How Good is CC-CEDICT?

[HedgePig]

I recently downloaded the CC-EDICT dictionary and I'd be interested to know how good this dictionary is. I'm impressed with what I've seen but my Chinese really isn't anywhere good enough to evaluate it properly. Is it riddled with mistakes that will trip up the ignorant (me) and unwary (again, me)? Or is it something that can pretty much be relied upon?

I really like the fact that it's free, available offline and is portable. Is there anything else I should be look at?

[renzhe]

I like it and I find it useful.

It is definitely inferior to a proper paper dictionary and to many commercial electronic dictionaries, though.

There are some errors, but it's generally OK. Make sure to download a recent version because it has been significantly expanded in the last few years.

The cool thing about CC-CEDICT is that it's free to download, and easy to write scripts for looking up words, looking up pinyin, and distribute the resulting lists. You're not allowed to do this with most other dictionaries.

[jbradfor]

As you've probably realized, CC-CEDICT is user-created. So words get added very quickly, meaning it is good for recent / technical / slang words. Which is good.

The bad part is, of course, that while I believe there are "editors" for all changes, since it is user-created, it is "riddled with mistakes". I've found about 20 over the last year or two, and if I can find them, they're not that hard to catch. However, most were quite minor. They still need to do some work to clean up their traditional character correctness, especially when one simplified character maps to two (or more) traditional characters.

So, I like it and use it (and like the MDBG interface at http://de.mdbg.net/), but like all user-created content, a bit of caution is in order.

[HedgePig]

Thanks for the feedback, renzhe and jbradfor.

I too like the fact that it can be easily searched and one can pull out lists of words.

For me the problems with the traditional characters aren't a big issue at all, although it's good to be aware of this.

HedgePig

[flash3435]

Seeing that CC-CEDICT is user-created, does that mean that it is updated from time to time with new content? I know MDBG has lot of new words like "podcast", "blog" etc. which older dictionaries probably wouldn't have. I'm curious how often the actual CEDICT files are updated.

[renzhe]

The MDBG guys are maintaining the current version of CC-CEDICT. You can edit all entries from the webpage, and submit new ones.

They have to go past an editor to get accepted, though, which is a good thing

I also believe that many of the mistakes come from the old CEDICT, before the checking of new words was strictly enforced. There used to be two versions, traditional (in Big5) and a simplified (in GBK) and they were synced manually, which is probably another source of errors.

I'd suggest simply clicking the "Edit" button if you see an error in MDBG.

Edited September 24, 2009 at 11:15 AM by renzhe

[flash3435]

Thanks! I think the review process is a good idea. My question was regarding the downloadable version of the CEDICT database that is available on MDBG. Once new entries are added to MDBG, is the downloadable database automatically updated as well? Or is that version only updated from time to time?

[renzhe]

Actually, here is the site I was looking for earlier, but couldn't find:

http://cc-cedict.org/editor/editor.php <- you can edit CC-CEDICT and add new stuff all day long. I added some of the stuff from the first episode project, actually, but not nearly all. Perhaps a later project...

Or is that version only updated from time to time?

The latest version is from today, so I guess it's updated often.

Anyway, a few years ago, CEDICT had 40,000 entries. Today, CC-CEDICT has 105,000. So the project has a bright future, IMHO, as long as the errors get weeded out over time by an active community.

So, if you see something wrong on the MDBG dictionary, edit it!