BABYL OPTIONS: -*- rmail -*- Version: 5 Labels: Note: This is the header of an rmail file. Note: If you are seeing it in rmail, Note: it means the file has no messages in it.  1, filed,, Summary-line: 12-Apr masahiko@sato.riec.tohoku #request for advice on Japanese input methods Received: by etlken.etl.go.jp (4.1/6.4J.6-ETL.SLAVE) id AA04183; Mon, 12 Apr 93 12:08:58 JST Received: by etlpom.etl.go.jp (4.1/6.4J.6-ETLpom.MASTER) id AA27383; Mon, 12 Apr 93 12:12:32 JST Received: from lime.sato.riec.tohoku.ac.jp by etlpost.etl.go.jp (5.67+1.6W/2.7W) id AA00326; Mon, 12 Apr 93 12:12:26 JST Received: by lime.sato.riec.tohoku.ac.jp (5.65+1.6W/2.7W) id AA07240; Mon, 12 Apr 93 12:12:23 JST Received: by venus.sato.riec.tohoku.ac.jp (4.1/6.4J.6) id AA22449; Mon, 12 Apr 93 12:12:22 JST Date: Mon, 12 Apr 93 12:12:22 JST From: masahiko@sato.riec.tohoku.ac.jp (Masahiko Sato) Return-Path: Message-Id: <9304120312.AA22449@venus.sato.riec.tohoku.ac.jp> To: ckclark@mit.edu Cc: mule@etl.go.jp In-Reply-To: Calvin Clark's message of Sun, 11 Apr 93 20:12:06 -0400 <9304120012.AA15270@auditorium> Subject: request for advice on Japanese input methods *** EOOH *** Date: Mon, 12 Apr 93 12:12:22 JST From: masahiko@sato.riec.tohoku.ac.jp (Masahiko Sato) Return-Path: To: ckclark@mit.edu Cc: mule@etl.go.jp In-Reply-To: Calvin Clark's message of Sun, 11 Apr 93 20:12:06 -0400 <9304120012.AA15270@auditorium> Subject: request for advice on Japanese input methods | Date: Sun, 11 Apr 93 20:12:06 -0400 | From: Calvin Clark | | Mule seems to support EGG, SJ3, CANNA, and SKK. I have tried out EGG | and SKK so far, and find them very different in feel. This is because their design philosophies are completely different. | In particular, I have the following questions: | Of these four input methods, which is the most popular? It seems | like EGG is the most recent, but I do not have a feel for how | widely used it is. I'd also like to know if any of these methods | are likely to become "obsolete" in the near future. EGG is by far the most popular. This is because both EGG and Nemacs were developed at the same institute and EGG is a built-in Japanese input system for Nemacs (and Mule). However, SKK is also very popular (I believe(^_^)). For example, at Chiba University, all the students use SKK under Nemacs as their Japanese input tool. | Which method would you recommend for beginning students of | Japanese? I recommend SKK. The reasons are: 1. Easy to learn. There is an English tutorial program, and they can learn the usage of SKK with very little knowledge of Japanese. 2. Extremely reliable. Unlike EGG, SKK does not depend on servers. So, SKK is always available as long as Mule is active. Don't use Mule without it(^_^) 3. Very fast. SKK is fast because of the following two reasons. (1) No grammatical analysis is done for conversion. (2) No overhead of the communication between the server and client. Everything is done within Mule. 4. SKK is good for Japanese beginners, because using SKK have good effects for writing Japanese by hand. There are two reasons fot this. (1) One has to correctly specify the hiragana strings that are to be converted to kanji strings. (2) One has to correctly specify the okurigana. For example, if you wish to input 「動く」(UGOKU), you have to type "UgoKu". Note that you have to do (1) and (2) when you write Japanese by hand. 5. SKK is easy for beginners. In addition to the reason 1 above, SKK is easy for beginners because it never backtracks. EGG depends on WNN and WNN does grammatical analysis of the given hiragana string. As it is impossible to always convert it into the kana-kanji string which the user expects, the user has to teach the system how it should analyze the given hiragana string. This never happens in SKK, and because of it SKK has no commands for backtracking. So there are less basic command to learn, and SKK always works as the user expected. 6. SKK has many unique features not available in EGG. (I am not an EGG user, so please forgive my ignorance if some or all of the features below are available in EGG.) I give only three examples here. (1) It is very painful if you have to type "konnpyuuta" when you wish to input 「コンピュータ」 (katakana representation of "computer"). In SKK, you just type "computer", and the SKK dictionary will correctly convert it to the correspoing katakana string. (2) There is a copletion mechanism for the string used for dictionary lookup. Suppose that you wish to input 「コ ンピュータ」 twice in a row. Then to input the second 「コンピュータ」, you just have to type "c" followed by TAB. SKK will complete it into "computer". (3) You can register any emacs lisp program in your dictionary. So, by typing "today", you can convert into "April 12, 1993". | Which method would you recommend for native speakers or | students in advanced Japanese language courses? Again SKK. In addition to the reasons given above, I would like to add that fact that SKK is equally effective for inputting written, spoken, dialect, old, modern Japanese. It is for example good for secretaries who have to type in the given document as it is. SKK is always getting better due mainly contributions of ehthusiastic SKK users. As SKK is entirely written in Emacs Lisp it is easy to improve SKK. | How do the available dictionaries compare? Among the freely availble dictionaries, SKK has the largest entries. It has currently 75000 entries and we are going to release a new SKK dictionary with 100000 entries soon. Dictionaries with large entries can have bad effects for the system like EGG which does grammatical analysis of the given hiragana string. The ratio of incorrect conversion may increse. This does not apply for SKK. Also in SKK you can register a new entry into your dictionary very easily. Because SKK dictionary requires no grammatical information for the entry, the user has no need to supply such informations. This is also nice for beginners who have only a limited knowledge about the Japanese grammar. As I said at the beginning of this mail, SKK and EGG have completely different design philosophies. It should therefore be natural that some people prefer SKK and others prefer EGG. It might be interesting to let the students try both SKK and EGG (or SJ3 or CANNA) and let them choose. masahiko  1, filed,, Summary-line: 13-Apr kon@quincy.d1.bs2.mt.nec. #Re: request for advice on Japanese input methods Received: by etlken.etl.go.jp (4.1/6.4J.6-ETL.SLAVE) id AA05058; Tue, 13 Apr 93 21:30:10 JST Received: by etlpom.etl.go.jp (4.1/6.4J.6-ETLpom.MASTER) id AA14168; Tue, 13 Apr 93 21:33:43 JST Received: from TYO.gate.nec.co.jp by etlpost.etl.go.jp (5.67+1.6W/2.7W) id AA10660; Tue, 13 Apr 93 21:33:38 JST Received: from mailsv.nec.co.jp (mailsv) by TYO.gate.nec.co.jp (5.65c/6.4J.6-TYO_gate) id AA12568; Tue, 13 Apr 1993 21:33:31 +0900 Received: from ideon.d1.bs2.mt.nec.co.jp (namesv.bs2.mt.nec.co.jp) by mailsv.nec.co.jp (5.65c/6.4JAIN) id AA26903; Tue, 13 Apr 1993 21:33:33 +0900 Received: from quincy.d1.bs2.mt.nec.co.jp by ideon.d1.bs2.mt.nec.co.jp (4.0/6.4J.6) id AA12784; Tue, 13 Apr 93 21:33:00 JST Received: by quincy.d1.bs2.mt.nec.co.jp (4.1/6.4J.6) id AA10155; Tue, 13 Apr 93 21:33:02 JST Return-Path: Message-Id: <9304131233.AA10155@quincy.d1.bs2.mt.nec.co.jp> To: ckclark@mit.edu Cc: mule@etl.go.jp Subject: Re: request for advice on Japanese input methods In-Reply-To: Your message of "Sun, 11 Apr 93 20:12:06 -0400." <9304120012.AA15270@auditorium> Reply-To: kon@d1.bs2.mt.nec.co.jp Date: Tue, 13 Apr 93 21:33:00 +0900 From: Akira Kon *** EOOH *** Return-Path: To: ckclark@mit.edu Cc: mule@etl.go.jp Subject: Re: request for advice on Japanese input methods In-Reply-To: Your message of "Sun, 11 Apr 93 20:12:06 -0400." <9304120012.AA15270@auditorium> Reply-To: kon@d1.bs2.mt.nec.co.jp Date: Tue, 13 Apr 93 21:33:00 +0900 From: Akira Kon > Which method would you recommend for native speakers or > students in advanced Japanese language courses? As I am a developer of Canna, it would be difficult to be fair and square. However, I am trying to be fair and I would like to give objective information. Egg: Egg is the oldest of the four. Egg was developed based on Wnn. Wnn is developed by Kyoto univ., Omron corp. and Astec inc. Egg is widely used among UNIX/Emacs users in Japan. SJ3: SJ3 is developed by Sony and it became a free software recently. Mule with SJ3 is based on Egg. That is, the user interface of Japanese input is the same as Egg, because SJ3 just replace the kana-kanji conversion part of Egg from jserver/Wnn to sj3serv/SJ3. You can find the source of SJ3 in the X11R5 distribution tape. Canna: Canna has similar user interface as Egg. The policy of Canna is to provide the same user interface of Japanese input method not only for Emacs but also for X environment and the other TTY environment. You can find the source of Canna in the X11R5, too. SKK: SKK is developed by Tohoku univ. SKK is the second oldest free Japanese input system for Emacs. As other mentioned, SKK has a different style of Japanese input. The following is the comparison table of these four. Egg SJ3 Canna SKK Written in elisp/C elisp elisp/C elisp or elisp kana-kanji itself conversion is jserver sj3serv cannaserver or done by (Wnn) (SJ3) (Canna) skkserv user interface is written in elisp <- C elisp same user interface is available - - kinput2(X) skkfep(TTY) also at canuum(TTY) grammatical connectivity between words dictionary dictionary dictionary user are specified by developed by ETL SONY NEC Tohoku univ > I'd also like to know if any of these methods > are likely to become "obsolete" in the near future. At least, Canna wil NOT become obsolete in the near future. :-) Please point out any errors of the above description to me． Thank you. Akira Kon