ishikawa
ishik****@yk*****
2018年 4月 10日 (火) 15:10:16 JST
(Sorry I sent an e-mail with the same content from an e-mail address NOT registered in the mailing list. Please ignore THAT message and drop it from distribution. すみません、誤って mailing list に登録してないアドレスからメイルを送ってしまったようです。 そちらのメイルがモデレーション待ちみたいですが、それは無視して削除してください。こちらのメイルが自動配送されるはずなんですが。） Hi, Alex. It is great to hear that you can pass the initial hurdle. As for "convert large clause" and "convert small clause", I was not even aware there were such phrases. Maybe someone in the know can help you. I will investigate the comments surrounding the jl_tan_conv() function in the mean time to learn the difference if possible. OK, I found out: Actually, there are somewhat outdated glossary (in Japanese) at [1] http://www.tomo.gr.jp/wnn/wnn-yogo.html Outdated since it does not seem to mention Wnn7, or Wnn8 (commercial versions), but I take that large clause is a series of small clauses. Small clause consists of "independent word" and "auxiliary word" (my translation: there may be official English terminology). See a description of Wnn6 server dictionary (in Japanese) [2] http://www.seinan-gu.ac.jp/~yositake/private/Wnn6-doc/man04.html Also, the Japanese paragraph about big and small clauses from [1]: 大文節と小文節 (large clause and small clause, that is) Wnn3からWnn4へのバージョンアップの際、「１つ以上の小文節から大文節が構成される」という概念が導入された。たとえば、「第一回」は、「第」「一」「回」という３つの小文節からなる大文節であり、「吉田さん」は、「吉田」と「さん」という２つの小文節からなる大文節だと考える。「第」「回」「さん」などは接頭接尾語で、Wnn4の候補には、接頭接尾語がついた単語を大量に並ぶことになった。これによって辞書にない単語も変換できるというメリットは、「Wnn4 はとんでもない単語を作り出してしまう!」という悪評と表裏一体だったようだ。 So 「第一回」 (large clause) .... series of 「第」「一」「回」 (each is small clause) 「吉田さん」(large clause) ... series of 「吉田」「さん」(each is small clause) 「第」is a prefix, 「回」and 「さん」are suffix (suffices?). Regards, Chiaki On 2018年04月08日 11:36, Alex Taylor wrote: > On Fri, 6 Apr 2018 19:46:17 +0900 ishikawa wrote: > >>> I'm trying to create a client for OS/2. This platform doesn't use >>> X11 so I assume porting kinput2 would be difficult... >> How interesting. Before I switched to linux (Debian distribution) >> on then main PC, I used OS/2 for 4-5 years (I started to use it >> before the release of Win98). >> >> It is true that converting kinput2 for OS/2 may be difficult (but >> maybe with clever usage of macros?). I remember there was a >> POSIX-like environment that runs on top of OS/2 that was offered by >> a few daemons and a linked library. I think there was even a port >> of X11 server. I am not sure if it is still available, though. >> (Well, I was not even sure of the availability of OS/2, but found a >> reference to ArcaOS 5.0 Blue Lion.) I am afraid that I threw away >> all the print copies and CDROMs of OS/2 magazine (in Japanese) a few >> years ago. So I can't even figure out what the software was called. >> But there was a guide about porting X11 applications to this >> environment. I remember writing my own implementation of getenv() >> so that I could port a program from Linux to OS/2. getenv() was >> missing although fork/spawn was somehow supported (I think by means >> of a clever macro, etc.) Sorry that I could not help on this aspect >> of porting. >> >> I hope kinput2 source code gives you better samples of using Wnn >> library functions. > Yes, the following (based on simplified logic from kinput2) seems to work > to connect and setup the environment+dictionaries: > > /* Connect to the server. > */ > bdata = jl_open_lang(user, server, NULL, NULL, hnd_err, hnd_msg, 0); > if ((bdata == NULL) || (jl_isconnect(bdata) == 0)) { > printf("Failed to establish session with server %s.\n", server); > return 1; > } > > if (jl_fuzokugo_get(bdata, fzk) == -1) { > /* Environment isn't active on server, so initialize it now. > */ > wnnenv = jl_env_get(bdata); > jl_set_env_wnnrc(wnnenv, "wnnenvrc", WNN_CREATE, NULL); > } > > Now I can play around with conversion. Thanks again for the tip! > > > By the way, I'm having trouble understanding the difference between > "convert large clause" and "convert small clause" (for example in the > description of jl_tan_conv). Is there a glossary somewhere that can > explain these terms? > > Thanks,
Fork