ReadTheKanji.com has been around for quite a while in a beta form, but with (fairly) recent announcements on the site going subscriptions-based I thought it would be a good time to review the system.
The ReadTheKanji.com team is offering a lifetime membership for a one-time payment of $10. Later, they’ll switch over to what looks to be recurring subscription fees, which you can bypass by jumping on the bandwagon right now. I was given a lifetime account for beta testing the system in its early stages, but I would jump on the offer if necessary.
ReadTheKanji.com has a couple of features that make it stand out over other free services like Smart.fm, on top of an expanding database of content. There are great customization options that let you choose how you are tested, and you can track your progress and print out your own color-boxed Kanji poster. But first, let’s look at the bulk of the system
The Quiz
ReadTheKanji.com has both pros and cons:
Pros
- Blanks instead of multiple choice questions
- A system that keeps track of your progress to reduce repetition of learned information
- Contextual clues to see a variety of usage
- Sentences that don’t dumb down the learning process for ambitious students of the language
The fill-in-the-blank style testing is a breath of fresh air. While there is something to be said about multiple choice questions for grammar testing, I prefer blanks for testing Kanji. The learner has a choice of study-level based on the old JLPT 4-level system, and it keeps track of progress on multiple levels – The Kana systems, individual Kanji, and vocabulary. It takes into account what you’ve already studied so you don’t have to wade through the repetitiveness of absolutely-randomized questioning.
Not only is the Kanji shown as a question, but learners also have the option to toggle contextual clues. (Or even only contextual clues as questions themselves! More on that later.) The sentences themselves aren’t dumbed down for the language learner, which fits into my philosophy of language study. If there’s a gripe I have with the contextual clues in the quiz, it’s that some of the sentences seem like textbook Japanese, which leads us to the cons.
Cons
- Some sentences could be reworked for more natural Japanese
- A micromanaged database that doesn’t allow users to provide their own sentences / terms
- No audio
- Could expand on testing styles
One major downside to the system is that content production is out of the learners’ hands. That makes sense when you’re focusing on database quality, but it makes a less flexible system. Study content is only provided by the man behind the curtain.
Another downside is the lack of audio for sentences. While the site doesn’t claim to ground you in conversational Japanese (its name is ReadTheKanji.com, afterall), it would be an added bonus to hear the sentences read out by a native, at native speed.
Finally, and this is more of a suggestion than an outright criticism, although I do prefer the fill-in-the-blanks testing style, I would like it if there were a few more testing styles thrown in the mix, and perhaps even some of the content expanded beyond just Kanji. Perhaps that’s something they’re considering for the announced subscription-based services in the future?
Prefs
The flexibility of the testing system comes from the preference settings. Learners can control how many cards they see per “round”, how often the cards repeat, whether they want to use the embedded IME romaji converter or use traditional IME input from their OS, show cards randomly or based on related Kanji, select the content that they’re tested on either in single levels or a combination of them, and change the way the quiz questions are displayed.
The real power of the system is that last point. There are currently four options for display components: Big Word (the Kanji), English definitions, Japanese sentences, and English sentences. As the system currently only tests on reading the Kanji, you can use any or all of these options to tailor the way you are tested. What I’m hoping to see in the future is an expansion on the system itself, where learners can go through it backwards to produce the meaning of Kanji, or even the Kanji itself based on, perhaps, a multi-radical lookup method.
Stats
The stats are really fun to look at, and inspire confidence in your progress. Not only that, but the huge sheet of Kanji could also be printed out in its multi-colored glory to be posted on your wall, in front of the toilet, or in your cubicle at work to review during your downtime.
You can also view your progress on vocabulary terms, and an overview of all stats broken down into JLPT levels. After a long session of studying, the stats page is where learners will most likely end their session.
In their own words:
Read The Kanji is great for reading practice and mastering kanji reading patterns, but because it uses actual Japanese sentences, it has the side benefits of helping with:
- Learning new vocabulary
- Benchmarking yourself for an upcoming JLPT test
- Expose yourself to new grammar
- Crazy awesome Japanese typing speed improvements
You can read up on the system a little bit more at the ReadTheKanji.com site, and take a JLPT4-limited test-drive of the system to decide if it’s something you can find utility in.
Related posts:
{ 2 trackbacks }
{ 23 comments… read them below or add one }
readthekanji.com, how limited. Writing the Kanji is the hard part, not reading.
For sure, it’s a limited system – But it’s not claiming to be anything more than that (at least not yet).
But you’re also a naysayer of Heisig, and with Heisig I only ever had to study each Kanji once or twice before I could write it from memory.
By the way, Tae – For nostalgic reasons, I had a look around your guide to grammar again, and it looks like it has really expanded nicely. Great work.
I’d like to see a useable web app that drills you in writing Kanji…
Criticizing ReadtheKanji for not testing writing is uncalled for!
It’d be nice, but good luck getting accurate stroke recognition via mouse input. Definitely, an OS like Snow Leopard with its native touchpad Kanji input offers the ability to create some software that does it.
What I would, instead, suggest for a webapp for something like “WriteTheKanji” would be a system based on the multi-radical input system on sites like Jisho.org and Nihongodict.com (Kanji lookup). It’s not ideal, but it’s better than nothing.
readthekanji.com uses the Tanaka Corpus for example sentences, which is far from ideal because many of the sentences are translations of English originals and you can’t tell which ones. There’s no point in using a Japanese sentence to learn from when it’s not something that was said or written in Japanese. This is a fundamental flaw in many pieces of Japanese learning software that use Jim Breen’s free resources and then charge customers.
See “A Warning to Users” on this page:
http://www.csse.monash.edu.au/~jwb/tanakacorpus.html
That’s what I was getting at when I wrote, “Some sentences could be reworked for more natural Japanese,” in the post.
Last I checked, they were fixing up and slimming down the Tanaka Corpus. No corpus is perfect, but there aren’t many other bilingual Japanese-English corpora freely available.
I admire the efforts being made to build a better corpus, but I think that they’re wasting their time. Especially since, as far as I’m aware, the people working on it aren’t native Japanese. They need to strip out everything that came from the Tanaka Corpus and build from whatever remains.
If you have a collection of sentences where half of them are worthless and you can’t tell which half, the collection in its entirety becomes worthless too. The whole point of example sentences is to see natural use of the language, but there’s no way for you to know whether any given entry is natural or not.
The Tanaka Corpus is the only free resource that I know of but just because it’s free, doesn’t mean you should use it. No corpus is perfect, but the Tanaka Corpus is unfortunately useless.
Why don’t you get a project started? I’d be happy to help out.
I’ve thought about it. The problem is copyright and data privacy. If you create a project where people can contribute sentences/examples then you should also ask for the source of the information. If the source is copyrighted material, then you risk the ire of the copyright holder. If the source is conversation/non-copyrighted material then the quality is often more dubious.
Copyright was the answer given to me when I asked the compilers of the Kenkyusha dictionary how they collected example sentences. They told me that all of their examples were written by Japanese native employees, which are then translated by English speaking natives, and then checked and rechecked by both groups. I asked them why they wrote their own examples and didn’t take real Japanese from books/films/TV/websites etc. The answer was copyright concerns.
Ideally, it would be great to have a resource in Japanese like the OED Corpus, which is regularly updated with the latest examples of English from numerous sources. The OED Corpus is based on a GBP 35 million research program though and is expensive to use.
http://www.askoxford.com/oec/mainpage/?view=uk
Personally I believe that there’s a case for copyright laws to be nullified when data use is for academic, non-profit purposes. Someone needs to petition the JP government!
Well, copyright isn’t so much of an issue when it’s derived from public discussion, like on forums written in Japanese. There are also sites like アルク, and I’m wondering if Kenkyusha sentences aren’t available to be compiled as they are, by definition, reference materials. I don’t have any sort of background in intellectual property, so I really don’t know.
Thanks Andrew for the background on the Tanaka corpus. Alex I also agree that not opening sentence construction up to the user is a real drawback. I’m less critical however than both of you because I believe that anything that encourages someone to study Japanese is a good thing. Chances are if you’re learning from a text book you’re also studying unnatural and outdated Japanese.
Most people learn very quickly on, that the language most people speak is very different from the Japanese you learn in the classroom, even when your teacher is Japanese. Your best efforts to lay the sentences in the Tanaka corpus end to end in any fashion would never come close to prose, this is just sentence level stuff.
As a learner you need to be aware of these things but not overly concerned, because it is only through communication, written and spoken, with human beings and not machines that one becomes fluent in a language.
“Chances are if you’re learning from a text book you’re also studying unnatural and outdated Japanese.”
If you’re learning from a text book then certainly the content may be contrived and may not correspond closely with the Japanese you commonly hear every day.
However, at least it was written by a Japanese person, who should be trying to write Japanese that will be useful to readers. Getting a Japanese translation of something that was originally written in English is much worse. In fact, it’s possibly the worst case scenario. You’re trying to learn Japanese and you’re studying from sentences that a Japanese person has written to have the same meaning as an English sentence. There is a large possibility that the resulting Japanese sentence is something that you would never hear from a Japanese person.
@wrightak
But at least it’s Japanese, even if bad Japanese. Obviously it’s not ideal, but it is Japanese. It’s a stepping stone to higher Japanese, kind of llike training wheels. Soon enough, as his skill grows, the learner will be exposing himself to better and more natural Japanese and will naturally discard the bad stuff.
After all, as we grow older we naturally discard all the dick and jane type speech that we were exposed to and read as children.
Alex, thanks for your straightforward review. I think this review, both the pros and cons listed are very much spot on as to how we think of ourselves. It’s great to read both the review and everyone’s comments because it just help show that the Japanese language community is also as excited about the potential of Read The Kanji as we are. So Let me take a second to try and respond to the current cons brought up.
1. It seems everyone, especially us, can agree that the sentences need work. We originally took them from the Tanaka Corpus because we just had to start somewhere; and while some sentences do have mistakes, the majority of them are pretty sound. We have been combing through the mistakes (both one by one and from user submitted reports), and are working towards replacing all the sentences with fresh, clean examples. So this is really more of an issue that just takes time.
2. We do have a micromanaged database. We have just decided to first focus on making the test the best it can be. From that point, it is much easier to add on extra features, such as custom decks, because the core will have already been built.
3. We have plans for adding in audio, but we are currently viewing this feature as just an extra for now. It would be a fantastic addition in my opinion, but since it isn’t an absolute requirement, it tends to get passed over for more pertinent features. As our subscription model gets set in place though, I’d expect this to be much more of a possibility as it could be easily outsourced while we’re work on other things.
4. Also agree about different types of tests; however, again, we’re still working on just the one test now. Once everyone seems happy with how that’s working, we can start on another and repeat the process. I’d just really rather have one great streamlined test than to have two mediocre tests.
5. I can agree that having a place to test kanji writing would be great, but then we start moving into areas beyond our scope. The same could be said for adding in a grammar section, or a listening section. These are all great ideas, but not necessarily in the scope of a site called Read The Kanji. So I can’t agree with Tae’s comment about it this being limited; it’s just the “do one thing and do it well” mentality rather than trying to do it all.
One thing I’d like to mention though is that the reason we recently made Read The Kanji a pay service was to directly fund all of these new feature requests and development. It’s not a grab for cash while providing a sub-standard service. We love working on Read The Kanji and are really excited about expanding into much more than what it is; but without funding (which we tried for a year), we realized it would be impossible add most of the requested features and expand Read The Kanji within any reasonable timeline. So our answer is to do this without having to clutter the site with distracting advertisements and popups.
I hope this answers most people’s questions!
Thanks for taking the time to respond here, Joe.
For the record, if it wasn’t already clear in the original post – I’m recommending ReadTheKanji.com to Japanese learners. I was just making sure that they know what they’re investing their $10 in, so that they’re not expecting the moon.
Thanks for building a great site, and I’ll continue to submit feedback on sentences as I go through my study rounds.
There are several open source library for recognizing written characters. I wonder why nobody has made a port that can be used for websites or iPhone apps.
http://tomoe.sourceforge.jp/cgi-bin/en/blog/index.rb?date=20060516
http://www.tegaki.org/
Definitely the tech is there, but writing characters with your mouse is a terrible task. I guess it would be better than nothing, though! I still like my idea of multi-radical selection as well which would make people focus more on radical groupings rather than individual strokes. Maybe someone could build a tiered system.
Honestly, if I had the programming skills, I’d put something together. I don’t, so I won’t – I’m just hoping someone else will come along and do it.
I just tried out tegaki that Tae pointed to. It actually works reasonably well, although some characters were really hard to recognize.
I have a Wacom tablet which completely removes the disadvantage of mouse input (tablets are very useful, and the Bamboo doesn’t cost a fortune). Same goes for smart phones, I guess. So +1 for hoping that someone with time and skills will make a web interface with a quiz element for these libraries.
Forgot to mention that I am also happily using ReadTheKanji. The sentences might not be 100% correct Japanese, but at my current (lower intermediate) level, I don’t think it matters too much.
One bad thing I noticed about my own way of studying there is that I often focus only on the reading of the kanji, not on the actual meaning of the word that I sometimes don’t remember at all. I guess this is natural, since we are only quizzed on the reading, but it is a bit unfortunate. I tried to turn off the big kanji to focus on the word in context, but not much difference. Instead, I recently started to import all difficult words into Anki, where I can test myself on the meaning as well. This seems like a reasonably productive approach so far – especially because I can run Anki on my keitai while commuting.
I can understand this kind of problem, and sometimes even experience it myself when using Read The Kanji. Lately I have been wondering if it’s just a result of the word being asked standing out too much, with it being highlighted and all the colors.
One approach I’ve been considering is just allowing the ability to turn off the highlighting. That way, you could potentially set the quiz settings to just show the definition together with a sentence without a word highlighted. It would ideally kind of force you to read the definition, and then try to actively find the word in the sentence. I think this would work even better with multiple example sentences per word, but as we are planning to do that anyway, it wouldn’t really be a problem for long.
Do you think something like this would allow users to shift the focus to the meanings better than what we can do now?
Highlighting might have something to do with it, but another factor is the break-neck speed that learners often pace themselves at out of habit. The problem might not only be with the system in that case. Maybe you could include a recommended study guide on ReadTheKanji.com where you introduce concepts like this: http://www.victorymanual.com/study-slower-learn-faster/
You might even show a few preset display options and how they might affect your studies.
I think multiple sentences will be a great addition as well and I look forward to their implementation. Thanks for all your hard work, Joe!
I had the same problem, here’s a way that could help you: first, hide the meaning in the prefs. You need to say the meaning in your head before even typing the reading. Next you validate the meaning by clicking on the “meaning” at the top. If you can’t remember the meaning or got it wrong, just press the enter key without filling the input box, you’ll then fail the kanji, both reading and meaning are now required to pass.
I think a combination of all those suggestions would work well. A more disciplined (self-enforced) study method is obviously helpful, but can be hard in practice. I think I will try out the method suggested by David – sounds like a good approach, and quite similar to how I am doing it now through Anki.*
And I am also looking very much forward to getting more example sentences – it will increase the understanding of the word as well as making it less likely to just race through at break-neck speed as Alex says. I am not really sure whether turning off the highlighting will actually work – but perhaps it will be an easy option to toggle in the preferences?
*Note: One thing that keeps me from doing what David suggests, revealing hidden elements before typing in the answer, is that this feature does not work in Opera. I think it has something to do with the implementation of the JQuery (or simply JavaScript) keypress event that is different from other browsers. So I have to switch to FireFox when I want to use that feature.
Note to note: Actually I use FireFox for ReadTheKanji, anyway, because I can then use the Peraperakun-plugin to quickly save words to a list, which can then easily be exported and imported into Anki.
Finally, also a big thank you from me to Joe for the great site you have built!