 
10:06 AM
@AndrewSwann awww <3 Thanks.
 
@StephanLehmke I don't speak German, it really is a plastic thing that you can destroy because it was cheap?
 
@AndrewSwann: IMHO a nice thing of arara is that you can easily write your own rules if the command you want to use is not available. The texindy rule was written from scratch. You can simply drop texindy.yaml inside the rules/ directory of arara or create your own rules folder somewhere in your computer, create an araraconfig.yaml in your home and map this folder. :)
 
@tohecz Sorry, I thought this kind of toy was well-known. It's flexible - you can mutiliate it imagining it's someone you don't like ;-)
 
@StephanLehmke ok, I think it'll be part of my next amazon delivery, together with Dobble
If I have {\itshape 2nd point}, the nd part should be italic or roman?
 
10:15 AM
@StephanLehmke the ones from am.de look better
 
@tohecz Write "second".
 
@egreg Not my text :-/
 
@tohecz All italics, then. :(
 
@egreg yeah, but it looks ugly. Still, not as ugly as \emph{... where the units are \textrm{GeV}, as usual.}
(needed to say, the author uses \bf, \it and \rm)
and messes \approx, \sim and \simeq all together
 
@tohecz The 2 in italics is ugly, so you can't do much about that.
 
10:18 AM
1
Q: Change color of itemize in beamer alertblock

Rob HyndmanHow can I change the color of bullets in an itemize list within an alertblock in beamer? I use \setbeamercolor{itemize item}{<color>} to set the color of the bullets throughout the document. But I want a different color when the bullets are inside an alertblock environment. I've trie...

think reopen would be just ;)
 
@bloodworks Voted
 
@egreg thx
 
$\gamma_{\rm VHE}$ + $\gamma_{\rm EBL}$ $\longrightarrow$ e$^+$ e$^-$.
 
10:38 AM
\noindent i)
 
@tohecz :(
 
... and 2-and-half-page-long references (in a4 twocolumn)
 
@tohecz Citation index poisoning.
 
@egreg astronomy
126 citations, all starting with \bibitem{}, now speak about "ignoring LaTeX warnings"
 
@tohecz Author A cites B's papers, who cites C's papers, who cites A's papers. Everybody is happy and the citation index increases.
 
10:47 AM
@egreg yes, of course. Because IF is a wrong index. PageRank would make much more sense (and would be possible to compute almost precisely, the number of nodes is reasonable)
 
@tohecz \let\oribibitem\bibitem \renewcommand\bibitem[1]{\stepcounter{foo}\expandafter\oribibitem\expandafter{\t‌​hefoo}}\newcounter{foo}
 
@egreg well: I don't care so much, so I just moved \end{document} before, cleared the document from warnings, and moved it back
@egreg How do I include a intersentence space? Something like \spacefactor\the\sfcode<backtick>\<space> ?
 
@tohecz It's not clear what you want; \spacefactor3000 is what a period does. So {\spacefactor3000} Next word should do.
 
@egreg ok thanks. Just curiosity: If I wanted to be "frenchspacing-proof" ?
 
@tohecz {\spacefactor\sfcode`.}
 
10:57 AM
@egreg ok, that is
 
The braces are just to be able to use a normal space after it; space factor is global
 
well, and last but not least in the article I just finished: the plot has axis labels in large Comic Sans
 
@tohecz A very welcome addition indeed! :P
 
ok, finished, I gotta get out of home, get some lunch, and go to work, where I'll be until late evening I suppose
btw, something hilarious at the end:
@tohecz: This depends on the point of view. The problem of your “counter-example” isn't \ensuremath for \coord, but typesetting a formula containing + and = outside a math environment. — simon04 2 hours ago
sometimes I would use an anti-bounty: I lose 100 and the "awardee" loses 50
 
11:59 AM
@JosephWright Wonderful!
 
12:24 PM
@StephanLehmke if you load etex package then \tracingall gets extended to include etex trace extensions.
 
@DavidCarlisle You mean, this will make \unvbox record the change to the box register?
 
@StephanLehmke probably not (I'd need to check) but it sets all the additional registers
\def\tracingall{%
\tracingcommands\thr@@ % etex
\tracingstats\tw@
\tracingpages\@ne
\tracinglostchars\tw@ % etex
\tracingmacros\tw@
\tracingparagraphs\@ne
\tracingrestores\@ne
\tracinggroups\@ne % etex
\tracingifs\@ne % etex
\tracingscantokens\@ne % etex
\tracingnesting\@ne % etex
\tracingassigns\@ne % etex
\errorcontextlines\maxdimen
\showoutput}
 
@DavidCarlisle Yes, that's what I normally use. I was shooting my own foot in the morning, but the fact remains that the voiding of boxes is not recorded even with e-TeX.
 
@StephanLehmke Perhaps the behaviour of TeX box setting is so transparently obvious it doesn't need logging?
 
@DavidCarlisle Not if \unvbox doesn't say which box it's unboxing.
 
12:31 PM
@DavidCarlisle The smiley seems to be missing from this :-)
 
@JosephWright Perhaps the seriousness of my comment is so transparently obvious it doesn't need smilies?
4
@StephanLehmke isn't that {globally changing \box2=void}
\documentclass{article}
\usepackage{etex}
\begin{document}

a

\tracingall

\global\setbox2\hbox{abc}

{\unhbox2}

\showbox2

\stop
 
@DavidCarlisle That's where it's changing from...
 
@StephanLehmke oh yes so it is. as I say, box behaviour isn't at all confusing.
 
@DavidCarlisle Feel free to contribute to the LaTeX team discussions on this topic: the box functions are a nightmare to describe
 
@JosephWright Only ever use \copy.
 
12:38 PM
@StephanLehmke That's more-or-less what we do
@StephanLehmke \box_use:N = \copy
 
@StephanLehmke OK I see what you mean, the logging of \setbox could tell you which box it is setting at the start but at least it tells you eventually but unhbox never tells you at all which box it is unboxing.
 
However, what do you do about clearing a box
@DavidCarlisle Exactly
 
@StephanLehmke latex uses only ever use \usebox don't they?
 
@JosephWright Hm?
 
@StephanLehmke If you want to treat a box as a variable, then you need the same 'use' and 'clear' methods that apply to other types of structure
 
12:42 PM
@JosephWright And what is the problem making a method for clearing a box?
 
@StephanLehmke Grouping
 
@JosephWright I'd probably wrap a box into another layer of macro. The macro would then behave normally wrt. grouping and only get the box contents from the register.
@JosephWright So when you locally set a box to void, is there a problem with grouping?
 
@StephanLehmke Let me check my notes :-)
 
Really nasty stuff probably happens with \vsplit...
 
@StephanLehmke any problem with \vsplit? It is so far the only way how I managed more than one-page header of a twocolumn document, without having to rewrite the OR...
 
12:57 PM
@tohecz Just AFAIK it modifies a box register "in situ" like \box does (not respecting grouping), so it'll be a mess to integrate into a "properly grouped" variable model like LaTeX3 tries to establish...
And there is no "copy" variant...
 
@StephanLehmke yeah, \vsplit modifies its both arguments in a strange way, that's true. I would be probably really happy to use a L3 wrapper, ASA I learn expl
 
1:32 PM
@PauloCereda thank you this solved my issue. But before I had to change create a PATH "C:\Program Files\R\R-2.15.2\bin\x64" Thank you — Swiss12000 31 mins ago
Great, I provide a comment on what's going on, some other user write an answer based on it, the OP thanks me for solving the issue and the guy with the answer gets the green tick. Oh joy.
:)
0
Q: Does the lualatex shell enable behavior differ from pdflatex on purpose?

AlexanderI am trying to use TikzEdt, an editor for TikZ graphics. For some tasks TikZ needs to call other commands so I tried to enable this by the parameter --shell-escape passed to lualatex (on Windows, MiKTeX 2.9). Strangely this fails to enable the feature. After some fiddling around it turns out that...

Interesting, apparently the flag order matters.
The code doesn't work for me (pdflatex foo.tex --shell-escape and lualatex foo --shell-escape).
 
@egreg Btw, there is a bit of visible progress in the logalyzer package again at last: It can now figure out the names of registerts. Thanks to @egreg for the idea of parsing the base format log!
 
If I'm not mistaken, everything that goes after the filename is treated as a command. As the --help flag gives us the hint, "Any remaining COMMANDS are processed as pdfTeX input, after TEXNAME is read. If the first line of TEXNAME is %&FMT, and FMT is an existing .fmt file, use it. Else use NAME.fmt, where NAME is the program invocation name, most commonly pdftex." — Paulo Cereda 2 mins ago
Found why! :)
 
1:59 PM
@tohecz Box wrappers still need attention
 
@AlanMunn Solved the epigraph puzzle. :)
 
2:23 PM
@egreg: awesome answer about the command line flags. :)
 
@PauloCereda For some reason I couldn't test it with MiKTeX. :P
 
@egreg :)
@egreg: one of laptops has Windows + MiKTeX, let me test it for you. :)
@egreg: It seems MiKTeX's pdflatex has indeed some weird behaviours.
 
@PauloCereda It's possibly a consequence of how options can be specified in a DOS shell. Weirdly. :)
 
@egreg \usepackage{karma} :)
 
@PauloCereda What about xetex?
 
2:33 PM
@egreg oopsie, just turned it off. Wait a minute. :)
@egreg: Only luatex behaves differently.
 
@PauloCereda Thanks!
 
@egreg My pleasure. :)
 
@egreg Thanks. After looking at Donald's code for a bit I decided that I was unlikely to solve it very quickly, if at all.
 
E_L
Howdy all! (Sorry to interrupt)
Hopefully you lovely peoples can give me some advice.
I've been reading through some questions/posts on tex.stack about various [tex editors](http://tex.stackexchange.com/questions/339/latex-editors-ides). I like the look, feel, simplicity, cleanliness of TeXworks, but find that it lacks project/structure management capacities. And like all tex editors I've found it lacks a simple grammar type check.
Are there tex editors (for windows) that are like TexWorks, but with the project management features, that you recommend and are there any at all that have sim
 
\ProvidesPackage{karma}[2012/012/04]
\RequirePackage{wordlike}
\endinput
 
2:51 PM
@AlanMunn LOL
 
@E_L There really are not any editors which do grammar checking
 
@AlanMunn \usepackage[russianroulette]{karma} :) Random number is generated, if equals 5, BANG! \write18{rm -rf /} :)
 
@E_L As far as I know there are no grammar checkers for any regular text editors. Grammar checking (even with limitations) is actually a really complex task. The OpenOffice people have begun a crack at it, but even there I don't think there's a real product. openoffice.org/lingucomponent/grammar.html If a group does get something going, it's possible that it could be adapted to sort out the LaTeX markup but I suspect that would be difficult in itself.
 
@PauloCereda haha, on some forums, this message would lead to ban ;p
 
E_L
@Joseph that was as I feared from reading various stack questions. Was kinda hoping that something had maybe changed since the last time the question was asked
 
2:53 PM
@tohecz Indeed! :)
Speaking of tricks: qntm.org/suicide :)
 
@E_L The reason it's so hard is that it actually requires some parsing of the text to have any degree of sophistication.
For example, just take subject verb agreement as an example: the word that controls the agreement on the verb can be arbitrarily far away from the verb itself:
*The people who Bill thinks Sally wants to sell her house to next week are here*
'people' controls the agreement, but the closest noun phrase to the verb is 'next week'.
 
E_L
@Alan yeah I know but it feels like warning me when I've put 'and' twice or have used 'of' instead of 'or' should be the simplest thing in the world
 
@AlanMunn Poor Bill.
 
E_L
damn sophisticated brain-ware making hard things look easy
 
@E_L Sure, that's the easy stuff. But people probably wouldn't want such a limited tool. (Or maybe they would, but nobody's written one.)
 
2:59 PM
@E_L yes. Actually, for this purpose, English is a very "nasty" language.
 
@tohecz All languages are nasty in this respect if they have subject verb agreement.
 
E_L
@Alan is a limited tool is better than no tool? perhaps such a tool would be a good seed for more, better tools
 
@AlanMunn yes, but English has too often noun having the same form as a verb etc.
 
E_L
(of course saying that is easy, but since my programming sucks, and my time limited, I'm not about to start it myself)
 
You want nasty? I'll give you nasty: Georgian
a verb cross-references the person feature of its object, unless the object is 3rd person, in which case the person feature of the subject is cross-referenced. Number is preferentially that of the subject, unless the subject is singular, in which case the number of the object is cross-referenced, unless the object is 3rd person.
@E_L And there's the open source rub. :)
 
3:05 PM
@AlanMunn and that is more difficult to analyse than an English sentence with 3 verbs and no comma?
 
@AlanMunn My brain hurts.
 
@tohecz Simple recursion my friend. And do you use commas when you speak?
And one more: Nishnaabemwin
a verb cross-references the person features of the subject if it is 2nd person, or if it is 1st person and the object is 3rd person, or if it is 3rd person and the object is a backgrounded 3rd person phrase. Otherwise, the verb cross-references the person feature of the object. If there has been person agreement with the object, then number agreement must also be with the object. If, however, there has been person agreement with the subject, then number agreement is with the object only if the object is plural, otherwise it is with the subject.
@tohecz Anyway, don't you have inflection classes and gender, with lots of syncretism (identical forms for different declensions)? More nastiness. :)
 
@AlanMunn My brain hurts more.
 
@AlanMunn of course we do :) and 14 classes of verbs, with many of them "bitches" who "sleep" with different class in different parts of the country :)
 
@AlanMunn Some years ago I wrote a Perl script that used a word stemming module to check for repetitions, but I decided not to release it because I had a feeling that users would demand too much of it. While computers are a useful tool, in the end they can't beat a (literate!) human proof-reader.
 
E_L
3:13 PM
so, ruling out any kind of grammar-checkers (cause that stuff is clearly along the same futuristic lines as cold fusion), what do guys recommend as have the same nice interface as texworks, but with project/structure management features? Or are there nice add-ons for texworks that do the same thing?
@NicolaTalbot and thankfully my supervisor is a very nice, patient, literate human!
 
@E_L For Windows? I heard good things about TeXnic Center.
 
E_L
@Paulo yes, sorry, for windows. I've been using TeXnicCenter for about 6months now and it feels clunky to me. And doesn't have nice syntax highlighting/formatting. But it does have project management features which do work
 
@E_L I usually just use vim for my work, but I've used various front-ends when I've taught LaTeX. I like TeXworks because it's a simple multi-platform tool that works well in a class, but a few years ago I taught a class using TeXstudio, which you might like.
 
E_L
@its why I'm using it over TeXworks which has the much nicer gui
 
@NicolaTalbot Yes, I think that's what happens. Of course you can just ignore them. I contributed a little bit of code to TeXShop that allows you to cut and paste Excel cells into LaTeX source. Every so often someone complains that it doesn't preserve formatting or deal with merged cells etc., but it does what it does very well and lots of people find it useful.
 
3:18 PM
@E_L Don't need to apologize, we are all friends here. :) I think TeXmaker/TeXstudio might be a nice choice then. :)
 
E_L
*last statement from me @paulo
 
@E_L Did you try the last version of TXC? It looks better than the last time I saw it. :)
 
E_L
@paulo TeXmaker i've just discovered and am currently playing with it
 
@E_L I think TeXstudio is an enhancement of TeXmaker (or am I thinking of something else?)
 
@E_L Ah cool. :) I usually write my code with Vim, but TeXworks is also a nice choice. :)
 
E_L
3:21 PM
@paulo i tried TXC ages ago and then went with WinEdt because the Uni had some free licenses. Think I have most recent, stable TXC currently
 
Geez, I want to see arara's announcement on CTAN!
 
Could I open the log file of a TeX process for reading while it is still running? Is that system dependent?
 
@NicolaTalbot Yep. Formerly TexmakerX, it is a fork of Texmaker.
 
@E_L Ah I see. :)
@StephanLehmke I can think of a tail flag I used to use when reading real-time logs.
!!/answer Weather in São Paulo, Brazil.
@PauloCereda Psmith, the TeX bot: temperature | 22 °C (wind chill: 23 °C) :: conditions | cloudy :: relative humidity | 83% (dew point: 19 °C) :: wind speed | 2 m/s :: (3 hours 23 minutes ago)
 
@PauloCereda I meant from the TeX process itself. What would I have to do to make sure I can see the log output of a specific macro in the log? I assume it's buffered?
 
3:23 PM
!!/answer Weather in East Lansing, Michigan
 
@AlanMunn Psmith, the TeX bot: temperature | -8 °C (wind chill: -12 °C) :: conditions | snow, overcast :: relative humidity | 79% (dew point: -11 °C) :: wind speed | 2 m/s :: (31 minutes ago)
@StephanLehmke ooh sorry, I have no idea. :(
@AlanMunn ooh snow! :)
 
@TorbjørnT. Okay, so they're parallel rather than one superseding the other?
 
@PauloCereda It's very pretty outside right now.
 
@AlanMunn I'll visit you someday. Save some snow for me. :)
 
@NicolaTalbot Yes.
 
3:29 PM
@TorbjørnT. Okay, that's useful to know, thanks. I like that TeXstudio has its own PDF viewer so you don't have to faff around with Adobe Reader. (Not that I've used Windows in a long while, but it was a hassle when I used to teach LaTeX on Windows machines.)
 
E_L
ooh. So far I'm liking texstudio. feels nicer than texmaker
 
!!/eightball did I miss anything exciting at the super bowl?
 
@DavidCarlisle Psmith, the TeX bot: The great 8-ball says: most likely.
 
E_L
(and texniccenter crucially, since for why that is i am why here)
 
@DavidCarlisle oooh there was a blackout! I heard that @AlanMunn was involved. :)
 
3:40 PM
@PauloCereda Yes, they were using my epigraph code to run the lights.
 
E_L
!!/eightball did @AlanMunn cause the super bowl blackout?
 
@E_L Psmith, the TeX bot: The great 8-ball says: youtu.be/dQw4w9WgXcQ
 
@PauloCereda That was the only thing reported on the radio as I drove in, I thought if the most noteworthy part of the game was the bit where nothing happened, the game must be very dull.
 
@AlanMunn LOL
@E_L Sorry, my bot is pesky sometimes. :)
 
@DavidCarlisle It actually was a pretty exciting game overall. Very close, and SF could have won it in the last 5 minutes in theory.
 
E_L
3:42 PM
@PauloCereda I still can't believe that after the years I spent on IRC I still clicked that link
 
@AlanMunn Indeed. Flacco played very well.
@E_L hehe :)
!!/eightball Are you mean, Psmith?
@PauloCereda Psmith, the TeX bot: The great 8-ball says: youtu.be/dQw4w9WgXcQ
Hey!
!!/eightball I insist, are you mean?
@PauloCereda Psmith, the TeX bot: The great 8-ball says: don't count on it.
The troll video doesn't work with Germans. Apparently GEMA protects them. :)
 
E_L
!!/eightball will there ever be a tex grammar-checker?
 
@E_L Psmith, the TeX bot: The great 8-ball says: it is decidedly so.
 
kan
Hello!
 
!!/choose superbowl, a bowl full of cereal
 
E_L
3:44 PM
@PauloCereda that is exactly what I wanted to hear
 
@PauloCereda Psmith, the TeX bot: The great oracle says: superbowl
 
kan
Is someone familiar with GNU TeXinfo?
 
@kan Hi! :)
 
@kan once upon a time
 
@DavidCarlisle and they lived happily ever after.
 
3:45 PM
@PauloCereda No it vanished into the wilderness and was never seen again
 
kan
@DavidCarlisle Hmm, I just began doing stuff. I am writing code for an assignment. But, for people to know what I am doing, I want to write documentation. (Octave code...) So, could you please help with some comments on what I have now? (Not the code but the docs bit...)
 
@DavidCarlisle ouch.
 
kan
%%-*- texinfo -*-
%% Kannappan_hw2_prob4_function returns an N x N matrix which has ones along the
%% diagonal, -1 in the strict lower triangle and ones in the right most column
%% and zeroes everywhere else. @sp 1
%% @group
%% @sc{Input} = @var{N} (a natural number) @*
%% @sc{Output} = A N x N matrix of the pattern discussed in the docstring.@sp 1
%% @endgroup
%% @example
%% @code{Kannappan_hw2_prob4_function(3)}
%% @result{} 1   0   1
%%   -1   1   1
%%   -1  -1   1
%% @end example
 
Take that, Disney.
 
kan
@David How does it look? I am unhappy with Input and Output being aligned weirdly.
 
3:48 PM
!!/answer Wheather in here
 
@tohecz Psmith, the TeX bot: temperature | 21 °C (wind chill: 22 °C) :: conditions | cloudy :: relative humidity | 100% (dew point: 21 °C) :: wind speed | 3 m/s :: (3 hours 49 minutes ago)
 
@kan I can run latex in my head and get an idea how it looks but I haven't used texinfo since 1988 or something so I have no idea how it looks without actually running it, not even sure I have it :-)
 
@PauloCereda now tell me, where am I?
 
kan
@DavidCarlisle Oops. :(
 
!!/answer weather in Paris
 
3:49 PM
@tohecz Since Psmith runs here, it uses a geolocation service. The result is from a city nearby me. :)
@tohecz Psmith, the TeX bot: temperature | 10 °C (wind chill: 8 °C) :: conditions | overcast :: relative humidity | 100% (dew point: 10 °C) :: wind speed | 4.6 m/s :: (3 hours 20 minutes ago)
 
kan
Kannappan_hw2_prob4_function returns an N x N matrix which has ones
along the diagonal, -1 in the strict lower triangle and ones in the
right most column and zeroes everywhere else.

   INPUT = N (a natural number)
OUTPUT = A N x N matrix of the pattern discussed in the docstring.

     `Kannappan_hw2_prob4_function(3)'
     => 1   0   1
       -1   1   1
       -1  -1   1
From the terminal ^^ @David. :)
 
@PauloCereda yeah, this seems just right. The crane next to our building shows 10.0C and wind between 10 and 20 km/h
 
@tohecz :)
Sacred bleu!
:)
 
:)
 
can't you just use a blank line so input and output both start a para? what you have is same as latex \par INPUT \\ OUTPUT with the first one indented
 
3:53 PM
@AlanMunn: The nightmare aka carnaval is coming next week. Can I hide in your basement, please? :)
 
@PauloCereda yes of course :) you don't like festivals?
 
@tohecz It's a living hell. :)
 
kan
@DavidCarlisle Hmm. Good Analogy. But, I think @* is a line break... (as you show by \).
 
@PauloCereda ... which brings large amount of money to the area
 
@tohecz Sadly, it doesn't. Only bars can get money from tourists. :)
 
3:59 PM
@PauloCereda and they do not pay taxes? And hotels do not have profits?
 

« first day (827 days earlier)      last day (4107 days later) »