- Have a way to identify gairaigo. Lsource tag is great but most gairaigo are from english and they don't have a Lsource tag.
- maybe allow an empty lsource for english gairaigo ?
Jim Breen : I have resisted this as the vast majority are from English and it would add to the clutter. Major kokugojiten similarly only tag non-English gairaigo. What do others think?
JLL : Two things : 1) kokugojiten do tag english gairaigo. The only difference is they don't (need to) specify the language source, as english is the default language. It is the same difference than between :
- <lsource xml::lang="eng" />
- <lsource />
I agree that the first form is not necessary, but the second one is.
2) If I want to build a Japanese/French dictionary from JMDict, I won't take the english gloss and then I will also lose the etymology of the word (unless I add something like (from eng: xxxxx) in each non english gloss). So maybe it is an advantage in a Japanese/English only dict, but not for a Multilingual one.
- In a way the first seems more logical. An empty lsource seems rather odd. It would really be shorthand for <lsource xml::lang="eng" />. I could live with either, as the EDICT/EDICT2 generator and dictionary client software that used JMdict could simply ignore it.JimBreen 05:30, 28 August 2010 (UTC)