@egreg I think that you are right. On the top-voted unanswered question, Seamus mentionned a newer version, but I think that he may have mistaken an update of the mh bundle for an update of breqn.

@egreg: Jap can be considered offensive, I believe you should invent another abbreviation for Japanese.

@AndreyVihrov I'll do it immediately

Do you think that \japanesetext is good?

@egreg It looks good. polyglossia would use \textjapanese, which is nearly the same.

UTF-8 in Vim gave me headaches until I found how to set it correctly. =P

@PauloCereda Doesn't it work out of box? Was your locale not UTF-8-enabled?

@AndreyVihrov In my Linux box, it works like a charm. But when I use Windows, it goes with iso-8859-1.

I use :set enc=utf-8, then everything works fine. ;-)

@PauloCereda Vim gives me headaches independently of the encoding. :)

@egreg One would expect Vim to give beeps

Friends, quick and dirty question: if I want to set \parskip 7.2pt to just a few paragraphs, is it correct to enclose the whole part with { ... }, like: { \parskip 7.2pt paragraph1 paragraph2 ... } ?

@PauloCereda Yes. Notice that \parskip is inserted before a paragraph: when a paragraph starts, TeX inserts \parskip glue, then the \everypartokens and finally it reexamines the token that started the paragraph. However, my position about spaced paragraphs is similar to Goudy's about letterspacing lowercase text.

Whoever would letterspace lowercase letters would also steal sheep.

@egreg Got it, thanks! =)

Love that sheep stealing thingy.

I'm on a modem connection these days and have disabled images in my browser accordingly — just to find out that there is no alternative text to tex.sx design elements like vote buttons and formatting icons. :-|

@AndreyVihrov Whoops! That's not good.

I sense Fedora. =)

Bother, bother, bother. pdftotext isn't reliable enough at figuring out the correct indentation of lines. Looks like I'll need to add a post-processing stage after all.

@AndrewStacey Empty spaces are not processed?

@PauloCereda Through font faces and rendering you can guess only the renderer, not the distribution. It's Gentoo :-)

@AndreyVihrov hehe you are right. That GTK engine reminded me of Nodoka, that's why I tried Fedora. But Gentoo is equally awesome! =)

@PauloCereda It would seem to me that PDF producers will not insert spaces at beginning of lines or if the space is very wide, and use some MoveTo operation instead. I concluded this from that I can't always select spaces in PDFs.

And line indentation would meet these criteria.

I see. It makes sense to optimize the text placement by moving it to the correct position instead of deliberately adding spaces.

My experiments show that pdftotext does see the empty space at the beginning of the line (with the -layout option) but that it is inconsistent as to how many actual spaces it translates that to. In one experiment, adding a tiny amount of indentation resulted in an extra 3 spaces on the input.

Anyone has a better idea of title for tex.stackexchange.com/q/9897/2707 (currently "All those braces (implicit, explicit, \iffalse{\fi}, etc)")?

@AndrewStacey I'm just starting a TeX package (which should perhaps rather be done as an engine) which reads ahead, expands tokens, performs assignments, and builds "text boxes" instead of boxes, outputting a text file. Similar to my answer to your question on "what is the next character typeset", but on a bigger scale, the aim being to support all primitives in some way or other.

(cont) Basically "\def\h{Hello}\def\w{world}\h, \w!\end" would give a text file with "Hello world!". Would that be a possible method for what you're trying to achieve? What kind of hooks would you need?

@AndrewStacey Could you provide a sample output?

@BrunoLeFloch That would be fantastic! Possibly the best thing would be for me to make my files available for you to take a look at. I have them in a BZR repository and usually just publish that, would that be okay?

@AndrewStacey I'll just have to learn/install one more version control system. But I guess it shouldn't be too hard, yes.

@PauloCereda Err, I can, but it might be simpler for me to describe what I did. I just did a loop that did \hspace{N pt}xxx with N iterating from 0 to 24 and ran pdftotext on it. Then I tried a load of similar stuff, with the indentation N ex or the width of x....x and ran pdftotext on the output and counted the spaces. I couldn't find any logic.

I already extracted (by hand :( ) a list of all the primitives I could find in TeX, eTeX and pdfTeX. I'll also have to do it for XeTeX and somehow think about LuaTeX, but that will come later.

@AndrewStacey \hspace{N pt}xxx\par?

@BrunoLeFloch Actually, you'd be able to just browse the files. The reason for using the repository would be that you could get my updates (since I'm working on this right now). If you want an idea of what I'm trying to do, take a look at ncatlab.org/doriath/show/converting%20latex%20to%20itex that's taken from a LaTeX source which I've converted to Markdown+itex via a LaTeX package.

@BrunoLeFloch Yes.

(I'll publish it tomorrow now, as it's getting late now)

@AndrewStacey Another interesting route is the opposite (Markdown->TeX), as done by Paul Isambert's interpreter package (tex.ac.uk/tex-archive/help/Catalogue/entries/interpreter.html).

@AndrewStacey Interesting, I got the same odd behaviour you described. :(

@BrunoLeFloch The Ruby Markdown interpreter (maruku) can export LaTeX, and since it can be combined with the itex2MML then it can handle maths as well. But I'm not so interested in that direction; that seems a silly direction to go in!

@AndrewStacey It depends on which language you like to write in. It could be cool to have both the power of macros and the ease of Markdown (e.g., when taking quick notes). One way may be to preprocess an almost-TeX file into a proper TeX file.

@BrunoLeFloch I just can't imagine how Markdown can be viewed as easier than LaTeX! I'm much faster at typing \item than 1. (partly because of how I have my keyboard configured, I'll admit). And I keep getting the indentation wrong, and nesting emphasis and bold is a nightmare ... shall I go on?

@AndrewStacey Do you often nest emphasis and bold? In a mixed system you could have "\emph{**this**}..." give "\emph{\textbf{this}}\dots", for instance. Mm, not very pretty anyways.

@BrunoLeFloch No I don't, except that Maruku's syntax allows you to put extra style on elements so *...* becomes a useful handle to hook these things on. Anyway, Markdown itself isn't so bad, but then there's not a lot you can do with it. Where I really miss the macro expansion is when it comes to mathematics. By the way, you might be interested in a sneak preview of a blog article: tex.blogoverflow.com/files/2011/07/se-blog.pdf

(and so to bed)

@AndrewStacey I'm reading.

@AndrewStacey Good night :).

@BrunoLeFloch ConTeXt can handle Markdown; this has been improved lately.

@MartinSchröder Does it handle mixed input? Is it possible to keep the power of macros together with some (subset of) Markdown?

@BrunoLeFloch It can generate XML and ePub. I haven't used it, but advise to give it a try. And of course there's pandoc.

@MartinSchröder The main question I think is to know which input formats remain programmable. That's one strength of TeX which remains regardless of the output format.

Pandoc doesn't seem to be programmable, but I can't tell for sure. I really agree with Andrew's point of view in his future blog post.