Belton wrote:heh? EJIRO has ten times the number of entries compared to JDICT.
Oops! I misread the original poster's quote, and because I had no idea what the actual number of entries was, I just took the mistaken quote at face value. I will point out, however, that if you include enamdict, compdict and example file entries, there are something like 800,000 entries in total. And EIJIRO includes conjugations that EDICT doesn't have. For example, they count the word 自閉症の as a separate entry, whereas EDICT usually just lists the root (自閉症) and the type of word. So the difference between the two isn't really the number of entries, it is the quality & usefulness of those entries.
Both EDICT and EJIRO are J to E resources. EDICT never claims to be E to J. I don't think EJIRO claims to be either. They just happen to be searchable in both directions.
EIJIRO is E-J all the way. "英辞郎とは、英和形式で入力されたデータベースです。" IMO, it's also much better for J-E than EDICT. EDICT is primarily a J-E dictionary, but it does have the intention of being used as E-J. This is why there are part-of-speech tags, and Jim encourages the use of tags such as "X" ("rude/x-rated"), "vulg" ("vulgar"), "col" ("colloquialism") and "sl" ("slang"). Sadly, submitters rarely, if ever, use these tags when they submit an entry.
I certainly have yet to find a good E to J dictionary, electronic or print. The markets seem to be in the other direction.
EIJIRO is the best on the market. Either that or the "Readers Plus" dictionary that is included in some of the handheld electronic dictionaries. The worst thing about EIJIRO, besides the crappy online search engine, are the English example sentences, which are often quite bizarre.
one thing that annoys me about EDICT is the same definitions for similar but different kanji with the same reading.
I agree. Again, this is part of the "primarily J-E dictionary written by English speakers" syndrome. Sometimes Japanese people make typos and select the wrong kanji. So English readers find these mistakes, assume that the kanji has been used properly, and then submit the new "alternate" meaning to EDICT.
But part of what you are talking about is the problem of trying to translate using a dictionary. You really need to know both languages for these nuances of meaning, and I wonder if a dictionary can really give that. Why translating's a very skilled job I suppose.
I think it can. Or at least it should strive to do so. EDICT contains all of the tools necessary to do this, but it falls short of the task. I think the reason for this is that it all comes down on one guy's head. And he's not a native speaker of Japanese. I think EDICT would be a vastly superior resource if he open-sourced it, or at least opened it to CVS access. But apparently, EDICT is just print-out from a database of his own design. So "CVS is not an option", I've been told.