Open Source Mandarin

[metal.lunchbox]

I've used many textbooks and audio methods for learning Mandarin Chinese and seen many many websites with learning resources. I have encountered two related problems that I believe many of you will be familiar with. Many of the textbooks are very well designed and effective but are both costly and very limited by their renaissance-era technology. The web resources are many but tend to be very poorly designed and extremely isolated- a flash card tool here, a low-quality audio of a brief dialog there. Some of the paid services are better but cost too much and remain fairly limited.

I propose developing a comprehensive open-source online course for Mandarin (simplified) chinese that would allow for user contributed content and would be entirely free. Content would be made avaiable for others to use and modify through appropriate liscenses (CC,GPL). I am a PHP/MySQL web developer and I know that there are many more possibilities for using the web as a learning resource than are being currently utilized. Anyone can participate, and I'll start right here in this forum. Have any ideas? think this is a good/bad idea? want to contribute content? want to help code ? Any contribution is appreciated.

Maybe a good place to start would be to discuss structure and any content out there that is public domain or otherwise free to use.

james

[renzhe]

Free and open content is very important, but language teaching is a difficult field that requires experienced professionals to create content. This is the big bottleneck, and there are many resources out there with sub-par contents because creating quality content is so difficult.

I personally think that it is possible to put together a good program built on many different resources. There are good CC-licensed podcasts out there, excellent flashcard tools with good databases, good dictionaries and open source dictionary applications, there are many TV shows that can be watched online, and we have some annotations here on the forum. We also have a thread in here listing some of the better study tools on the web.

With a textbook, I'd stick with proven printed stuff. It's very hard to match that quality right now, and if you buy the books only, the cost is reasonable.

[metal.lunchbox]

Thank you so much for you input. I think the points that you raise are quite valid. I think that there is a middle way. Beginning from scratch to write a Mandarin textbook is surely very hard to do well enough to compete with the multitude of high-quality existing textbooks. And yes there are certainly many exisiting materials on the web that can be colllected into a kind of directory, but between those two extremes one can easily take inspiration from existing sources to generate original work, not working from scratch and providing more than a mere collection. I'd like to take existing content and form it together into an original, coherent whole, with some original content or modifications or existing works. I think collectively we have enough knowledge and skill to come up with something useful.

Does anyone know of public domain or free-liscensed (CC or GPL) material that might provide a useful starting point?

I know that FSI is public domain, any others?

[wrbt]

I'd love to have a site that consisted of:

1. Conversations between native speakers about 10-15 minutes long.

2. Transcript to conversation, with hover annotation

3. Click on item in transcript as you're reading it builds a vocabulary list

Yeah that kinda sounds like lingq but it would actually work properly instead of just hoping to get lucky on a word, plus content from more sources that have different regional accents.

I think it would be fairly easy to devise and algorithm that runs thru the content and checks against work frequencies to assign an arbitrary difficulty level.

[metal.lunchbox]

dialog is certainly the center of any effective language learning method. As for buidling vocab lists I think it would be easy enough for the essential vocab to be selected for the user and then allow the user to select any supplemental vocab that they would like to keep. Vocabulary excercises would be cumulative, eliminating entries through intervals. That is to say, words that a user consistently responds correctly for will not be shown as often. This much is easy enough to program using a database. I've pretty-much already programmed just such a tool using a flash-card excercise.

The basic structure of our method would probably be just like almost any other course available today: An appropriate dialog between native speakers, an introduction to the essential new vocabulary, an explanation of the grammar concepts used, a review of the dialog , excercises to allow user to integrate the concepts and vocabulary, some kind of evaluation (probably not every lesson).

So if dialog is so central to any project like this, how do we make one ?

[metal.lunchbox]

excuse me for replying to myself:

MIT's OpenCourseWare has materials for beginner to intermediate Mandarin Chinese available under a Creative Commons liscence, meaning they are free for us to use. The textbook and audio could be improved but I think they would be an excellent starting point for such a project.

http://ocw.mit.edu/OcwWeb/Foreign-Languages-and-Literatures/21F-101Spring-2006/Readings/index.htm

james

[imron]

1. Conversations between native speakers about 10-15 minutes long.

2. Transcript to conversation, with hover annotation

3. Click on item in transcript as you're reading it builds a vocabulary list

Two out of three ain't bad.

[metal.lunchbox]

Thats interesting but I don't see anything that says their material is free to use. I mean its obviously intended for the general public to have accesss to these recording but it doesn't look like I could reproduce them or modify them, or distribute them how I like. Are they freely available? doesn't seem like it.

james

[imron]

Yeah, my answer was more in response to wrbt's "I'd love to have a site that consisted of:", rather than something that could be directly used for this project.

[wrbt]

imron - thanks for the link

metal - I'm a web developer (SQL/Python/C#) but don't have time to take on more right now, maybe later in the summer I can volunteer to help you out

[anon6969]

I will help however I can.

I think a way to import the material into Anki will be good.

[metal.lunchbox]

Spaced repetition, the concept behind Anki, is certainly the key to effectively learning lots of bit of info , like chinese vocabulary. Just a week ago I wrote a simple web-app that seems to do roughly the exact same thing as anki, without ever having heard of the program before. So something like that will sure be a part of how our open-source mandarin course helps us acquire new vocabulary.

The real issue isn't so much the learning tools as the course materials themselves.

james

[trevelyan]

1. Conversations between native speakers about 10-15 minutes long.

2. Transcript to conversation, with hover annotation

3. Click on item in transcript as you're reading it builds a vocabulary list

We actually solved all three of these, and I remember we had a similar discussion a few months ago when Popup Chinese launched and we were playing around with user generated content. The major problem we discovered (the reason we de-emphasized the feature) was that most people didn't actually create good content and it reflected negatively on the stuff we were doing.

The UGC features are still active and we have one or two users who are quietly building content, incidentally, so if anyone is serious about building lessons for themselves or their fellow students I'd encourage them to get in touch. That said, it's a tough tightrope to walk.

[metal.lunchbox]

The open source mandarin project is not only about user generated content, but about having a truly open and free platform. Popupchinese is an interesting project with what I can only assume to be worthwhile content. Not only is all of that content strictly copyrighted but I was not even able view it without paying for a subscription. It is encouraging to hear that you have favorable experiences with user generated content in this field. Thank you for your comment.

james

[imcgraw]

FYI,

MIT has an open-source toolkit where we expose our speech recognizers (Mandarin and English) as web-services: http://wami.csail.mit.edu/

Right now, the setup requires an understanding of the Apache Tomcat and AJAX technology, but we're going to be releasing a pure Javascript API shortly.

[exasperation]

As a sort of pointer for what I have found te be lacking with regard to Chinese is like this site for Japanese www.guidetojapanese.org

A thorough "Chinese mindset" oriented grammar with lots of examples! That'd be lovely.

[metal.lunchbox]

RE: exasperation

That looks like such a great resource. I think the OpenMandarin project can certainly be a home for something like that for mandarin. I'd like to see some other examples of freely available grammar resources for chinese.

RE: all

I'd like to announce the premature start of the OpenMandarin project at dev.openmandarin.org . Its a collaborative project run on a wiki and a subversion repository to develop an open source mandarin course for the web. You don't have to be a proggrammer to get involved, check out our source materials page and add whatever free resources you know of. even if all you have are vague ideas, add them wherever you think them appropriate on the wiki. Take a look. I just started this so its pretty small right now.

james

[renzhe]

I admire your enthusiasm.

The site looks very nice so far, great design.

[pubmo]

Hi, just found this via google. I was thinking why there isn't a video-dialogue w/transcripts complete free chinese e-textbook. Kind of like "integrated chinese" (book i use) but completely free and with video and audio, etc.

nciku.com is pretty cool but i think it should have video and be in a level-graded textbook format with hyperlinks and so forth.

also wondering if there is such a thing for any other languages (learning english, learning spanish, etc).

also did OpenMandarin make any progress..

thanks

[markwong]

this is a really cool idea. I do software design by trade and if i can scrounge some time, might be able to help. keep me posted, and keep up the good work!