« first day (1176 days earlier)      last day (3757 days later) » 
00:00 - 17:0017:00 - 00:00

5:00 PM
@Raphael So am I, if I can get some progress made
 
@JosephWright But "nobody" can program in this paradigm.
 
@Raphael Lots of people can, it's just rather specialist :-)
 
Just look at this question; many requests are of the form "let me program properly, for heavens' sake!"
 
I'd find other languages awkard
 
@JosephWright Is there another relevant language using this paradigm?
 
5:01 PM
@Raphael No idea
@Raphael There's no evidence that would work well for typesetting
 
@JosephWright :) I can relate to that. (La)TeX is the only language where I know more of than the mere basics...
 
I guess the paradigm would be fine if there were easy ways to work with it. For instance, why isn't there a \newcommandE that expands all parameters fully, no questions asked? (That's probably because the modern LaTeX ecosphere went way beyond what TeX was built for; I guess we abuse it quite hard.)
@cgnieder Because you wanted to, or you needed to in order to be productive? (I move that the latter would be bad.)
 
@Raphael Can't be done in TeX: not everything is expandable
@Raphael Fair point about TeX versus LaTeX (or indeed ConTeXt): Knuth's model was much more that you'd be doing typesetting with minimal programming to support human decisions
 
@Raphael Both. It just developed that way over the years
 
@JosephWright Try, it probably has never been tried. I can not think of any reason why e.g. a package abstraction with namespaces, proper scoping and a more "natural" parameter interface (cf xparse) could hurt. In a sense, I'm talking syntactic sugar here; I don't care how TeX works, I just don't want to have to know/bother.
 
5:06 PM
@Raphael I've never understood namespacing
@Raphael You still have to agree on some scheme for assigning names, so you ultimately are back with the TeX convention but with a different description
@Raphael Scoping I do understand but the TeX model works fine so where is the issue
 
@JosephWright It resolves conflicts. If two packages A and B define command x, using \x would throw an error, A.x and B.x would work.
@JosephWright packages don't have a fixed scope; everything can interact with everything.
 
@Raphael Yes, but you still have to agree on A and B, and anyway \newcommand does give an error
@Raphael Of course: it can in any language
 
@JosephWright At declaration point (if nobody uses renewcommand), not a usage point.
@JosephWright No. One Java class can not interfere with the implementation of another (disregarding reflection).
 
@Raphael At the end of the day, it's just expansion. :)
 
@Raphael To me, point of use looks far too late
@Raphael Exactly, if you ignore the fact it is doable :-)
 
5:09 PM
Anyway, I'm more concerned about the symptoms: crappy tool support, crappy error messages.
@JosephWright The point is that the programmer has to specifically write "interfere with that other class", it can not happen by accident.
If I wanted to implement a "simple" IDE feature like automatic import and command name completion, that would be nearly impossible.
 
@Raphael Anyone can define new commands, so 'complete' name completion is impossible
 
Parsing LaTeX logs, well, I've done that and it's not pretty. Everybody does what they want. (That's probably easy to fix.)
 
@Raphael One of the LaTeX3 things we do have is that \cs_new:Npn and so on do check names and are more 'natural' to use in code than \newcommand is.
@Raphael Fix for LaTeX logs doable if you mean making them machine-readable, but TeX messages are out of our control
 
@tohecz: Tom! How are things now?
 
@PauloCereda well, I'm alive and the building is here :)
 
5:12 PM
@Raphael I'm on the LaTeX3 team because I think things can be done better :-)
 
I really start believing that it's the problem of the pump and boiling water
 
@JosephWright Yes and no. I could tell "there are 17 variants, pick: ..." (cf automatic imports of Java IDEs).
 
I also think I don't have the skills to do much if the underlying approach is not TeX-based
 
@JosephWright Hm, you'd have to postprocess those messages, I guess.
 
@Raphael No, I meant within a document
 
5:13 PM
@JosephWright Noted, appreciated.
 
kan
@PauloCereda where do you download songs from on mac and how do you maintain them? :-)
 
@JosephWright In "proper" programming languages, that's just the same.
 
@Raphael That's what many IDEs already do: I was thinking of the 'make this easier' point
 
@JosephWright Yes, I meant post-process them inside the LaTeX compiler.
 
@kan Hello! Er... what?!
 
5:14 PM
@Raphael Sorry, don't follow: messages are for the user
 
@JosephWright A colleague and I have thinking about how we would implement LaTeX from scratch. The language, not the typesetting.
 
@Raphael So it's not LaTeX. :)
 
@JosephWright Yes. pdflatex could collect TeX messages, reformat them according to the assumed standards, and then output.
@PauloCereda Sure.
 
@Raphael Been there. :P
 
@Raphael (sorry for stepping in mid-discussion): There's one problem with typesetting: you have 4 options I think: (1) Do it in Word/HTML/LibreOffice, it'll be quick, but very bad. (2) Hire a professional, it will be good if you choose a good one, but very expensive (however, that's the way how it was done 100 years ago). (3) Learn LaTeX, it will be painful, with all the problems you express, but it will be good, if you want it to be.
 
5:16 PM
We are computer scientists, so language design comes naturally to some extent. ;)
 
(4) Learn some other "professional" typesetting tool, it'll solve most of the issues you speak about, nevertheless it will be much more painful and likely won't produce as good result.
 
kan
@PauloCereda Hi Paulo! Hmm, I was asking about how you are maintaining your library of songs on your mac and where do you download them from in the first place? :-)
 
@Raphael You can't alter how TeX messages appear: that's fixed
 
@kan Ah now I got your question!
 
@tohecz Those are the options that are. We are talking about removing pains from (3).
 
5:16 PM
Only LaTeX ones
@Raphael Tarpit
 
@JosephWright Huh? How so? Just compile them.
 
@Raphael That's engine-level changes: have you got permission from DEK?
 
The one thing we were not able to ascertain with our knowledge was: is it absolutely necessary for the lexer to be rewritten at runtime? Not for current TeX, but in principle.
 
@Raphael yep, I know, I just somehow want to say that there are tries and there are other projects, but the tries are complicated and the other projects usually make a bad job, mostly because they are not well thought out.
 
@Raphael LaTeX is a macro collection relying on a stable binary
 
5:17 PM
@JosephWright Does he care how LaTeX presents messages to users?
 
@Raphael He doesn't, but TeX is his baby
 
@kan: I let iTunes handle my music collection (actually just a part of it), but I hear there are other apps that can do the job. :) And I usually rip my CD's. :)
@Raphael He didn't care about us when writing his books, but the message is sent. :P
 
I have only the utmost respect for DEK but I don't think his ideas of usability should influence every TeX-related project. ;)
 
@Raphael People have tried less flexible approaches: to date none have worked
 
kan
@PauloCereda Hah! So, no downloading from youtube?
 
5:19 PM
@Raphael I was more meaning that as LaTeX is a macro package, we have to take what TeX (or pdfTeX, or ...) gives us message-wise
 
@JosephWright I guess I'd have to learn more in order to form an opinion. The idea that is should be necessary for symbols to change meaning feels very, very wrong to my CS-brainwashed mind.
 
@Raphael It doesn't: he after all would never consider using LaTeX :-)
 
@JosephWright I get that. But you don't have to forward that blindly to users, or do you?
 
@kan Oh, sometimes! There's an utility named youtube-dl which downloads videos from YouTube (command line). If you want some graphical aid, there's JDownloader, which has the option to convert the video to .mp3. :)
 
@Raphael Don't have any option for engine messages: they just appear
 
5:20 PM
@JosephWright He'd probably have a heart attack if he tried, and we don't want that.
 
@Raphael well, you want to do crazy stuff: For instance you want to input code listings verbatim. And since there's clearly only one way to parse files in LaTeX (i.e. source code and input files are the same for the core), you can't do it in other way than by disabling the active features of the characters locally.
 
@Raphael Macro packages can issue their own messages but not intercept those from the engine (apart from for example altering the \hfuzz to suppress some overfull box warnings).
 
@JosephWright So instead of pdftex > log, why not do pdftex | msgreformatter > log? (I seem to be missing a point that is very basic to you.)
 
@Raphael pdfTeX or whatever writes to the .log file directly, there's no redirect involved
@Raphael Are you thinking of the terminal output?
 
@tohecz True. Local changes are not what worries me as long as there is one syntactic element triggering verbatim mode. Trouble starts is a package, say babel, reconfigures the lexer for everybody.
@JosephWright Yes.
 
5:24 PM
@Raphael Much less useful than the .log file: far less detail
 
One step would be to separate "programming" space from "markup" space. I understand that LuaTeX tried to enable that.
 
@Raphael well, unless you want to write \breakablehyphen and \nonbreakablehyphen inside multipart words, there's at least a bit of a need to do that.
 
@Raphael Not really: more of a LaTeX3 thing :-)
 
@JosephWright I guess if you come from the experience of a Java or Scala compiler, the level of detail pdftex throws at you feels just unnecessary. Given that it's hard to get a meaningfull error message at the position of error, it certainly feels that way.
 
@Raphael LuaTeX offers what you'd think of of 'proper' programming, but at the engine level that doesn't separate markup out: needs to be done by the format (LaTeX, ConTeXt, ...)
 
5:26 PM
@tohecz Hm, true.
 
@Raphael that's at least ad babel's " shorthands ;)
 
@Raphael You're thinking of languages where if they work right it's probably likely you want to keep things 'tidy'. In the TeX case, the idea is to be informative.
 
@JosephWright I have never had the feeling that TeXs log messages were informative for a LaTeX writer. Maybe that's me knowing to little.
 
We do have some thoughts on the babel business, but that would need a new format and unless/until we get LaTeX3 done that's not happening.
 
@tohecz As I said, I don't have a solution. :/
 
5:28 PM
@Raphael well, some of them are not. But seeing the typical log, I wouldn't hide so much, and I wouldn't add so much, probably :)
 
@tohecz I have been using my own flavor of trimmed down log for some months now. Can't say I have to look into the full log all that often.
 
@Raphael well, if things go right, I only need bad boxes and missing references from the log. But if things do not ...
 
kan
@PauloCereda OK I just setup youtube-dl. I hope -x option works without complaints on MAC OS.
 
@tohecz The most common errors (when writing documents, not packages) usually point you at the right place. At least the first error does, the rest is usually crap.
If not, e.g. if the error got "swallows" by an environment and reported add \end{...}, the full log does not give me more.
 
@Raphael yep, only 1st error matters, that's the rule of thumb for LaTeX
 
5:34 PM
@tohecz So sad... :(
(but unavoidable, given the paradigm)
 
@Raphael I dunno. It's not a programming language, it's a text writing program. You shouldn't be making too many errors. (I mean, I'm maybe equally good in C++ and LaTeX, and I get significantly less errors in the latter)
 
sigh I guess the bigger issue is: assume I (or anybody) would come up with The Solution that allows typesetting as we like and gives "nice" programming and document-writing abstractions. Would anyone care? Probably not: it'd probably be completely incompatible with LaTeX (or any other flavor) and moving the whole ecosphere is ... impractical.
@tohecz But we do program and use programmed packages. Maybe we should not, but them's the facts. And these are usually the issues that trip me over. (1, 2) I don't know how many hours I lost because of issues with expansion.
@tohecz (That's probably because modern IDEs don't train you to get it right the first time. Errors are severly punished in LaTeX, so you learn to avoid them.)
 
@Raphael maybe, but remember that I started coding when IDEs were stupid, and I started in Pascal ... )
 
@tohecz (Oh, poor you. ;))
Well, I guess in meantime I can work on my business plan for "You write Markdown and I get what you want using pandoc". :D
It's so, so painful to watch indie publishing enterprises struggle no end because they spend time on fixing links, indices, etc with Word, Illustrate or others. :/
One guy (RPG publisher) seriously told me something like "we have to feature-freeze now because redoing the PDF index is too time consuming to do twice".
 
ppr
Question: what is the correct way to refer to two paragraphs: § 1,6 or §§ 1,6 ?
 
5:45 PM
@ppr I'd write "§ 1, § 6" with half spaces between § and number. But don't take my word for anything.
 
ppr
@Raphael thanks for the input. §§ is perhaps more useful when you indicate a range: §§ 1-6
 
@ppr Whatever you do, make sure that it's legally valid (e.g., is "1-6" unambiguous?). I hear that law works in wicked ways.
 
ppr
In my opinion, 1-6 is not ambiguous: it indicates paragraph 1 to 6 (so 1,2,3,4,5,6).
 
@ppr Oh, I see that German Wikipedia supports your double- §. Local standards may vary.
 
@ppr well, the counter-question is: In what language? :)
 
5:51 PM
@ppr Yea. That's the common sense reading in many western cultures. What about 1-5; do you mean 1,2,3,4,5 or 1,3,5? ;) Just saying that law is one domain you don't want to take chances.
 
ppr
@tohecz in english
 
@ppr well, if you decided to use § for paragraphs, then you should go for §§ 1 and 6. I would write out ` and ` rather than use ` a comma, to make the text clearer and more fluent.
 
So, gotta scramble. Thanks for the discussion, some food for thought for me. Keep up the good work!
 
§§ simply means Sections if § means Section, in whatever contect you use it. It's similar to p. and pp., just it comes with no period after it.
@Raphael well, you're welcome :) We know that LaTeX is far from being perfect, however, no other projects seem to be more successful.
 
ppr
@Raphael bon appétit
 
5:55 PM
@JosephWright Regarding our original item of discussion, I guess I'll let the editor question be about that; build tools have their own questions.
 
@Raphael OK :-)
 
@ppr Good guess, home-made burgers it'll be. :]
 
@Raphael oh I should have dinner myself! :)
 
ppr
@tohecz thanks. It will be "§§ 1 & 6" so.
 
@tohecz That's true, and today a bit of a self-fulfilling prophecy. How can any contender hope to reach the features of LaTeX before funds run out? (So, last remark.)
@ppr (Or write a macro so you can change it later.)
 
5:57 PM
@ppr well, looks cryptic and old-school, so if you wish, why not ;)
 
@Raphael Very little money supports LaTeX development :-)
 
@JosephWright that's true, and it's true that the "concurrent" projects are usually free of many as well.
btw, does "concurrent" influence "competitive" in English? Because it does so in Czech.
 
@tohecz Don't follow
 
ppr
I don't remember how to write & in LaTeX (it seems to be not \and)
 
@ppr \&
 
6:00 PM
@ppr \& :)
 
@JosephWright well, the word "concurrent" literally means "running at the same time". But does English use it exactly in this way?
 
@tohecz No implication of competition in English, just 'at the same time'
 
\& (I just want to be cool)
 
@JosephWright ok, good not now :)
@PauloCereda in order to be cool, you need to move to the Andes :)
 
@tohecz That's cold, not cool. :)
 
6:03 PM
@PauloCereda ok, s/cool/&ed/g ;)
 
@tohecz <3
 
ppr
@JosephWright @cgnieder thanks it was quite obvious... :-)
And do someone know the exact meaning of & 2ff?
& 345ff
and following ?
for clarification it is about the meaning of "ff"
 
@PauloCereda when thinking of coolness I always hear this song in my mind youtube.com/watch?v=WEUw1t8RcZ0
 
@cgnieder LOL
 
6:18 PM
@ppr “and following pages” en.wiktionary.org/wiki/ff.
 
ppr
@cgnieder thanks
 
6:31 PM
@kan Hm I never tried. Did you get it via homebrew?
@JosephWright: TeXhax link? :)
 
 
1 hour later…
7:59 PM
@PauloCereda It's only by chance that I saw that message. I saw you make the repository on GitHub and I immediately cried over it not using expl3 tools… it's turning out to be a great pain to use it, though. Oi.
 
@SeanAllred @Joseph will probably rewrite it. :)
 
@PauloCereda I got everything but color working :( something something \xglobal not doing its job for whatever reason. But that was a really great hummingbird!
 
@PauloCereda One thing about your logo stuff: a polite request not to use l3... for the name :-)
 
@SeanAllred Thanks. :)
@JosephWright lalalalala I'm not hearing you lalalalalala Now serious: I actually know that, I'll rename it ASAP. I'm just waiting for Frank and his new feature requests. :)
 
@PauloCereda Cool
I'll answer @SeanAllred's questions in a bit :-)
 
8:04 PM
@JosephWright I'll rename it, don't worry. :)
 
Short answers: not much can be done about the keyval one, and I would say 'no' on the pgf one
@PauloCereda What was it you wanted about TeXhax?
 
@JosephWright The Call for Brothers in Arms.
 
kan
@PauloCereda I wgot from their github page. :-)
 
@JosephWright Thank you. :)
 
 
1 hour later…
9:29 PM
Gold badge on AskUbuntu! askubuntu.com/questions/128413/…
 
9:49 PM
@JosephWright kool :)
Btw, I can't help myself, this comment is simply ridiculous, given my previous one got 5 stars:
@tohecz I find more useful teach the user how to catch the fish, so maybe a question about "How I find the change made in the TeX?" would be more productive and relevant, no? — Braiam 29 mins ago
 
> Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life
 
@JosephWright I get his point, but it doesn't seem to me to apply in this case.
anyways, gotta go, I get up in 6 hours
 
10:15 PM
@JosephWright What is Ubuntu? :)
Now I'm not afilliated with L3. Bah. :)
 
10:34 PM
@PauloCereda Now what's that in the readme ...
 
@TorbjørnT. Fixed. :) Check commit message. :P
 
@PauloCereda Hey, don't shoot the messenger.
 
@TorbjørnT. <3
Elton John had a song about don't shoot the piano player, I guess. :)
 
10:53 PM
@PauloCereda Look at the project I started, here's the temporary first page
user image
4
 
@egreg awwwww <3
@egreg I can't wait to read this book! :)
 
@PauloCereda Will it be finished?
 
@egreg Will the be a chapter about xor? :)
 
@PauloCereda When @DavidCarlisle finishes it.
!!/eightball O Great Oracle, will David finish xor before 2020?
 
@egreg I won't talk about it, otherwise David backfires telling me the chapter will be the next to dandelion. :)
@egreg Oh wait a minute!
!!/eightball O Great Oracle, will David finish xor before 2020?
@PauloCereda Psmith, the TeX bot: The great 8-ball says: not in this lifetime.
 
10:56 PM
!!/eightball Will I finish my project?
 
@egreg Psmith, the TeX bot: The great 8-ball says: totally.
woohoo!
!!/eightball Will I finish dandelion?
@PauloCereda Psmith, the TeX bot: The great 8-ball says: been there, done that.
 
@PauloCereda Good!
 
@egreg How was the weather today? :)
 
@PauloCereda Awful, thanks.
 
@egreg Ouch.
Ours is terrible, very, very hot.
@egreg: you know that a L3 guide would be a milestone, right? :)
I'm really looking forward to reading it. :) (wrong smiley)
41
Q: What's new in TeX, version 3.14159265?

ShadowThe new Knuth TeX has been released in early this month, and its version number is 3.14159265 now. See here. What's new in this release?

Is "The version number is new." a valid answer? :)
Oh boy, I'm getting the grip of British humour. :)
Where's my top hat and monocle? I'm feeling fancy. :)
 
11:08 PM
@PauloCereda ?
 
@DavidCarlisle He had no editor to pick, we were afraid you'd suggest emacs. :)
 
@PauloCereda no need to state the obvious, when choosing from a set of 1 possible editor.
 
@DavidCarlisle :)
@David: in other news, how's Minecraft going? :)
 
@PauloCereda the green things explode
 
@DavidCarlisle Ah the creepers. :)
They hiss then explode. :)
 
11:17 PM
@PauloCereda only if you have the sound on
 
@DavidCarlisle :)
 
@PauloCereda we lost at cricket again:(
 
!!/cricket
Psmith, the TeX bot, in fixed font mode: Let's take a look at the last cricket results:

- Pakistan 341/10  v Sri Lanka 428/9 &  133/5 *
- Bengal 114/10 &  16/1 * v Maharashtra 455/10
- Punjab 270/10  v Karnataka 39/2 *
- Ireland Women v South Africa Women
- Ireland Women v Pakistan Women

Our cricket expert David might explain these results later on.
@DavidCarlisle Oh no!
 
@PauloCereda apparently it's no longer news but having lost the ashes test series 5-0 we are now 3-0 down in a 5 match 1-day series against Australia, which isn't a great position to be in.... bbc.co.uk/sport/0/cricket/25798897
 
11:45 PM
@DavidCarlisle Oh my!
@Joseph: Done, offered my help to TUG.
1
Q: How to create a Leontief-Gozintograph with tikz?

user39628I'm looking for a way to create a leontief-gozintograph that looks similar to this one: I searched at http://www.texample.net/tikz/examples/ without success. Is there anyone who already have been creating something like this?

How to create a WHAT?!
Gotta love some questions. :)
 
@PauloCereda I want a Godzillagraph.
3
 
@egreg LOL
 
This has a somewhat off-topic flavor ;-)
0
Q: Python script does not work for a LGA with space

SabrinaAttached is my Python script, I add it to ArcToolBox to run. If the LGA name has no space, it works fine, the report file created. The problem is: when the LGA name has a space, it run but no report file is created. If manually run the script with space in LGA name, no report created, but all the...

 
00:00 - 17:0017:00 - 00:00

« first day (1176 days earlier)      last day (3757 days later) »