You think Japanese is hard, try LaTeX

I haven’t been posting lately because I’ve been trying to focus on my book which I’ve decided to call “Tae’s Complete Guide to Japanese”.

I’m having some hiccups because TeX, LaTeX, XeTeX, whatever the hell you want to call it SUCKS! The fact that you can’t even come up with a single name to identify what you’re talking about is a perfect example of the ass shit this monstrosity has become. This whole hodgepodge of crap is what you get when you have absolutely no API, no architecture, nor any sort of standard and instead have a bunch of people do whatever the hell they want. There’s all sorts of packages doing god knows what to each other with no sort of hierarchy, inheritance, black box protection, or namespace. Don’t even think about a single source of documentation. Documentation? Whoa, don’t get ahead of yourself with this fancy pants documentation. We ain’t gonna tolerate no stinkin document-thingy round here, boy.

Here’s what I’m struggling with. I can’t get bloody italics to show up in a Japanese font!

I’ve posted more details on my dilemma on a programming site here.

In the meantime, I’m going to try to reinstall my Tex package because I could have sworn my italics were working at some point. Was I seeing things again? Or did some other package just break it?

At least it’s not docbook, thank god!

25 thoughts on “You think Japanese is hard, try LaTeX

  1. Why don’t you write the book in Word? Exactly, because Tex does have the upper hand!

    I wouldn’t say that Tex sux. It sucks (it being the general situation, not Tex) if you run into a wall while using Tex, but all the other time, it is superior to using MS Office in my opinion. In the Office programs, you also might run into problems. I for one always have problems with changing or resetting page numbering in Office/OpenOffice, which is really easy to do in Tex.

    A few hints: Try to keep your \usepackage section clean, use only what you really need. Most packages do have good documentation, at least the more common ones, try to look it up on the net. If you can’t accomplish something within a certain time frame, post the problem on one of the Tex help groups and turn to something else for the time being. I can say that at least the German groups are quite helpful, I guess the international groups are similar.

    I’ve been writing my diploma thesis with Latex, which was my first real Tex project and my largest writing project. The outcome is quite nice. While I did have some problems on the way, most of them could be resolved within an appropriate timeframe. And now that I’m finished (yay), I must say that I’d do it this way again any time.

    Concerning the problem at hand, I haven’t used Japanese writing yet in Tex, but I think that the answers at StackOverflow are quite sound. The problem could really be in the font itself.

    By the way, which IDE are you using? TexnicCenter? I’ve been using it together with the MikTex package manager, the autocomplete and code insertion features are really nice.

  2. I like how the responses basically denied the usefulness of what you were trying to do and also denied the existence of italisised Japanese.
    I’m sure I’ve seen text be italisised on DVD subtitles before… (whether that’s the norm or it makes Japanese people feel uncomfortable, I have no idea).
    Shouldn’t there be a way to apply a slant to text via latex code?

  3. I didn’t try it out myself, but the \textsl{ text } command should give you slanted text. 🙂

  4. Word is great for doing what it’s designed to do, which is office word processing not making documents for publishing. I’m sure making mailing labels in Tex would be a royal pain as well. That’s like trying to compare Photoshop with Picasa.

    The \textsl, \textit, \emph, none of it works. I think it needs a separate italics font even though it works fine in any modern application. The fontspec package looked promising as the documentation claimed that I can pick another font just for italics and bold. However, this did not work as documented.

    This stuff is like 30 years old. I’m sure it was great in the 70s but I think it’s fine time somebody create a new platform that actually uses the advances of modern operating systems and computer science principles.

    In the end, I’m just going to use underlines (isn’t it pathetic that I can’t use ITALICS??). The ulem package overrides \emph with underlines anyway if you don’t call \normalem. This is just another example of how poorly designed this is or more accurately, what you get if you DON’T design software.

    I’ve been using NotePad++ but I’ll try TexnicCenter. Thanks.

  5. Have you considered visiting the nearest university bookstore and buying QuarkXPress at an educational-discount price?

  6. Wow, brings back memories of the research papers I had to write using *TeX. Was a royal pain in the ass to learn but I had to give the good example for the ACM’s sake. Anyway, luckily I didn’t have to use much Japanese except for that Shogi specification paper.

    I for one though, don’t like reading Japanese in italics. Anyway, good luck with it =) btw is it a commercial project? You can also go the Adobe way though I don’t personally like Adobe products that much, you know, I’ve got my rea$on ^^

  7. TeX is known for two things. One is that it can handle even really esoteric typesetting needs, like, for instance, if you’re typesetting a higher mathematics text in Hebrew and really *need* to be able to inscribe an aleph in a zero, put three dots and a hat over it, a bar to the left, and a qamets below, and then use that whole construct as a superscript under a root.

    However, the other thing it’s known for is a learning curve that tends to put off even veteran computer geeks if they don’t absolutely have to learn it.

    And yeah, most of the readily available fonts I’ve seen that can handle Japanese and Chinese characters are pretty lame in other ways. For instance, a lot of nominally serif/mincho fonts only have *some* of the characters in that style, and other characters are, to all appearances, plain sans/gothic in style. I suspect this is because there are just so MANY characters, the typeface designers run out of steam. I mean, with a bit of experience, you can slap together a font that’ll handle every character you need for English in, what, a week? Whereas, with Japanese, you could easily spend that long just getting hiragana done.

  8. Regarding the editor, nothing beats Emacs, and I’ve heard good things about AucTex. But, of course, that’s yet more learning curve. (For the record: I’m comfortable working with Emacs but moderately afraid of trying to learn TeX.)

  9. Oh, one more thing: to typesetting geeks the word “italics” has a completely different meaning from what the rest of us think it means. I think the trick word processors do when the font doesn’t have an italic version might be called “oblique” in typesetting parlance, but it is *not* what they call “italic”. “Italic” means (to them) that the shape of the letters is actually different and more stylized, not just slanted. Look at the difference between Georgia regular and Georgia italic for instance. Software can’t do that automatically; getting it to look decent is AI-complete.

    Also, when the guy on the TeX forum said that the concept of italics is “not useful” for Japanese text, I suspect he meant not useful given the current state of available fonts.

    If you can reliably get both gothic/sans and mincho/seriffed text to display as such, you might be able to use one of those for normal text and the other for emphasis. They’re quite different, visually.

  10. Yeah, no such thing as italics with CJK. And oblique CJK text makes baby Jesus cry. It’s ugly, it’s an abomination, and demons will drag you down into a crack in the earth should you attempt to use it. …well, probably. Don’t hold me to that last one.

    You are much better off using more traditional methods of emphasis, such as using different weights, or wakisen/wakiten.

    In the end, you probably shouldn’t knock a software package just because it isn’t letting you commit a crime against aesthetics and centuries of typographical tradition so easily.

  11. just a quick note to say that i am just starting to learn japanese, and have been finding your ‘japanese guide to japanese grammar’ very useful. I’m only up as far as starting to learn the hiragana alphabet- but you’ve set out the information i need clearly, and its been a great resource so far. maybe in a few years i’ll pop onto your blog again and write you a note in japanese 🙂

  12. Good luck with the textbook, TeX sounds a pain.
    I’d also like to say thanks for writing the Grammar Guide, it’s been very useful to me so far.
    Don’t know if you’ve seen, but the BBC has linked to it as one of the resources for Japanese, as part of its Japan Season:
    (Main page here, though I’m afraid I’m not sure anyone outside Japan will be able to see the videos:
    Hopefully more people will be encouraged to take an interest in Japanese.

  13. I really glad to think that many people interested in Japanese. I have study japanese for 4 years since I was 15. I even have join the JLPT test last year. I like to study it b’coz i find it unique. If there is another chance 4 me, i would like to study other language too. I just like it. Someday i hope to be a pro translater. I am no one that you know but i just want to let u know the fans of japanese have increase one person. Hehehe…. bye….

  14. @Jonadab the Unsightly One
    TeX is cool, and I’ve used it mostly for my science-related work in English. But I’ve never been able to get it to work in Japanese. So I just use Word. As you said Jonadab, TeX can do alot of things in Japanese I suppose, but I don’t think it can do furigana with tategaki text. And tables can be a real pain, in either English or Japanese…

  15. I love latex, but then again I’ve never used it for Japanese and I’d assume you need a lot of Japanese specific packages (for furigana, vertical writing and like, possibly even kerning and splitting information).

    Look forward to hearing more about your book!

  16. Sorry if I’m wrong, but I thought the replacement for italics in Japanese was just using Katakana to emphasize.

    • Katakana can be used to emphasize as well as placing ・ in between every letter but italics can be used for many other things besides just emphasis.

      In the end, I solved my problem with the MS Meiryo font. This font is the only free Japanese font I know has true italics for the English alphabet that doesn’t cost thousands of dollars. Thanks Microsoft!

  17. I agree that you shouldn’t have had the problem in the first place, but I must say I really hate the way italics look in Asian languages. Italics were designed for European writing, where slanting letters looks natural (and many people’s handwriting is naturally slanted). By contrast, Chinese characters, kana, and hangul were designed to fit neatly into square blocks, so they don’t really look so good to me when you try to slant them that way.

    It may indeed be the case that I care about this more than native Chinese/Japanese/Korean speakers do, but eh, I’m obsessed with aesthetics, what can I say? 🙂

    – Kef

  18. (La)TeX may be hard to learn, but it could be used well with some experience, there’s a lot of documentation even if it’s distributed among classes and packages. Because it’s open source I can solve problems where I could not even guess a cause with closed source apps. Not to forget there are many people in discussion boards and forums helping with user’s questions.


  19. Hi

    First, I’m impressed with your site. I’m not a speaker or learner of Japanese but still. Good luck with your book.

    You might want to try LYX ( Apparently it is latex with a GUI on top so you can click on buttons for headlines, index words, notes, etc. instead of typing the tags. Other advantage is that it can be downloaded as a complete package that should work. The package includes tex, latex and all the other necessary files. You can download binary installers for different operating systems.

    If I remember well, you can chose from many languages and fonts when downloading or installing so you might find the solution for the italic Japanese font there.

    From what I saw, documentation is pretty good ( and it has an extensive help module, accessible by clicking the help-button in the user interface.

    For me, I couldn’t get LYX to print on my PC and didn’t want to spend too much time fixing that. I went back to using WordPerfect 8, which I got with a motherboard I bought. WordPerfect is still far the best word processor for me, and I have little reason to change as long as I need it to write handouts for my English and Netherlands classes, which I print on my PC.

    Hope to be of help,


  20. I like how people deny that TeX has all these problems you’ve mentioned. Like it doesn’t have absolutely terrible packaging. Like it was designed with i18n in mind. Etc, etc.

    I think some kind of a newTeX would be a great thing. Just rewrite the system from scratch, with modularity in mind, with i18n support from the beginning, with native support for OpenType, normal tables, and so on. Also put a scripting language (like luaTeX). Don’t keep it backwards compatible. Write good documentation. That’d be a good thing. I don’t know if i have enough skill for that though, and others are just blatantly deny that TeX has problems. Because it’s so much better than Word, that is. Heh, if we’re talking about real typesetting QuarkXPress beats TeX easily for general texts, though it IS expensive.

    I think that TeX nowadays is really superior only for scientific documents. Esp. maths.

  21. The reason the wrong font is used is — I assume, not having tested it — because the Google site in question didn’t supply lang=”ja” anywhere on the page. Therefore, the browser has no way of knowing that the text is Japanese and has to guess. Hypothetically, Chrome could say to itself, “OK, this page uses a large number of kana, so it must be in Japanese”, but for some reason it doesn’t — probably to keep things simple. After all, the guy making the website is supposed to get it right in the first place. I find it amusing that Google’s own site didn’t do that, though.

    – Kef

  22. Quite late to the party, but in case someone finds this page and hopes the answer is in the comments (or xelatex is used for a later revision): \usepackage{xeCJK}

    This is basically the LaTeX “CJK” package, made suitable for use with XeLaTeX (where the latex package has to do all sorts of crazy encoding tricks, xelatex needs non of that because it actually natively understands unicode, so the plain CJK doesn’t actually work that well in xelatex). This package comes with a slanting and bolding options for CJK fonts, which can be either applied integrally to the document, or on a “per strech of text” basis.

Comments are closed.