03:00 - 21:0021:00 - 00:00

3:01 AM
@GonzaloMedina Re: the chemical compound question. You're way too kind. :-)

3:29 AM
@AlanMunn so it seems ;-)
Alan, are you still around here?

@AlanMunn It made perfect sense to me.

@GonzaloMedina Mais o menos.
@Canageek Yes, but you needed to do quite a bit of reading between the lines. And the formats of the images have nothing to do with the labelling.

@AlanMunn No, but one could see how the OP would think that after reading about the diffrencees between what works in PDFLaTeX and LaTeX

@AlanMunn I have an English orthography question that I would like to ask you, but if you want, we can postpone it.

3:38 AM
I know it's English 101, but I still have this doubt and I want to solve it: do words ending in an "s" receive the apostrophe and an additional "s" when forming the possesive? I was taught that they do not receive the "s", but then I read somewhere that they can receive the "s".
@AlanMunn The above message contains my question.
Now I feel ashamed! Bothering you with such trivial questions!

3:50 AM
@GonzaloMedina I believe that one depends on which style guide you use,

It's not a problem. And this is a kind of crazy rule. First of all, it's not really driven by spelling, but by pronunciation. So here it is. Generally, a word that ends in a sibilant /s/ or /z/ (orthographically "s", "se", "z", "ze", and sometimes even "x" and "ce" will just have the apostrophe and no "s": so the the dogs' tails. But sometimes they do: for example, The Jones's house but Socrates' students.

@AlanMunn Thanks! And is there any rule allowing to determine whether they will or will not have the "s"?

Unfortunately knowing the pronunciations is something that is automatic for me as a native speaker, and so the rule isn't very difficult, but I'm not sure I actually understand it explicitly enough to explain it. I browsed the Wikipedia entry which looks solid enough on the topic (although it says some wacky things about the 'ce' cases, which I've never seen spelled with just the apostrophe.
So Socrateses (I'm avoiding IPA here) doesn't sound right, so you don't add the 's', but "Jesuses" does sound right, so you can add the "s".
It's not such a trivial question after all.

I will have a look at the Wikipedia entry. Thanks again. Just a quick test example (for me): "parentses" do sound right to me, so "parents" will have the final "s". Am I right?

Nope. In that case, no "s": My parents' house

4:04 AM
@AlanMunn arghh! Initially that was my veredict, but then I repeated "parentses" again and again and it sounded right, so I changed my mind.
I guess it's difficult for non-native English speakers.

@GonzaloMedina Yes. Let me check with a phonologist friend who seems to be online at the moment. If he responds to my chat I might have a better explanation for you.

@GonzaloMedina He disappeared. You've got me puzzled now on what the phonological condition is. I'll ask around and see if I can come up with something. For the meantime feel free to keep asking your "trivial" questions. :-)

@AlanMunn Thank you very much for taking the time! If I find another triviality, I'll ask you :-)

4 hours later…
8:14 AM
@JosephWright: could you take a look in this ad? I'm not sure I got it right: meta.tex.stackexchange.com/a/2111/3094

8:42 AM
@PauloCereda See my comment there - I think we're sticking with the 2100 slot

What's up with the old questions popping up on main due to "Community" and no visible change?

@JosephWright OK, I'm working on a new ad. :)

@wh1t3 That's one of Community's jobs. It occasionally moves questions to the top of the list (I think unanswered ones)

yeah, guess that's it

4

8:46 AM
@PauloCereda Nice

@JosephWright Updated. :)
@JosephWright: I mentioned our sessions here: blog.stackoverflow.com/2011/12/own-your-community-2/#comments
They seem to like them. :)

9:31 AM
The Santa Claus question was closed. I bet Santa is watching that thread. :P

1 hour later…
10:39 AM
My voting pattern probably matched some sort of sockpuppet tactical voting algorithm and now I'm demoted to #5. Gotta love those algorithms! :)
Santa won't give me presents because of this. Damn.
Help Al Gore! Oh wait.

11:28 AM
@PauloCereda I too lost 100 rep: I was at 50023 tonight and am at 50138 now, having got 215. Well, no double birthday, this time. :)

@egreg Ah. :)
@egreg: there's an evil plot against us here. :)

@egreg Automated recalc: ScheduledController.InvalidateVotes() old rep = 50063, new rep = 49963

(slow motion) noooooooooooooooooo!

Less than the 600 rep I lost the other time. :)

11:44 AM
I love this stamp. :)

12:10 PM
@PauloCereda I feel like posting a big "HARUMPH" comment on that blog since the ideas attributed to the travel site were all ones that we did first! I think that your comment should have mentioned this ...

@AndrewStacey Oh my bad. :( TBH I didn't pay attention to the mentions.

12:25 PM
@AndrewStacey: I did it!

@PauloCereda You should ask @barbarabeeton what colour she would prefer

@AndrewStacey Ah true. :)

2 hours later…
2:10 PM
Not sure on the reasons, but I see that my answer to
14

While not directly related to TeX, I'd say that a good read of Bringhurst's The Elements of Typographic Style is relevant here. There is a lot of very careful detail in the book, and for example the discussion on altering font kerning is very useful. What it suggests more than anything is that yo...

got accepted!

@GonzaloMedina: Do you understand the problem in this question: http://tex.stackexchange.com/questions/38345/environment-for-text-to-run-into-margins
I can't and so I can not improve the package.

@JosephWright But you didn't get the bounty.

@egreg True, and part of the reason I'm a little confused :-)

1 hour later…
3:27 PM
Just now I had 55555 rep, now it's gone again.
3

Somebody should go and downvote some of Martin's answers or questions!

@egreg Or upvote you answers ;-)

@MarcoDaniel No, it was only to see again that 55555!

@egreg I think Martin made a picture

Well we could just wait until @egreg is so far.

3:42 PM
@MartinScharrer He only needs 5000 -- ok ~200 Reputation per day -- maximal 25 days ;-)

4:04 PM
@MarcoDaniel Achilles and the tortoise. :)
Not that I want to say that Martin is a tortoise as far as rep catching is concerned.
This only makes the overtake harder. :)

7

As in the title, I'd like to be able to select whether TeXShop uses biber or BibTeX in the document rather than using the command line or changing the default in the TeXShop preferences. For TeX engines this is easy; i.e. you can put the following somewhere near the top of your TeX file % !TEX ...

For biblatex users who still use bibtex as well, Mark's solution is very useful. (And applicable to any editor/platform)

@AlanMunn I guess this could be put into TeXShop as part of its "magic lines" features.

4:19 PM
@egreg I've asked Dick about doing that. It probably wouldn't be done exactly that way, but it would be a really useful thing to have as more people switch between bibtex and biber.

@AlanMunn I'd like also that TeXShop and TeXworks used the same style for the "magic lines". I know that TeXworks understands TeXShop's, but those for TeXworks are easier

@egreg What's the difference? (I never use TeXworks).

@AlanMunn For example, to declare UTF-8 encoding, with TeXworks it's sufficient to say %! TEX encoding = UTF-8, instead of %! TEX encoding = UTF-8 Unicode. Also the TS- prefix for program could be dropped.

I see. I suspect this will be a low priority feature, especially since there is a macro to insert the line for you in TeXShop. But it would make sense.

4:36 PM
@AlanMunn %! TEX bibprogram = biber would be probably the choice, since TEX is only for announcing what follows to the program. I wouldn't favor %! BIB program=...
But I'd like also the possibility to write TeX instead of TEX (TeXworks accepts it).

1 hour later…
5:54 PM
Looking forward into a possible future; Should these features all be implemented independently of the editor?

@MarkSEveritt How would that work? In the ConTeXt world, things are run by a script almost all of the time, but for plain or LaTeX users the editor just runs the engine

@JosephWright Except for those of us who use latexmk

@Canageek I think @MarkSEveritt's solution shows a way in which many of these things could be implemented independently of the editor. (Or maybe I'm misunderstanding what you meant.)

@Canageek Yes, but that is not something that can be applied to everyone

@JosephWright Also, I find latexmk very annoying; if you compile often, it slows things down quite a bit. I prefer to run latex and bibtex separately and use latexmk only towards the very end of a project.

6:03 PM
@JosephWright What I don't get is the difference between latexmk, which he doesn't want, and a script: Isn't one just a script YOU wrote vs one that someone else wrote?

@Canageek See my previous comment. latexmk is slow, and for intermediate compilations (which are the vast majority) not really necessary. So separating bibtex runs from latex runs manually is more efficient IMO.

@AlanMunn Ah, I don't change my citations until the end most of the time, so BibTeX doesn't run that often. Also since sumatra PDF updates as the file does, I can start reading while BiBTeX is still running anyway

@AlanMunn: you were right about the chem. compounds question, after all!

@JosephWright Speaking of chem, did you ever make that title page package?

@Canageek latexmk always wants to resolve references, so it runs bibtex or biber each time that a bibliographic reference is added; it runs LaTeX twice or more every time a \label is added or it has changed its value. During document preparation it is almost always irrelevant if a cross-reference is incorrect.

6:07 PM
@GonzaloMedina Yes, I noticed that. I hate to say "I told you so" :-)

@egreg Ah, I've not worked with it heavily yet-- I was writing a history essay, so it only run it when I added a reference, in which case I would want to see that the reference had been added.

@egreg Also I find that latexmk impedes error detection in your source.

I guess I'm still in the word mindset where I want to see exactly what it looks like every time.

This question was actually prompted by using Latexmk. I prefer to control when the bibliography is built separately.

@AlanMunn That's another thing that I hate in most front ends which display only "there were errors". Beginners tend to ignore them. :(

6:11 PM
TeXStudio has an error pane. HOWEVER, it sometimes messed things up. It however will take you right to the relevent line in the log file, most of the time anyway.

@egreg Since I only use TeXShop, I've never had that problem.

When I was on linux and using emacs I used a grep script to find errors. 'grep " !" *.log' I think.

Especially as I'm just getting into biblatex, and I'm experimenting with the external resource and new regular expression stuff.

Oh, plus an argument to include the next few lines.
@MarkSEveritt My version of using BiBLaTeX is that I past the *.bib file the paper gives me into my *.bib file, then edit out the cruft and change the keyword to something I will remeber >.>

@Canageek Emacs has the very handy C-c   feature (but some errors fool it).

6:14 PM
@egreg I never understood what that was trying to tell me. Also I hated having to close the pane it opened again afterwards

@Canageek C-x 1

@Canageek I've evolved a fairly elaborate system involving citeulike, clearing up the crud citeulike does wrong and then the usual bits.

@MarkSEveritt What is citeulike?

@MarkSEveritt Oh, like RefWorks

6:16 PM
@Canageek Bibliography management.
Yeah

@MarkSEveritt Is it free? Are there upsides over refworks? I found that one almost more of a pain then just saving the PDF to my hard disk.

@Canageek It's free. I've never used refworks though so I can't comment on how they compare.
@Canageek I've been told good things about bibsonomy, which is similar.
@Canageek I'm just pretty heavily invested in citeulike now, so I have to suffer the occasional weirdness.

I'm tempted to write a question on "How do I format my Christmas List to get Santa's attention"
But I can't think of how to actually make it useful, other then asking how to have mistletoe bullet points

@Canageek I'd be tempted to close it as a duplicate. ;-)

6:27 PM
@Canageek Working on it. Expect something early in the new year

@JosephWright Sweet, thanks.

@Canageek Not about Santa, but about how to get attention with the help of formatting. It's a Great Question. :-)

@lockstep Don't we only have 2 of those? I don't see a list looking anything like either of those, though I suppose both my question and a CV want to get attention.

@Canageek It was your question I was thinking of.

@lockstep Ohhhh. >.>
I should actually get the time to use that answer over Christmas actually. My group got too busy to play right after I asked it.
So I only ever made one sample document to make sure it worked.

6:32 PM
@lockstep That's "How do I make my document look like it was written by a Santa worshipping present-hungry kid".
4

Actually I have been contemplating a "How do I not go #)(*$insane when making a multilevel indented list?" question, as mine tend to be -Main -Subpoint -Main -Subpoint -Subpoint which means I'm writing a LOT of \begin{itemize} \end{itemize} statements. A LOT of them. @Canageek Sounds like you want markdown. ;) @MarkSEveritt Markdown? The syntax that SE uses for questions and answers. I was actually thinking of an environment where - was top level, -- was 2nd level and so on. Or even \item1 \item2 and so on for what indent level. 6:38 PM @Canageek That's an interesting idea. It would be rare to need a nested itemise environment, but when it does come up it's annoying. @MarkSEveritt It comes up all the time when making presentations in Beamer. It was one of the reasons that I stopped. Anyway, if you have pandoc installed you can use that with markdown to generate the LaTeX for your list. In theory. I barely understand the markup here :( I just get it close enough and then wait for the edit patrol to clean it up. @Canageek See this for an example in markdown. I know now bits of wikimarkup (Forgotten most of it, haven't been a regular editor for ages), markdown (tiny, learning), LaTeX (Decent, improving), html (Not used really since elementary school, so very tiny bits).... 6:42 PM Hehehe. Same here. Could we just all agree to use LaTeX for everything and make a LaTeX engine for infinite pages and that compiled everything to 90% then finished when it found out the width of the page? When computers get fast enough to pass it in (really) real time, deal. @MarkSEveritt If you stuck to a limited subset and put Mozilla level money behind engine optomization they probably are now. It'd probably come down to restricting it to a subset which can be more-or-less directly translated to XML. Which is supposed to be fairly crippling for LaTeX. @MarkSEveritt Really, I'd like it if web browsers just implemented TeX's text layout algorithms. Justified web pages? SURE, why not? Wouldn't even need the rest of it; that one change would make the web so much more readable. 6:46 PM @Canageek However, advancements in HTML and associated technologies along with exciting new LaTeX things could conceivably bridge the gap one day soon. Let's just get LaTeX to render to SVG. ;) @MarkSEveritt Or we will hit the point at which you can compile LaTeX faster then the transfer time of the graphics, so you could compile it before they arrive without hte user noticing @MarkSEveritt We can't even get LaTeX to take svg as input. Well, someone mentioned you could with TikZ, but I have no idea how. @MarkSEveritt This came up at the UK-TUG AGM @Canageek Remember though, the speed of a web browser is really determined by the slowest, crappiest browser on the planet. So some old Nokia phone from 2002. @JosephWright What came up? SVG? @MarkSEveritt Jonathan Fine was very keen on the idea of SVG as the way to tackle the layput-on-the-web problem @MarkSEveritt I'd say anything older then IE6 is fine to ignore now that MS is pulling the pulg on that. But yeah, compiling LaTeX on an android would be painful 6:50 PM @Canageek Remember there are licensing issues too! @JosephWright I agree with him then. SVG isn't so mature on the web yet, but it's a lot like EPS and PDF in principle. @MarkSEveritt For LaTeX? Right, how would MS implement it closed source? They'd have to ue an open plugin or something. @MarkSEveritt The biggest issue for me is that high-quality layout can't be reflowed. For example, The LaTeX Companion needed hand-adjustment of about 10% of the pages @Canageek You could implement TeX in javascript if you like! @JosephWright Ouch @MarkSEveritt I'm not clear on the reflowing issue: Jonathan was essentially talking about non-reflowable elements @JosephWright Is that a fundamental thing with SVG, or just a lack of maturity? 6:53 PM @JosephWright Still, which would be better: LaTeX without hand adjustment or HTML? @Canageek HTML @MarkSEveritt Why WOULD you? That would run about as fast as molasses. @JosephWright How so? @Canageek I dunno. HTML is getting very nice, and it's only on a screen, which is all blocky and pixellated. @Canageek The two tasks are fundamentally different. Read up on book design, for example Bringhurst, and see how much effort is needed to choose the font, page layout, margins etc to match the text. @MarkSEveritt It still uses ragged left, which I didn't notice how bad it was until I started using LaTeX. 6:54 PM @Canageek That Mozilla money you mentioned has been pumped into getting JS really really really fast for quite a while now. Google too. @MarkSEveritt It is still never going to beat real compiled code. @CIf someone took the time to implement TeX in oopd oops @Canageek If you took the time to implement TeX in JS, I reckon you wouldn't see a significant increase in the time it takes to run. I'm not suggesting anyone do that mind. @MarkSEveritt Of course, I'm one of those crazy people who thinks the cloud is a terrible idea. hehehe @MarkSEveritt I think you would, vs equally optimized C. You are adding multiple levels between the CPU and the code, which take time to get through. @MarkSEveritt Even Python implemented with PyPy and Pshyco doesn't claim to be as fast as compiled code. 6:57 PM Not as many levels as you think. Of course, C will be faster, @Canageek Python is sloooooooooooooooooooooooow. Even with Cython. (newer than psycho) Compared with C anyway. @MarkSEveritt Yet oddly I suspect still faster then Java, if only due to less bloat. I have an idea, lets write a MIPS assembly code version of it and see how fast we can make it run Hides @Canageek The crux of it is that you have several huge companies in a JS engine arms war. @Canageek I want to try it with Haskell. @MarkSEveritt Yes, becuse everyone goes "I can put this on the web, since not-bad web devs are cheap right now, then I don't have to pay a real software designer" @Canageek just to be exotic. Javascript is the new VB.Net @MarkSEveritt That would be cool. Saddly the class I took on computer theory stopped 1 unit before lambda calculus 7:00 PM Yup. It's nicer though. @Canageek I'm more of an old school C person. Perhaps if Google's native client takes off the point will be moot. @MarkSEveritt It is odd, when I was being taught C I was annoyed at having to use malloc and no dynamic arrays, etc. Then I learn C++ and get sick of having to do weird things and all the OO people's snobbiness. @Canageek I keep meaning to learn Objective-C for that reason, but it's hard because everything out there is for Mac or iPhone coding. Well, that and finding the time. @MarkSEveritt Here is a radical idea: How about a sucessor to Fortran? IE Something modern, easy to use but still great for math? They tried that already. It was called fortress. Why didn't it catch on? 7:04 PM Fortress is a programming language designed for high-performance computing. It was created by Sun Microsystems with funding from DARPA's High Productivity Computing Systems project. One of the language designers is Guy L. Steele Jr., whose previous work includes Scheme, Common Lisp, and Java. Language features The name "Fortress" is intended to connote a secure Fortran, i.e., "a language for high-performance computation that provides abstraction and type safety on par with modern programming language principles". Its improvements include implicit parallelism, Unicode support and concr... Funding got cut. Plus it was very exotic. @MarkSEveritt and who wants a math langage that is interpreted? The whole point of a math langage is to be fast as heck, so you don't waste supercomputer time. I mean, who is going to pay to run Java on a supercomputer? Ah, this was designed to be really fast. Eventually anyway, at the moment it's only a reference implementation in Java. @MarkSEveritt Then why is it using the JavaVM? Java is slow. Oh, I see. So, now I just need to learn that, and C, write my own compiler, kidnap a really good programmer and make him write me one that doesn't suck, and hope it catches on enough to have decent library support. Or learn F95. I've used F90. It felt like an odd blend of BASIC and C 7:09 PM I like the way it handles matrices. In C that means using GSL or something like that, or worse my own code for matrices. @MarkSEveritt Wait, does it have a 'find solution to matrix' function? So I don't have to implement cramer's rule each time? droool @Canageek Not that I'm aware of. I'm thinking mainly of dot and Hadamard products. booo Solving a 3x3 matrix in C is a painnnnn C just wasn't meant to make them easy. To make matters worse GSL uses it's own implementation of complex numbers for legacy reasons. @Canageek And that bugs the crap out of me. Then, then, the ode suite only supports real numbers. Pointless. heh Thus why everyone still uses Fortran? I wonder why they chose C++ for GEANT then? 7:15 PM Everyone? Perhaps in HPC, but I'm the only one I know who's even dabbled in it in my line of work. C++ has the Boost libraries and lots of other things going for it. For better or worse, it's got a lot of momentum. Oh, a lot of profs I know use it, though probably becuse that is what they learned in the 70s as undergrads @Canageek Anyways. Nice chatting! It's gone 4am here so time to turn in. @MarkSEveritt I have an exam tommorow, so I shouldn't be chatting anyway Good luck! ;) Screams from the other room. My SO using Word and trying to number things. 5 7:21 PM @AlanMunn Words treatment of large docuemnts and figures has done more to sell the rest of my Lab on giving LaTeX a try then anything else One of our labmates just finished a 200+ page Ph.D. thesis in LaTeX. Took her TWO DAYS to get the figures and tables of contents to work. I would love a program where I could place the figures for LaTeX in a GUI, then rerun LaTeX and make adjustments. Or even a 'place figure XX cm from left and YY cm from top of page, flow text around it' command. @Canageek We may have that with the new OR, but this is actually pretty rare in published books @Canageek I suspect this depends on one's definition of "work". But the relevant comparison is how many weeks the same task would have taken in Word. @JosephWright OR? Also what is rare, exact placement of figures? Or text flowing around them? It is common in papers and RPG books to have text flow around figures, ditto textbooks. @AlanMunn Have the numbers in the right format, have the numbers in the right order, have the table of figures match the figures, have the references not show "ERROR: FIGURE NOT FOUND" @Canageek Most papers float figures to the top/bottom of pages (chemistry is a bit odd in this regard) @Canageek OR = Output Routine @JosephWright Or a 'Don't be a dick LateX, I don't want my figure halfway down the next page, away from the text talking about it. No, I also don't want it at the bottom of the page with a huge amount of whitespace in front of it. Ok, moving it to the appendix with no warning and putting it in with my source code doesn't help either' command. 7:30 PM @Canageek That's the 10% that can't be automated, I'm afraid The OR is for next year, I hope, at which stage we may have something exciting in this area. @JosephWright Is there a way to insert images that isn't a float? Someone mentioned something about that. @Canageek To insert images you only need \includegraphics @Canageek Of course, several. You don't have to put graphics inside a figure, can use \captionof from the caption package or the H specifier from the float package. @Canageek Yes, this is one of the biggest misconceptions about images (and tables) in LaTeX. You can always just use \includegraphics or \begin{tabular} anywhere in your document and things will show up exactly there. @GonzaloMedina ....why does Wikibooks tell me to use a figure then? Can I put in a caption and label if it isn't a figure? 7:32 PM @Canageek Yes, that's what the \captionof does. head desk, head desk I've gone and added title pages and other such things to get images to show up in the right place. opens up CTAN @Canageek Yes, you can use \captionof, as Joseph and Alan have mentioned. @GonzaloMedina I tend to use H foams at mouth, feels betrayed by wikibooks @JosephWright I personally don't like H very much (although in some cases it can be useful); my opinion is that if you don't want objects to float, then don't use float environments in the first place. 7:36 PM @GonzaloMedina I guess I take a slightly different view: the fact that figure floats by default does not stop it being useful mark up for 'contains a figure' even if you don't want a float. It's not called float, after all is going to fail the section of this exam on pipelining and forwarding feels LaTeX isn't going to help me when most of this will be done by hand @JosephWright it's, indeed, a different view; of course, using figure to contain a figure can be seen as more intuitive than using minipage; perhaps the name figure for the environment was not the better choice... @Canageek It's not 'everything', it's 'some work'. Now, I'm not saying LaTex2e is perfect (I'm on the LaTeX3 team for a reason) @GonzaloMedina We have some ideas for LaTeX3 on this, which might help. @JosephWright Ah, I was about to ask exactly that ;-) Oh I just needed to tell someone about how doomed I am. No idea why I thought a class on how CPUs work would be a good elective to take in my last year. Nothing to do with the above. 7:41 PM @GonzaloMedina At the moment, it's all very much a few sketches Wait until I've sorted the OR, as I said next year I hope Fonts first: my LaTeX3 task for the holiday @JosephWright What are the chances I'll ever be able to use LaTeX3 when publishing? I'm gussing not for at least 10 years after it is finalized? @Canageek Down to publishers, but as many don't have e-TeX (finalised 1999) I'm not holding my breath. What we need to do is make sure there are 'killer' features A proper design layer should help, I hope Design layer? It also depends on which publishers we are talking about @Canageek Currently, changing design in LaTeX2e is a bit of a pain, partly because the kernel fixes some values and partly because there is nothing like CSS for LaTeX We want to separate out programming, design and use We are making progress (finally) on this @brunolefloch has threatened to write his thesis in LaTeX3: he's supposed to finish in 2014! No pressure, then @JosephWright ACS, PNAS, Elsevier, wiley, etc. Noooo, CSS is evil, then I have to work in two documents, and crazy bad things happen *hides under the bed* 7:47 PM @Canageek CSS is fine provided you see the two processes separately: one is creating the page content, one is designing the page. We want the same idea: really going for 'LaTeX is about content' when you are using it. @Canageek Hmm, some chance at ACS, none at Wiley, good chance at Elsevier, I'd say @JosephWright Will I have to go to another document each time I want to bold something, or add an image or stuff? @Canageek That's not what I mean by design @JosephWright Oh, my mistake, what did you mean? Like, page margins and such? What we're after is the LaTeX document staying ~ the same, but the preamble being very different @Canageek Yes. For example, currently it's very hard to alter the way that sections look unless you load a package, which is effectively providing a custom design layer. Try using \@startsection! @JosephWright Aren't things that start with @ raw TeX? Those tend to scare me when I see them. 7:50 PM It's also hard to change design unless you understand LaTeX: taking a design brief and turning it into a layout is not easy. @Canageek That's one of the problems: you have to use the programming internals to alter design. This is A Bad Thing. @JosephWright So I'm not going to need a quarter page of \loadpackage{} to make my document work? I approve, largely as then I won't have to turn them all off every time I want to make a MWE @Canageek Not 100% clear at present. One discussion still-to-have is what should be 'kernel' and what should be 'load optionally'. Frank Mittelbach has said we need 'all of The LaTeX Companion' in the kernel. There have been some big new packages since then, and I guess they'd also be 'in'. For example, I'll push very hard for citations to be managed by an approach similar to biblatex @JosephWright (cough) biblatex @JosephWright about fonts, will I still be unable to understand what LaTeX is doing with them? I tried reading the NFSS section in The LaTeX Companion but my head hurt, and I didn't see anything I could actually use. Question: Since all LaTeX is just marcos on TeX, couldn't you just compile it down to pure TeX, the way some programming languages compile to C? @Canageek There are competing issues in this area. I'm in favour of easy naming, similar to a GUI, but I know that there are issues with that. @Canageek No, since the definition of TeX macro can change during the run. This is a vital TeX programming technique, but means that we can't simply expand everything to it's definition exhaustively. 7:55 PM @JosephWright Could there at least be a way to look up the names easily? Like a list all fonts command? The hardest part of using fonts without loading a pcakge right now is looking up that name. It's a very common misunderstanding from people with a C background @Canageek That's one area I hope we can address. As I say, there are competing demands and it's not obvious at the moment how to satisfy them all. Karl Berry, for example, is not keen on loading system fonts as this breaks document portability. @JosephWright I don't care so much about system fonts, as being able to figure out how to use whatever fonts there are. On the other hand, for someone writing a thesis on a single PC it would be much better if they could have easy access to system fonts @Canageek Not just the LaTeX team's domain, which makes life more complicated For traditional TeX fonts there is a back-end needed, and it's not realistic for the team to manage providing every single piece of font support Sorry @JosephWright By that I mean, if we stick with the existing special LaTeX fonts, could there be a clearer 'use this font here' command then what there is now? @Canageek Well yes, but then we get into 'most end users know nothing about font selection, and so should leave well alone' 8:00 PM How hard would it be to write a 'Write out the name of every font my engine can use and the 4 letter code thingy' macro? Like a libsum, but of fonts? I'm imagining something a bit like fontspec but built into the kernel So if I know I want to use TeX Gyre Pagella I don't have to read through its package? @Canageek The Berry notation, I think you mean. There is some documentation about it, but it depends on which fonts you actually own (This isn't something I think should be in the kernel, just an idea) @Canageek The free fonts are a bit different. I can imagine having core documentation on those. 8:01 PM Where are are only 26^4 choices, right? Could the marco just check them all and ignore the ones tha don't exist? Currently, my idea is that we might have a 'friendly' layer which will translate 'Palladio' into 'TeX Gyre Pagella' and then into the file name, unless of course you own Palladio in which case it will load the paid-for version. What we'll need for that is a database of fonts @JosephWright That would be cool. Anything that makes more sense of At least the common ones \title{\fontfamily{antt}\selectfont A Log of blah blah blah} @Canageek This is the sort of 'bad idea' I don't want to encourage. 8:03 PM If I want the title of my document in a different typeface then the rest of it. @JosephWright What IS the right way to set the title in a different typeface? Or is that the bad idea? @Canageek That should be done by setting up the design of the title. Now, at present the base classes don't provide an interface to do that, but things like KOMA-Script do Back with the need for a proper design layer @JosephWright So is there a way I'll be able to do something like \titleface{lmodern} or something? Oh oh, with the desing layer will I be able to make my section titles smaller without having to hunt down a package to do it for me? If you manage that then I'll owe you a second beer. @Canageek Something like that: needs more than 2 minutes thought! @Canageek That's the plan @JosephWright That is why I said 'something like' @Canageek I'd think a bit more \SetupDocumentPart { title } { fontface = Latin Modern } or something 8:07 PM So I could have {fontface=Latin Modern, fontsize = 16, fontstyle = roman} or such? That would be great If Frank M is not busy, we are hoping to get on with a lot of 'design layer' work in the next year or so. That may, of course, depend on 'real life'. (For example, my postdoc ends in February and I currently need to find a new job) @Canageek That's the general idea. Deciding on the keys will be interesting! Would it be possible to exact every TeX marco that my document is using, place them all in a file and have a single document that can compile in TeX? You couldn't change it after that, as it might then need more marcos, but for an archival purpose for example. @Canageek Well, sort of, but it would be tricky. Essentially you'd want to take your format, etc., and put them all together. @JosephWright I was thinking more a 'I don't care how old the TeX install is, I want this to compile at the other end, no matter if they are on TeX86 or ran tlmgr this morning' @Canageek That's what plain TeX users do, as the format has been stable for a long time 8:11 PM So grab every marco is used, from any package, cat them on the front end, and make a single, massive *.tex? @Canageek The problem is that you can't get the macros back out of TeX except as a dump, and that is tied to the version of the engine you have. Perhaps LuaTeX can do that: I'm not sure @JosephWright Oh, I thought LaTeX was literally a collection of TeX code. @Canageek It is, but you can't get a list of the names in the hash table out of TeX other than in a dump You have to know the names to do, say, \meaning \foo This is another thing I'm worrying about for LaTeX3. What we'd like is that, once we do reach a truly usable kernel, we'd like to have a mechanism which says 'Compile with the kernel of date X', while still allowing changes as needed Couldn't you run a script that found every \foo, then goes through LaTeX and finds the definition of \foo and puts it at the start of the document? I have some ideas, but we are not quite there yet @Canageek Catcodes 8:15 PM @JosephWright What is a catcode? @Canageek TeX assigns each character a catcode - how it is to be interpretted A control sequence name is made up of the escape char (catcode 0) followed by either one non-letter or one or more 'letters' (catcode 11) Then there are names constructed by \csname ... \endcsname What TeX thinks of as a letter can vary. See \makeatletter @.@ And I thought C was odd in some of its choices. @Canageek TeX programming is certainly interesting Search for xii.tex to see catcodes put to their most extreme use @JosephWright I'm trying to figure out how to calculate cache size right now, I'm going insane quite quickly enough, thanks. \let~\catcode~76~A13~F1~j00~P2jdefA71F~7113jdefPALLF PA''FwPA;;FPAZZFLaLPA//71F71iPAHHFLPAzzFenPASSFthP;A$\$FevP
AGGFRruoPAqq71.72.F717271PAYY7172F727171PA??Fi*LmPA&&71jfi
Fjfi71PAVVFjbigskipRPWGAUU71727374 75,76Fjpar71727375Djifx
RrhC?yLRurtKFeLPFovPgaTLtReRomL;PABB71 72,73:Fjif.73.jelse
B73:jfiXF71PU71 72,73:PWs;AMM71F71diPAJJFRdriPAQQFRsreLPAI
I71Fo71dPA!!FRgiePBt'el@ lTLqdrYmu.Q.,Ke;vz vzLqpip.Q.,tz;

8:26 PM
I've seen assembly that was more easily human readable then that.
3

8:44 PM
@egreg: I don't believe that the explanation of the "Bug in subcaption" is correct. I don't know it but I tested the example with subfig and everything works well:
\documentclass{article}
\usepackage{subfig}

\usepackage{tabularx}

\begin{document}
\begin{figure*}
\subfloat[Fig1a\label{fig:1a}]{%
}
\hfill
\subfloat[Fig1b\label{fig:1b}]{
\begin{tabularx}{0.5\textwidth}{r}
Figure 1b
\end{tabularx}
}
\caption{Fig1}
\label{fig:1}
\end{figure*}
\begin{figure}
Figure 2
\caption{Fig2 caption}
\label{fig:2}
\end{figure}
Figures \ref{fig:1a}, \ref{fig:1b}, and \ref{fig:1} should be 1a, 1b
and 1; Figure \ref{fig:2} should be 2.
\end{document}

@MarcoDaniel subcaption and subfig are very different packages.

@egreg Of course. I only guess ;-)

@MarcoDaniel Actually, with subfig, asking for \showthe\value{figure} just before the first \caption command gives 0; with subcaption TeX answers 1.
If I ask for \showthe\value{figure} in the argument of \subfloat` I get 1.

0

My paper includes several graphs that are all much too long to make as tables, so I made them figures (copy and paste from spreadsheet to paint). My paper has to include a list of figures and a list of tables, and I would like to know if there is a way that I can trick LaTeX into thinking that th...

Yet another reason for a "Do not do this" tag.

03:00 - 21:0021:00 - 00:00