[PATCH] Add an option to not reduce vocabulary of the Japanese

[PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

[-- Attachment #1: Type: text/plain, Size: 630 bytes --]

Hi,

The Japanese dictionary bundled with Emacs has a small vocabulary.

For example, to convert "なごや" to "名古屋" (Nagoya) in Kanji, I would
enter "なご" and convert it to "名古", then enter "や" and convert it to
"屋".
Because the Japanese dictionary bundled with Emacs does not have "名古屋
".

The skkdic-convert function in the ja-dic-cnv package generates the
Japanese dictionary, but the logic includes the dictionary vocabulary
reduction process.

So I have created a patch to add an option to skip this reduction
process. I would be happy to receive your review and feedback.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-Add-an-option-to-not-reduce-vocabulary-of-the-Japane.patch --]
[-- Type: text/x-patch, Size: 6669 bytes --]

From 8afafacf87af38ef0fd3193d5133cf06de365c65 Mon Sep 17 00:00:00 2001
From: Taiju HIGASHI <higashi@taiju.info>
Date: Thu, 2 Jun 2022 23:24:13 +0900
Subject: [PATCH] Add an option to not reduce vocabulary of the Japanese
 dictionary.

* configure.ac: Add "with-ja-dic-reduction" configure argument.
* leim/Makefile.in: Add "JA_DIC_NO_REDUCTION_OPTION" variable.
* lisp/international/ja-dic-cnv.el (skkdic-convert-okuri-nasi): Add
"no-reduction" argument.
(skkdic-convert): Add "no-reduction" optional argument.
(batch-skkdic-convert): Add "--no-reduction" command line argument.
---
 configure.ac                     |  7 +++++++
 leim/Makefile.in                 |  4 +++-
 lisp/international/ja-dic-cnv.el | 26 ++++++++++++++++++--------
 3 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/configure.ac b/configure.ac
index ed8ec890ac..e28715ad43 100644
--- a/configure.ac
+++ b/configure.ac
@@ -491,6 +491,7 @@ OPTION_DEFAULT_ON([threads],[don't compile with elisp threading support])
 OPTION_DEFAULT_OFF([native-compilation],[compile with Emacs Lisp native compiler support])
 OPTION_DEFAULT_OFF([cygwin32-native-compilation],[use native compilation on 32-bit Cygwin])
 OPTION_DEFAULT_ON([xinput2],[don't use version 2 of the X Input Extension for input])
+OPTION_DEFAULT_ON([ja-dic-reduction],[don't reduce the Japanese dictionary])
 
 AC_ARG_WITH([file-notification],[AS_HELP_STRING([--with-file-notification=LIB],
  [use a file notification library (LIB one of: yes, inotify, kqueue, gfile, w32, no)])],
@@ -6491,6 +6492,7 @@ AS_ECHO(["  Does Emacs use -lXaw3d?                                 ${HAVE_XAW3D
   Which dumping strategy does Emacs use?                  ${with_dumping}
   Does Emacs have native lisp compiler?                   ${HAVE_NATIVE_COMP}
   Does Emacs use version 2 of the the X Input Extension?  ${HAVE_XINPUT2}
+  Does Emacs reduce the Japanese dictionary?              ${with_ja_dic_reduction}
 "])
 
 if test -n "${EMACSDATA}"; then
@@ -6589,6 +6591,11 @@ SUBDIR_MAKEFILES_IN=`echo " ${SUBDIR_MAKEFILES}" | sed -e 's| | $(srcdir)/|g' -e
 
 AC_SUBST(SUBDIR_MAKEFILES_IN)
 
+if test "$with_ja_dic_reduction" = "no"; then
+  JA_DIC_NO_REDUCTION_OPTION=--no-reduction
+fi
+AC_SUBST([JA_DIC_NO_REDUCTION_OPTION])
+
 dnl You might wonder (I did) why epaths.h is generated by running make,
 dnl rather than just letting configure generate it from epaths.in.
 dnl One reason is that the various paths are not fully expanded (see above);
diff --git a/leim/Makefile.in b/leim/Makefile.in
index 3b4216c0b8..f1a476a035 100644
--- a/leim/Makefile.in
+++ b/leim/Makefile.in
@@ -32,6 +32,8 @@ leimdir = ${srcdir}/../lisp/leim
 
 EXEEXT = @EXEEXT@
 
+JA_DIC_NO_REDUCTION_OPTION = @JA_DIC_NO_REDUCTION_OPTION@
+
 -include ${top_builddir}/src/verbose.mk
 
 # Prevent any settings in the user environment causing problems.
@@ -134,7 +136,7 @@ generate-ja-dic: ${leimdir}/ja-dic/ja-dic.el
 ${leimdir}/ja-dic/ja-dic.el: $(srcdir)/SKK-DIC/SKK-JISYO.L
 	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
 	  --eval "(setq max-specpdl-size 5000)" \
-	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" "$<"
+	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" $(JA_DIC_NO_REDUCTION_OPTION) "$<"
 
 ${srcdir}/../lisp/language/pinyin.el: ${srcdir}/MISC-DIC/pinyin.map
 	$(AM_V_GEN)${RUN_EMACS} -l titdic-cnv -f pinyin-convert $< $@
diff --git a/lisp/international/ja-dic-cnv.el b/lisp/international/ja-dic-cnv.el
index 704f1a1ae6..7d3103fd8d 100644
--- a/lisp/international/ja-dic-cnv.el
+++ b/lisp/international/ja-dic-cnv.el
@@ -295,7 +295,7 @@
       (setq skkdic-okuri-nasi-entries-count (length skkdic-okuri-nasi-entries))
       (progress-reporter-done progress))))
 
-(defun skkdic-convert-okuri-nasi (skkbuf buf)
+(defun skkdic-convert-okuri-nasi (skkbuf buf no-reduction)
   (with-current-buffer buf
     (insert ";; Setting okuri-nasi entries.\n"
 	    "(skkdic-set-okuri-nasi\n")
@@ -311,7 +311,9 @@
           (setq count (1+ count))
           (progress-reporter-update progress count)
 	  (if (setq candidates
-		    (skkdic-reduced-candidates skkbuf kana candidates))
+		    (if no-reduction
+                        candidates
+                      (skkdic-reduced-candidates skkbuf kana candidates)))
 	      (progn
 		(insert "\"" kana)
 		(while candidates
@@ -322,10 +324,12 @@
       (progress-reporter-done progress))
     (insert ")\n\n")))
 
-(defun skkdic-convert (filename &optional dirname)
+(defun skkdic-convert (filename &optional dirname no-reduction)
   "Generate Emacs Lisp file from Japanese dictionary file FILENAME.
 The format of the dictionary file should be the same as SKK dictionaries.
-Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
+Saves the output as `ja-dic-filename', in directory DIRNAME (if specified).
+When NO-REDUCTION is t, then the dictionary is not reduced.
+"
   (interactive "FSKK dictionary file: ")
   (let* ((skkbuf (get-buffer-create " *skkdic-unannotated*"))
 	 (buf (get-buffer-create "*skkdic-work*")))
@@ -389,7 +393,7 @@ Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
 	(skkdic-collect-okuri-nasi)
 
 	;; Convert okuri-nasi general entries.
-	(skkdic-convert-okuri-nasi skkbuf buf)
+	(skkdic-convert-okuri-nasi skkbuf buf no-reduction)
 
 	;; Postfix
 	(with-current-buffer buf
@@ -427,15 +431,21 @@ To get complete usage, invoke:
 	(message "To convert SKK-JISYO.L into skkdic.el:")
 	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L")
 	(message "To convert SKK-JISYO.L into DIR/ja-dic.el:")
-	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L"))
-    (let (targetdir filename)
+	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L")
+        (message "To convert SKK-JISYO.L into skkdic.el without reduction:")
+        (message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L --no-reduction"))
+    (let (targetdir filename no-reduction)
       (if (string= (car command-line-args-left) "-dir")
 	  (progn
 	    (setq command-line-args-left (cdr command-line-args-left))
 	    (setq targetdir (expand-file-name (car command-line-args-left)))
 	    (setq command-line-args-left (cdr command-line-args-left))))
+      (if (string= (car command-line-args-left) "--no-reduction")
+          (progn
+	    (setq no-reduction t)
+	    (setq command-line-args-left (cdr command-line-args-left))))
       (setq filename (expand-file-name (car command-line-args-left)))
-      (skkdic-convert filename targetdir)))
+      (skkdic-convert filename targetdir no-reduction)))
   (kill-emacs 0))
 
 
-- 
2.36.1


[-- Attachment #3: Type: text/plain, Size: 2020 bytes --]


By the way, if I may be honest, I would like to remove this reduction
process.

"名古屋" (Nagoya) [0] is the name of one of Japan's major cities and is a
proper noun.

I don't think most people, myself included, recognize that the word is a
composite of "名古" and "屋".

I am Japanese, so my sense may be different, but I recognize "New York"
as one word and "Spider-man" as one word.
In other words, instead of converting "名古" and "屋" respectively, we
want to convert "名古屋" as it is. It is stressful to have to separate
the words I imagine in my head from the words I use in Kanji
conversion. I would like to reduce that frequency at least a little.

Although the skkdic-reduced-candidates function mechanically eliminates
words that can be entered by combining them with other words, it does
not judge the importance of words, so even frequently used words like "
名古屋" are eliminated. That is very inconvenient.

My concern is that Emacs' standard Kanji conversion engine will be
regarded as useless.
Despite being based on a dictionary with a sufficient vocabulary
(SKK-JISYO.L), it generates an inconvenient dictionary by the reduction
process.
Most of the people who rated Emacs' standard kanji conversion engine as
useless are probably unaware of this fact.
I also rated the standard Emacs kanji conversion engine as
useless. Because I did not know that fact.
However, when I learned the facts, I realized that this was a
misunderstanding and that I had disrespectful feelings toward Emacs.
This is simply a disrepute due to misunderstanding.

The reduction of dictionaries would reduce the file size by less than
half. While significant, how important is this in today's computing
environment?
In my personal opinion, I feel that reducing the vocabulary of the
dictionary has more disadvantages than advantages.

My English is not very good, so I apologize if I did not convey my
intentions.

[0]: https://en.wikipedia.org/wiki/Nagoya

Best Regards,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> CC: higashi@taiju.info
> Date: Fri, 03 Jun 2022 12:16:06 +0900
> 
> The Japanese dictionary bundled with Emacs has a small vocabulary.
> 
> For example, to convert "なごや" to "名古屋" (Nagoya) in Kanji, I would
> enter "なご" and convert it to "名古", then enter "や" and convert it to
> "屋".
> Because the Japanese dictionary bundled with Emacs does not have "名古屋
> ".
> 
> The skkdic-convert function in the ja-dic-cnv package generates the
> Japanese dictionary, but the logic includes the dictionary vocabulary
> reduction process.
> 
> So I have created a patch to add an option to skip this reduction
> process. I would be happy to receive your review and feedback.

Thank you for working on this, and for your interest in Emacs.

We don't have a lot of people on board who speak Japanese, so I CC
Kenichi Handa in the hope that he could have some comments on your
patch.

Meanwhile, would you like to start the legal paperwork of assigning to
the FSF the copyright for your changes?  Your changes are small, but
they are still borderline larger than we can accept without the
copyright assignment.  If you agree, I will send you the form to fill
and instructions to go with the form.

> * configure.ac: Add "with-ja-dic-reduction" configure argument.

In addition to a configure-time option, I think it would be a good
idea to have a special Makefile rule to regenerate the Japanese
dictionary while skipping or not skipping the vocabulary reduction.
Is such an option available with your changes?  I think it is, but I'm
not certain.  So if needed, could you please add such an option to
leim/Makefile.in?

> +  Does Emacs reduce the Japanese dictionary?              ${with_ja_dic_reduction}

I guess this wording is better:

 Should Emacs reduce Japanese dictionary vocabulary?

> By the way, if I may be honest, I would like to remove this reduction
> process.
> 
> "名古屋" (Nagoya) [0] is the name of one of Japan's major cities and is a
> proper noun.
> 
> I don't think most people, myself included, recognize that the word is a
> composite of "名古" and "屋".
> 
> I am Japanese, so my sense may be different, but I recognize "New York"
> as one word and "Spider-man" as one word.
> In other words, instead of converting "名古" and "屋" respectively, we
> want to convert "名古屋" as it is. It is stressful to have to separate
> the words I imagine in my head from the words I use in Kanji
> conversion. I would like to reduce that frequency at least a little.
> 
> Although the skkdic-reduced-candidates function mechanically eliminates
> words that can be entered by combining them with other words, it does
> not judge the importance of words, so even frequently used words like "
> 名古屋" are eliminated. That is very inconvenient.
> 
> My concern is that Emacs' standard Kanji conversion engine will be
> regarded as useless.
> Despite being based on a dictionary with a sufficient vocabulary
> (SKK-JISYO.L), it generates an inconvenient dictionary by the reduction
> process.
> Most of the people who rated Emacs' standard kanji conversion engine as
> useless are probably unaware of this fact.
> I also rated the standard Emacs kanji conversion engine as
> useless. Because I did not know that fact.
> However, when I learned the facts, I realized that this was a
> misunderstanding and that I had disrespectful feelings toward Emacs.
> This is simply a disrepute due to misunderstanding.

This is something which would need an expert to respond to.  I admit
that I don't even understand the issues you are describing, as I don't
read Kanji and don't speak Japanese.  I hope Handa-san will comment on
that.

> The reduction of dictionaries would reduce the file size by less than
> half. While significant, how important is this in today's computing
> environment?

It isn't too important, IMO.  The reduction in Emacs's memory
footprint, if that is significant, is probably more important.

> My English is not very good, so I apologize if I did not convey my
> intentions.

There's absolutely nothing wrong with your English, so no need to
apologize.

Thanks!

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Hi Eli,

Thank you for reviewing!

> Meanwhile, would you like to start the legal paperwork of assigning to
> the FSF the copyright for your changes?  Your changes are small, but
> they are still borderline larger than we can accept without the
> copyright assignment.  If you agree, I will send you the form to fill
> and instructions to go with the form.

With pleasure, of course!
This is the happiest moment since I became a supporting member of FSF.

> In addition to a configure-time option, I think it would be a good
> idea to have a special Makefile rule to regenerate the Japanese
> dictionary while skipping or not skipping the vocabulary reduction.
> Is such an option available with your changes?  I think it is, but I'm
> not certain.  So if needed, could you please add such an option to
> leim/Makefile.in?

There were no options in the original code. As you suggested, I will add
a new target.

>> +  Does Emacs reduce the Japanese dictionary?              ${with_ja_dic_reduction}
>
> I guess this wording is better:
>
>  Should Emacs reduce Japanese dictionary vocabulary?

Thank you, I'll fix it and there is a wrong commit comment
(s/with-ja-dic-reduction/without-ja-dic-reduction/), so I'll fix too.

>> By the way, if I may be honest, I would like to remove this reduction
>> process.
>>
>> "名古屋" (Nagoya) [0] is the name of one of Japan's major cities and is a
>> proper noun.
>>
>> I don't think most people, myself included, recognize that the word is a
>> composite of "名古" and "屋".
>>
>> I am Japanese, so my sense may be different, but I recognize "New York"
>> as one word and "Spider-man" as one word.
>> In other words, instead of converting "名古" and "屋" respectively, we
>> want to convert "名古屋" as it is. It is stressful to have to separate
>> the words I imagine in my head from the words I use in Kanji
>> conversion. I would like to reduce that frequency at least a little.
>>
>> Although the skkdic-reduced-candidates function mechanically eliminates
>> words that can be entered by combining them with other words, it does
>> not judge the importance of words, so even frequently used words like "
>> 名古屋" are eliminated. That is very inconvenient.
>>
>> My concern is that Emacs' standard Kanji conversion engine will be
>> regarded as useless.
>> Despite being based on a dictionary with a sufficient vocabulary
>> (SKK-JISYO.L), it generates an inconvenient dictionary by the reduction
>> process.
>> Most of the people who rated Emacs' standard kanji conversion engine as
>> useless are probably unaware of this fact.
>> I also rated the standard Emacs kanji conversion engine as
>> useless. Because I did not know that fact.
>> However, when I learned the facts, I realized that this was a
>> misunderstanding and that I had disrespectful feelings toward Emacs.
>> This is simply a disrepute due to misunderstanding.
>
> This is something which would need an expert to respond to.  I admit
> that I don't even understand the issues you are describing, as I don't
> read Kanji and don't speak Japanese.  I hope Handa-san will comment on
> that.
>
>> The reduction of dictionaries would reduce the file size by less than
>> half. While significant, how important is this in today's computing
>> environment?
>
> It isn't too important, IMO.  The reduction in Emacs's memory
> footprint, if that is significant, is probably more important.

Thank you. I believe there is room for discussion about it.

--
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: Kenichi Handa <handa@gnu.org>,  emacs-devel@gnu.org
> Date: Fri, 03 Jun 2022 15:43:51 +0900
> 
> Hi Eli,
> 
> Thank you for reviewing!

You are welcome, and thanks for working on this to begin with.

> > Meanwhile, would you like to start the legal paperwork of assigning to
> > the FSF the copyright for your changes?  Your changes are small, but
> > they are still borderline larger than we can accept without the
> > copyright assignment.  If you agree, I will send you the form to fill
> > and instructions to go with the form.
> 
> With pleasure, of course!
> This is the happiest moment since I became a supporting member of FSF.

Thanks, I sent the form off-list.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

[Please use Reply All when replying, to keep the mailing list and
other people on the CC list.]

> From: Taiju HIGASHI <higashi@taiju.info>
> Date: Sat, 04 Jun 2022 00:25:20 +0900
> 
> > Thanks, I sent the form off-list.
> 
> Thank you, I sent an email to assign@gnu.org.
> 
> I created a v2 patch.
> 
> 
> >From 42932432cac237f3f32aac05377636af4c97da22 Mon Sep 17 00:00:00 2001
> From: Taiju HIGASHI <higashi@taiju.info>
> Date: Thu, 2 Jun 2022 23:24:13 +0900
> Subject: [PATCH v2] Add an option to not reduce vocabulary of the Japanese
>  dictionary.
> 
> * configure.ac: Add "without-ja-dic-reduction" configure option.
> * leim/Makefile.in (generate-ja-dic-without-reduction): Add rule.
> (generate-ja-dic): Always rerun even if the target file is generated.
> (${leimdir}/ja-dic/ja-dic.el): Do not reduce dictionary vocabulary if
>   The --without-ja-dic-reduction configure option is specified.
> * lisp/Makefile.in (generate-ja-dic): Change the make target in the
>   leim directory.
> * lisp/international/ja-dic-cnv.el (skkdic-convert-okuri-nasi): Add
>   "no-reduction" argument.
> (skkdic-convert): Add "no-reduction" optional argument.
> (batch-skkdic-convert): Add "--no-reduction" command line argument.
> ---
>  configure.ac                     |  5 +++++
>  leim/Makefile.in                 | 21 ++++++++++++++++++---
>  lisp/Makefile.in                 |  2 +-
>  lisp/international/ja-dic-cnv.el | 26 ++++++++++++++++++--------
>  4 files changed, 42 insertions(+), 12 deletions(-)
> 
> diff --git a/configure.ac b/configure.ac
> index ed8ec890ac..43af4416d3 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -491,6 +491,7 @@ OPTION_DEFAULT_ON([threads],[don't compile with elisp threading support])
>  OPTION_DEFAULT_OFF([native-compilation],[compile with Emacs Lisp native compiler support])
>  OPTION_DEFAULT_OFF([cygwin32-native-compilation],[use native compilation on 32-bit Cygwin])
>  OPTION_DEFAULT_ON([xinput2],[don't use version 2 of the X Input Extension for input])
> +OPTION_DEFAULT_ON([ja-dic-reduction],[don't reduce the Japanese dictionary])
>  
>  AC_ARG_WITH([file-notification],[AS_HELP_STRING([--with-file-notification=LIB],
>   [use a file notification library (LIB one of: yes, inotify, kqueue, gfile, w32, no)])],
> @@ -6491,6 +6492,7 @@ AS_ECHO(["  Does Emacs use -lXaw3d?                                 ${HAVE_XAW3D
>    Which dumping strategy does Emacs use?                  ${with_dumping}
>    Does Emacs have native lisp compiler?                   ${HAVE_NATIVE_COMP}
>    Does Emacs use version 2 of the the X Input Extension?  ${HAVE_XINPUT2}
> +  Should Emacs reduce Japanese dictionary vocabulary?     ${with_ja_dic_reduction}
>  "])
>  
>  if test -n "${EMACSDATA}"; then
> @@ -6589,6 +6591,9 @@ SUBDIR_MAKEFILES_IN=`echo " ${SUBDIR_MAKEFILES}" | sed -e 's| | $(srcdir)/|g' -e
>  
>  AC_SUBST(SUBDIR_MAKEFILES_IN)
>  
> +JA_DIC_REDUCTION=$with_ja_dic_reduction
> +AC_SUBST(JA_DIC_REDUCTION)
> +
>  dnl You might wonder (I did) why epaths.h is generated by running make,
>  dnl rather than just letting configure generate it from epaths.in.
>  dnl One reason is that the various paths are not fully expanded (see above);
> diff --git a/leim/Makefile.in b/leim/Makefile.in
> index 3b4216c0b8..c59f32c765 100644
> --- a/leim/Makefile.in
> +++ b/leim/Makefile.in
> @@ -32,6 +32,12 @@ leimdir = ${srcdir}/../lisp/leim
>  
>  EXEEXT = @EXEEXT@
>  
> +JA_DIC_REDUCTION = @JA_DIC_REDUCTION@
> +JA_DIC_NO_REDUCTION_OPTION =
> +ifeq ($(JA_DIC_REDUCTION), no)
> +	JA_DIC_NO_REDUCTION_OPTION = --no-reduction
> +endif
> +
>  -include ${top_builddir}/src/verbose.mk
>  
>  # Prevent any settings in the user environment causing problems.
> @@ -129,13 +135,22 @@ ${leimdir}/leim-list.el: ${srcdir}/leim-ext.el ${TIT_MISC}
>  
>  ${leimdir}/ja-dic/ja-dic.el: | $(leimdir)/ja-dic
>  
> -.PHONY: generate-ja-dic
> -generate-ja-dic: ${leimdir}/ja-dic/ja-dic.el
> -${leimdir}/ja-dic/ja-dic.el: $(srcdir)/SKK-DIC/SKK-JISYO.L
> +.PHONY: generate-ja-dic generate-ja-dic-without-reduction
> +generate-ja-dic: $(srcdir)/SKK-DIC/SKK-JISYO.L $(leimdir)/ja-dic
>  	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
>  	  --eval "(setq max-specpdl-size 5000)" \
>  	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" "$<"
>  
> +generate-ja-dic-without-reduction: $(srcdir)/SKK-DIC/SKK-JISYO.L $(leimdir)/ja-dic
> +	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
> +	  --eval "(setq max-specpdl-size 5000)" \
> +	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" --no-reduction "$<"
> +
> +${leimdir}/ja-dic/ja-dic.el: $(srcdir)/SKK-DIC/SKK-JISYO.L
> +	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
> +	  --eval "(setq max-specpdl-size 5000)" \
> +	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" $(JA_DIC_NO_REDUCTION_OPTION) "$<"
> +
>  ${srcdir}/../lisp/language/pinyin.el: ${srcdir}/MISC-DIC/pinyin.map
>  	$(AM_V_GEN)${RUN_EMACS} -l titdic-cnv -f pinyin-convert $< $@
>  
> diff --git a/lisp/Makefile.in b/lisp/Makefile.in
> index 8728467977..191f022a84 100644
> --- a/lisp/Makefile.in
> +++ b/lisp/Makefile.in
> @@ -162,7 +162,7 @@ $(lisp)/finder-inf.el:
>  
>  # This is the OKURO-NASI compilation trigger.
>  generate-ja-dic: main-first
> -	$(AM_V_at)$(MAKE) -C ../leim generate-ja-dic EMACS="$(EMACS)"
> +	$(AM_V_at)$(MAKE) -C ../leim ${srcdir}/../lisp/leim/ja-dic/ja-dic.el EMACS="$(EMACS)"
>  	$(AM_V_at)$(MAKE) compile-targets TARGETS="./leim/ja-dic/ja-dic.elc"
>  
>  org-manuals: main-first
> diff --git a/lisp/international/ja-dic-cnv.el b/lisp/international/ja-dic-cnv.el
> index 704f1a1ae6..7d3103fd8d 100644
> --- a/lisp/international/ja-dic-cnv.el
> +++ b/lisp/international/ja-dic-cnv.el
> @@ -295,7 +295,7 @@
>        (setq skkdic-okuri-nasi-entries-count (length skkdic-okuri-nasi-entries))
>        (progress-reporter-done progress))))
>  
> -(defun skkdic-convert-okuri-nasi (skkbuf buf)
> +(defun skkdic-convert-okuri-nasi (skkbuf buf no-reduction)
>    (with-current-buffer buf
>      (insert ";; Setting okuri-nasi entries.\n"
>  	    "(skkdic-set-okuri-nasi\n")
> @@ -311,7 +311,9 @@
>            (setq count (1+ count))
>            (progress-reporter-update progress count)
>  	  (if (setq candidates
> -		    (skkdic-reduced-candidates skkbuf kana candidates))
> +		    (if no-reduction
> +                        candidates
> +                      (skkdic-reduced-candidates skkbuf kana candidates)))
>  	      (progn
>  		(insert "\"" kana)
>  		(while candidates
> @@ -322,10 +324,12 @@
>        (progress-reporter-done progress))
>      (insert ")\n\n")))
>  
> -(defun skkdic-convert (filename &optional dirname)
> +(defun skkdic-convert (filename &optional dirname no-reduction)
>    "Generate Emacs Lisp file from Japanese dictionary file FILENAME.
>  The format of the dictionary file should be the same as SKK dictionaries.
> -Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
> +Saves the output as `ja-dic-filename', in directory DIRNAME (if specified).
> +When NO-REDUCTION is t, then the dictionary is not reduced.
> +"
>    (interactive "FSKK dictionary file: ")
>    (let* ((skkbuf (get-buffer-create " *skkdic-unannotated*"))
>  	 (buf (get-buffer-create "*skkdic-work*")))
> @@ -389,7 +393,7 @@ Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
>  	(skkdic-collect-okuri-nasi)
>  
>  	;; Convert okuri-nasi general entries.
> -	(skkdic-convert-okuri-nasi skkbuf buf)
> +	(skkdic-convert-okuri-nasi skkbuf buf no-reduction)
>  
>  	;; Postfix
>  	(with-current-buffer buf
> @@ -427,15 +431,21 @@ To get complete usage, invoke:
>  	(message "To convert SKK-JISYO.L into skkdic.el:")
>  	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L")
>  	(message "To convert SKK-JISYO.L into DIR/ja-dic.el:")
> -	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L"))
> -    (let (targetdir filename)
> +	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L")
> +        (message "To convert SKK-JISYO.L into skkdic.el without reduction:")
> +        (message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L --no-reduction"))
> +    (let (targetdir filename no-reduction)
>        (if (string= (car command-line-args-left) "-dir")
>  	  (progn
>  	    (setq command-line-args-left (cdr command-line-args-left))
>  	    (setq targetdir (expand-file-name (car command-line-args-left)))
>  	    (setq command-line-args-left (cdr command-line-args-left))))
> +      (if (string= (car command-line-args-left) "--no-reduction")
> +          (progn
> +	    (setq no-reduction t)
> +	    (setq command-line-args-left (cdr command-line-args-left))))
>        (setq filename (expand-file-name (car command-line-args-left)))
> -      (skkdic-convert filename targetdir)))
> +      (skkdic-convert filename targetdir no-reduction)))
>    (kill-emacs 0))
>  
>  
> -- 
> 2.36.1
> 
> 
> I had a little trouble in creating the new rules, so please point out
> any problems.
> 
> I implemented with the following policy.
> 
> 1. When re-running make from the project root, skip dictionary
>    generation if it has already been generated
> 2. In the Makefile in the leim directory, allow re-runs to regenerate
>    the dictionary.
> 
> It was difficult for me personally to distinguish between using
> leim/Makefile and lisp/Makefile.
> 
> Since leim/Makefile seems to be executed via lisp/Makefile, I modified
> leim/Makefile for convenience of dictionary generation and lisp/Makefile
> for convenience when make is executed in the project root.

I thought no changes were needed, because one could already say
something like

  make ja-dic.el JA_DIC_REDUCTION=''

Wouldn't that work?  Perhaps in lisp/Makefile.in one should make sure
the JA_DIC_REDUCTION value was propagated to the sub-make running in
leim/, but that's all.

Am I missing something?

Thanks.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

> I thought no changes were needed, because one could already say
> something like
>
>   make ja-dic.el JA_DIC_REDUCTION=''
>
> Wouldn't that work?  Perhaps in lisp/Makefile.in one should make sure
> the JA_DIC_REDUCTION value was propagated to the sub-make running in
> leim/, but that's all.
>
> Am I missing something?
>
> Thanks.

I may not have understood it properly.

Did you imagine that the special rules for re-generating would be
implemented in lisp/Makefile?
In that case, are you saying that there is no need to modify the
leim/Makefile?
Certainly, if that is the case, you may be right.

For example, what about the idea of adding regenerate-ja-dic and
regenerate-ja-dic-without-reductoin as new special rules in
lisp/Makefile?

These rules are intended to be independent of the existing state and are
always intended to rules that can reliably regenerate dictionaries with
reduced vocabularies or without vocabulary reduction.

If that is the case, there should be no changes to the previous patch,
only new rules added.

Sorry if I missed the point.

Thank you,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: handa@gnu.org,  emacs-devel@gnu.org
> Date: Sat, 04 Jun 2022 20:46:06 +0900
> 
> > I thought no changes were needed, because one could already say
> > something like
> >
> >   make ja-dic.el JA_DIC_REDUCTION=''
> >
> > Wouldn't that work?  Perhaps in lisp/Makefile.in one should make sure
> > the JA_DIC_REDUCTION value was propagated to the sub-make running in
> > leim/, but that's all.
> >
> > Am I missing something?
> >
> > Thanks.
> 
> I may not have understood it properly.
> 
> Did you imagine that the special rules for re-generating would be
> implemented in lisp/Makefile?

The rules for re-generating with and without reduction should be in
one place, in the same Makefile.

> In that case, are you saying that there is no need to modify the
> leim/Makefile?
> Certainly, if that is the case, you may be right.
> 
> For example, what about the idea of adding regenerate-ja-dic and
> regenerate-ja-dic-without-reductoin as new special rules in
> lisp/Makefile?

I'm saying that the same rule could serve both kinds of re-generation,
because the only difference is the single option passed to the Emacs
invocation.  We can rely on the user who want to re-generate the
dictionary to supply the option manually, when invoking Make.

So no separate rules should be necessary.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> > I thought no changes were needed, because one could already say
>> > something like
>> >
>> >   make ja-dic.el JA_DIC_REDUCTION=''
>> >
>> > Wouldn't that work?  Perhaps in lisp/Makefile.in one should make sure
>> > the JA_DIC_REDUCTION value was propagated to the sub-make running in
>> > leim/, but that's all.
>> >
>> > Am I missing something?
>> >
>> > Thanks.
>>
>> I may not have understood it properly.
>>
>> Did you imagine that the special rules for re-generating would be
>> implemented in lisp/Makefile?
>
> The rules for re-generating with and without reduction should be in
> one place, in the same Makefile.
>
>> In that case, are you saying that there is no need to modify the
>> leim/Makefile?
>> Certainly, if that is the case, you may be right.
>>
>> For example, what about the idea of adding regenerate-ja-dic and
>> regenerate-ja-dic-without-reductoin as new special rules in
>> lisp/Makefile?
>
> I'm saying that the same rule could serve both kinds of re-generation,
> because the only difference is the single option passed to the Emacs
> invocation.  We can rely on the user who want to re-generate the
> dictionary to supply the option manually, when invoking Make.
>
> So no separate rules should be necessary.
I see. If that is the case, then the v1 patch may be sufficient. With
the current rules, the dictionary file has to be deleted beforehand by a
gen-clean rule or something to regenerate it, but should I add a rule to
force regeneration even if the dictionary file has already been
generated?

Sorry for my lack of understanding :’(

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: handa@gnu.org,  emacs-devel@gnu.org
> Date: Sun, 05 Jun 2022 01:39:00 +0900
> 
> 
> > So no separate rules should be necessary.
> I see. If that is the case, then the v1 patch may be sufficient. With
> the current rules, the dictionary file has to be deleted beforehand by a
> gen-clean rule or something to regenerate it, but should I add a rule to
> force regeneration even if the dictionary file has already been
> generated?
> 
> Sorry for my lack of understanding :’(

It is I who should be apologizing for my inability to explain myself
clearly.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> From: Taiju HIGASHI <higashi@taiju.info>
>> Cc: handa@gnu.org,  emacs-devel@gnu.org
>> Date: Sun, 05 Jun 2022 01:39:00 +0900
>>
>>
>> > So no separate rules should be necessary.
>> I see. If that is the case, then the v1 patch may be sufficient. With
>> the current rules, the dictionary file has to be deleted beforehand by a
>> gen-clean rule or something to regenerate it, but should I add a rule to
>> force regeneration even if the dictionary file has already been
>> generated?
>>
>> Sorry for my lack of understanding :’(
>
> It is I who should be apologizing for my inability to explain myself
> clearly.

Then, based on the v1 patch, I will wait for Handa-san's review and then
I will make the final patch.
Is this policy correct?

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: handa@gnu.org,  emacs-devel@gnu.org
> Date: Sun, 05 Jun 2022 02:01:33 +0900
> 
> 
> >> From: Taiju HIGASHI <higashi@taiju.info>
> >> Cc: handa@gnu.org,  emacs-devel@gnu.org
> >> Date: Sun, 05 Jun 2022 01:39:00 +0900
> >>
> >>
> >> > So no separate rules should be necessary.
> >> I see. If that is the case, then the v1 patch may be sufficient. With
> >> the current rules, the dictionary file has to be deleted beforehand by a
> >> gen-clean rule or something to regenerate it, but should I add a rule to
> >> force regeneration even if the dictionary file has already been
> >> generated?
> >>
> >> Sorry for my lack of understanding :’(
> >
> > It is I who should be apologizing for my inability to explain myself
> > clearly.
> 
> Then, based on the v1 patch, I will wait for Handa-san's review and then
> I will make the final patch.
> Is this policy correct?

Yes, thanks.

We also wait for your paperwork (which already started) to be
completed.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[handa]

In article <878rqce51u.fsf@taiju.info>, Taiju HIGASHI <higashi@taiju.info> writes:

> Then, based on the v1 patch, I will wait for Handa-san's review and then
> I will make the final patch.

It seems that your patch has no prolem.

> The reduction of dictionaries would reduce the file size by less than
> half. While significant, how important is this in today's computing
> environment?
> In my personal opinion, I feel that reducing the vocabulary of the
> dictionary has more disadvantages than advantages.

I agree.

---
K. Handa
handa@gnu.org

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Hi Handa,


>> Then, based on the v1 patch, I will wait for Handa-san's review and then
>> I will make the final patch.
>
> It seems that your patch has no prolem.

Thank you for reviewing!

>> The reduction of dictionaries would reduce the file size by less than
>> half. While significant, how important is this in today's computing
>> environment?
>> In my personal opinion, I feel that reducing the vocabulary of the
>> dictionary has more disadvantages than advantages.
>
> I agree.
I know newcomers like me shouldn't be making important decisions, but do
you think it would be realistic to remove the ability to reduce the
vocabulary of the Japanese dictionary or make the ability to reduce
the vocabulary of the Japanese dictionary optional, in a future version
of the dictionary?

Even if we apply this patch this time, we should consider what it should
be in the future.

I feel that many Japanese users do not use the standard Emacs Kanji
conversion function, but I feel that this may just be an unfair
assessment due to the limited vocabulary of the dictionary.
If the dictionary vocabulary is not reduced, I expect that more users
will appreciate the standard Emacs Kanji conversion function as being
sufficient for their needs.

Best Regards,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[handa]

Hi,

In article <87czfncifi.fsf@taiju.info>, Taiju HIGASHI <higashi@taiju.info> writes:

> I know newcomers like me shouldn't be making important decisions, but do
> you think it would be realistic to remove the ability to reduce the
> vocabulary of the Japanese dictionary or make the ability to reduce
> the vocabulary of the Japanese dictionary optional, in a future version
> of the dictionary?

I myself think that ability is now useless, but as Emacs development
takes a conservative approach, and, unfortunately, Emacslisp does not
distinguish public and private functions, it is better to keep the
ability but just make the default configure not use it.

---
K. Handa
handa@gnu.org

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> I know newcomers like me shouldn't be making important decisions, but do
>> you think it would be realistic to remove the ability to reduce the
>> vocabulary of the Japanese dictionary or make the ability to reduce
>> the vocabulary of the Japanese dictionary optional, in a future version
>> of the dictionary?
>
> I myself think that ability is now useless, but as Emacs development
> takes a conservative approach, and, unfortunately, Emacslisp does not
> distinguish public and private functions, it is better to keep the
> ability but just make the default configure not use it.

Thank you for your opinion.

Your opinion is close to what I sent in my first email, and I would like
to do so if I may.

I have also received a suggestion to make it a run-time option instead
of a build-time option, is it possible to get some opinion on that?

Because I cannot see how many users would want a dictionary with a reduced
vocabulary.

I can say, at least for me, that I don't need it at all for Emacs in any
environment.

Thanks,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Lars Ingebrigtsen]

Taiju HIGASHI <higashi@taiju.info> writes:

> I have also received a suggestion to make it a run-time option instead
> of a build-time option, is it possible to get some opinion on that?
>
> Because I cannot see how many users would want a dictionary with a reduced
> vocabulary.

I think if everybody agrees that nobody would want a reduced dictionary,
then we should just stop reducing the dictionary -- no
configuration/run-time options necessary.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: handa <handa@gnu.org>,  eliz@gnu.org,  emacs-devel@gnu.org
> Date: Mon, 06 Jun 2022 16:14:01 +0200
> 
> Taiju HIGASHI <higashi@taiju.info> writes:
> 
> > I have also received a suggestion to make it a run-time option instead
> > of a build-time option, is it possible to get some opinion on that?
> >
> > Because I cannot see how many users would want a dictionary with a reduced
> > vocabulary.
> 
> I think if everybody agrees that nobody would want a reduced dictionary,
> then we should just stop reducing the dictionary -- no
> configuration/run-time options necessary.

I think a configuration option, by default off, is a good way of
removing the reduction, but still leaving someone who may want that a
"fire escape".  The cost in complexity is quite small in this case.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> From: Lars Ingebrigtsen <larsi@gnus.org>
>> Cc: handa <handa@gnu.org>,  eliz@gnu.org,  emacs-devel@gnu.org
>> Date: Mon, 06 Jun 2022 16:14:01 +0200
>>
>> Taiju HIGASHI <higashi@taiju.info> writes:
>>
>> > I have also received a suggestion to make it a run-time option instead
>> > of a build-time option, is it possible to get some opinion on that?
>> >
>> > Because I cannot see how many users would want a dictionary with a reduced
>> > vocabulary.
>>
>> I think if everybody agrees that nobody would want a reduced dictionary,
>> then we should just stop reducing the dictionary -- no
>> configuration/run-time options necessary.
>
> I think a configuration option, by default off, is a good way of
> removing the reduction, but still leaving someone who may want that a
> "fire escape".  The cost in complexity is quite small in this case.

I feel that the run-time option is unnecessary if a dictionary with a
reduced vocabulary can be selected at installation if desired, and if it
can be rebuilt later.
(This package can also regenerate dictionaries with the Emacs command.)

Based on the totality of the discussion so far, would the following
policy be the best?

1. make the build-time option selectable to install a dictionary with a
   reduced vocabulary
2. install dictionaries without reduced vocabulary by default
3. make it possible to regenerate dictionaries from make or Emacs
   command.

Thanks,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: Lars Ingebrigtsen <larsi@gnus.org>,  handa@gnu.org,  emacs-devel@gnu.org
> Date: Tue, 07 Jun 2022 00:08:01 +0900
> 
> > I think a configuration option, by default off, is a good way of
> > removing the reduction, but still leaving someone who may want that a
> > "fire escape".  The cost in complexity is quite small in this case.
> 
> I feel that the run-time option is unnecessary if a dictionary with a
> reduced vocabulary can be selected at installation if desired, and if it
> can be rebuilt later.

I agree.

> Based on the totality of the discussion so far, would the following
> policy be the best?
> 
> 1. make the build-time option selectable to install a dictionary with a
>    reduced vocabulary
> 2. install dictionaries without reduced vocabulary by default
> 3. make it possible to regenerate dictionaries from make or Emacs
>    command.

I think 2+3 is the best.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

[-- Attachment #1: Type: text/plain, Size: 402 bytes --]


>> Based on the totality of the discussion so far, would the following
>> policy be the best?
>>
>> 1. make the build-time option selectable to install a dictionary with a
>>    reduced vocabulary
>> 2. install dictionaries without reduced vocabulary by default
>> 3. make it possible to regenerate dictionaries from make or Emacs
>>    command.
>
> I think 2+3 is the best.

I attached the v3 patch.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: v3 patch --]
[-- Type: text/x-patch, Size: 6906 bytes --]

From 615f07c0a6c565b70838834d319ff97407d1e59c Mon Sep 17 00:00:00 2001
From: Taiju HIGASHI <higashi@taiju.info>
Date: Tue, 7 Jun 2022 09:21:10 +0900
Subject: [PATCH v3] The vocabulary in ja-dic.el should not be reduced by
 default.

* configure.ac: Add "with-small-ja-dic" configure option.
* leim/Makefile.in (${leimdir}/ja-dic/ja-dic.el): Change the build
method depending on whether or not the with-small-ja-dic option is
specified.
* lisp/international/ja-dic-cnv.el (skkdic-convert-okuri-nasi): Add
"with-reduction" argument and change "skkdic-reduced-candidates" not
to be called if the "with-reduction" argument is
unspecified. (breaking changes)
(skkdic-convert): Add "with-reduction" optional argument.
(batch-skkdic-convert): Add "--with-reduction" command line argument.
---
 configure.ac                     |  5 +++++
 leim/Makefile.in                 |  8 +++++++-
 lisp/international/ja-dic-cnv.el | 26 ++++++++++++++++++--------
 3 files changed, 30 insertions(+), 9 deletions(-)

diff --git a/configure.ac b/configure.ac
index 313a1436b5..3e6eab94f8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -491,6 +491,7 @@ OPTION_DEFAULT_ON([threads],[don't compile with elisp threading support])
 OPTION_DEFAULT_OFF([native-compilation],[compile with Emacs Lisp native compiler support])
 OPTION_DEFAULT_OFF([cygwin32-native-compilation],[use native compilation on 32-bit Cygwin])
 OPTION_DEFAULT_ON([xinput2],[don't use version 2 of the X Input Extension for input])
+OPTION_DEFAULT_OFF([small-ja-dic],[generate a small-sized Japanese dictionary])
 
 AC_ARG_WITH([file-notification],[AS_HELP_STRING([--with-file-notification=LIB],
  [use a file notification library (LIB one of: yes, inotify, kqueue, gfile, w32, no)])],
@@ -6492,6 +6493,7 @@ AS_ECHO(["  Does Emacs use -lXaw3d?                                 ${HAVE_XAW3D
   Which dumping strategy does Emacs use?                  ${with_dumping}
   Does Emacs have native lisp compiler?                   ${HAVE_NATIVE_COMP}
   Does Emacs use version 2 of the the X Input Extension?  ${HAVE_XINPUT2}
+  Should Emacs use a small-sized Japanese dictionary?     ${with_small_ja_dic}
 "])
 
 if test -n "${EMACSDATA}"; then
@@ -6590,6 +6592,9 @@ SUBDIR_MAKEFILES_IN=`echo " ${SUBDIR_MAKEFILES}" | sed -e 's| | $(srcdir)/|g' -e
 
 AC_SUBST(SUBDIR_MAKEFILES_IN)
 
+SMALL_JA_DIC=$with_small_ja_dic
+AC_SUBST(SMALL_JA_DIC)
+
 dnl You might wonder (I did) why epaths.h is generated by running make,
 dnl rather than just letting configure generate it from epaths.in.
 dnl One reason is that the various paths are not fully expanded (see above);
diff --git a/leim/Makefile.in b/leim/Makefile.in
index 3b4216c0b8..a256ca539b 100644
--- a/leim/Makefile.in
+++ b/leim/Makefile.in
@@ -32,6 +32,12 @@ leimdir = ${srcdir}/../lisp/leim
 
 EXEEXT = @EXEEXT@
 
+SMALL_JA_DIC = @SMALL_JA_DIC@
+JA_DIC_REDUCTION_OPTION =
+ifeq ($(SMALL_JA_DIC), yes)
+	JA_DIC_REDUCTION_OPTION = --with-reduction
+endif
+
 -include ${top_builddir}/src/verbose.mk
 
 # Prevent any settings in the user environment causing problems.
@@ -134,7 +140,7 @@ generate-ja-dic: ${leimdir}/ja-dic/ja-dic.el
 ${leimdir}/ja-dic/ja-dic.el: $(srcdir)/SKK-DIC/SKK-JISYO.L
 	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
 	  --eval "(setq max-specpdl-size 5000)" \
-	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" "$<"
+	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" $(JA_DIC_REDUCTION_OPTION) "$<"
 
 ${srcdir}/../lisp/language/pinyin.el: ${srcdir}/MISC-DIC/pinyin.map
 	$(AM_V_GEN)${RUN_EMACS} -l titdic-cnv -f pinyin-convert $< $@
diff --git a/lisp/international/ja-dic-cnv.el b/lisp/international/ja-dic-cnv.el
index 7f7c0261dc..7451773912 100644
--- a/lisp/international/ja-dic-cnv.el
+++ b/lisp/international/ja-dic-cnv.el
@@ -295,7 +295,7 @@
       (setq skkdic-okuri-nasi-entries-count (length skkdic-okuri-nasi-entries))
       (progress-reporter-done progress))))
 
-(defun skkdic-convert-okuri-nasi (skkbuf buf)
+(defun skkdic-convert-okuri-nasi (skkbuf buf with-reduction)
   (with-current-buffer buf
     (insert ";; Setting okuri-nasi entries.\n"
 	    "(skkdic-set-okuri-nasi\n")
@@ -311,7 +311,9 @@
           (setq count (1+ count))
           (progress-reporter-update progress count)
 	  (if (setq candidates
-		    (skkdic-reduced-candidates skkbuf kana candidates))
+		    (if with-reduction
+                        (skkdic-reduced-candidates skkbuf kana candidates)
+                      candidates))
 	      (progn
 		(insert "\"" kana)
 		(while candidates
@@ -322,10 +324,12 @@
       (progress-reporter-done progress))
     (insert ")\n\n")))
 
-(defun skkdic-convert (filename &optional dirname)
+(defun skkdic-convert (filename &optional dirname with-reduction)
   "Generate Emacs Lisp file from Japanese dictionary file FILENAME.
 The format of the dictionary file should be the same as SKK dictionaries.
-Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
+Saves the output as `ja-dic-filename', in directory DIRNAME (if specified).
+When WITH-REDUCTION is t, then reduce dictionary vocabulary.
+"
   (interactive "FSKK dictionary file: ")
   (let* ((skkbuf (get-buffer-create " *skkdic-unannotated*"))
 	 (buf (get-buffer-create "*skkdic-work*")))
@@ -389,7 +393,7 @@ Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
 	(skkdic-collect-okuri-nasi)
 
 	;; Convert okuri-nasi general entries.
-	(skkdic-convert-okuri-nasi skkbuf buf)
+	(skkdic-convert-okuri-nasi skkbuf buf with-reduction)
 
 	;; Postfix
 	(with-current-buffer buf
@@ -427,15 +431,21 @@ To get complete usage, invoke:
 	(message "To convert SKK-JISYO.L into skkdic.el:")
 	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L")
 	(message "To convert SKK-JISYO.L into DIR/ja-dic.el:")
-	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L"))
-    (let (targetdir filename)
+	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L")
+        (message "To convert SKK-JISYO.L into skkdic.el with reduce dictionary vocabulary:")
+        (message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L --with-reduction"))
+    (let (targetdir filename with-reduction)
       (if (string= (car command-line-args-left) "-dir")
 	  (progn
 	    (setq command-line-args-left (cdr command-line-args-left))
 	    (setq targetdir (expand-file-name (car command-line-args-left)))
 	    (setq command-line-args-left (cdr command-line-args-left))))
+      (if (string= (car command-line-args-left) "--with-reduction")
+          (progn
+	    (setq with-reduction t)
+	    (setq command-line-args-left (cdr command-line-args-left))))
       (setq filename (expand-file-name (car command-line-args-left)))
-      (skkdic-convert filename targetdir)))
+      (skkdic-convert filename targetdir with-reduction)))
   (kill-emacs 0))
 
 
-- 
2.36.1


[-- Attachment #3: Type: text/plain, Size: 306 bytes --]


I thought that if reducing vocabulary is an option, some people might
question what reducing vocabulary means. So, to make it easier to convey
the intent, I changed the configure option to with-small-ja-dic and also
changed the description.

Please point out if the original is better.

Thanks,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>>> Based on the totality of the discussion so far, would the following
>>> policy be the best?
>>>
>>> 1. make the build-time option selectable to install a dictionary with a
>>>    reduced vocabulary
>>> 2. install dictionaries without reduced vocabulary by default
>>> 3. make it possible to regenerate dictionaries from make or Emacs
>>>    command.
>>
>> I think 2+3 is the best.
>
> I attached the v3 patch.
>
>
>
> I thought that if reducing vocabulary is an option, some people might
> question what reducing vocabulary means. So, to make it easier to convey
> the intent, I changed the configure option to with-small-ja-dic and also
> changed the description.
>
> Please point out if the original is better.
>
> Thanks,

Sorry, I had forgotten that handa-san had advised me to add functions
and commands separately and avoid making backward incompatible changes.

I will make another patch later.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

> Sorry, I had forgotten that handa-san had advised me to add functions
> and commands separately and avoid making backward incompatible changes.
>
> I will make another patch later.

Sorry for sending several emails.

I was going to go into the implementation, but I am beginning to feel
that it is inconsistent that the default behavior has changed, but the
behavior of commands and functions has remained the same.

For example, if the batch-skkdic-convert-without-reduction command is
added separately, users who dare to regenerate a Japanese dictionary
with a reduced vocabulary will execute batch-skkdic-convert.
It is disconcerting that a command that we can assume is used by default
behaves the same as commands in the past.

I would like to get other people's opinions on whether I should do that,
since it has the side effect of making the code less maintainable.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
> Date: Tue, 07 Jun 2022 12:50:31 +0900
> 
> 
> > Sorry, I had forgotten that handa-san had advised me to add functions
> > and commands separately and avoid making backward incompatible changes.
> >
> > I will make another patch later.
> 
> Sorry for sending several emails.
> 
> I was going to go into the implementation, but I am beginning to feel
> that it is inconsistent that the default behavior has changed, but the
> behavior of commands and functions has remained the same.
> 
> For example, if the batch-skkdic-convert-without-reduction command is
> added separately, users who dare to regenerate a Japanese dictionary
> with a reduced vocabulary will execute batch-skkdic-convert.
> It is disconcerting that a command that we can assume is used by default
> behaves the same as commands in the past.
> 
> I would like to get other people's opinions on whether I should do that,
> since it has the side effect of making the code less maintainable.

We don't need to change the default behavior of the existing
commands.  We can leave the default behavior the same, and introduce
optional arguments to produce the new behavior.  We just want the
build process to use this new behavior by default, and allow users to
request the old behavior at configure time.

OK?

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Lars Ingebrigtsen]

Taiju HIGASHI <higashi@taiju.info> writes:

> Sorry, I had forgotten that handa-san had advised me to add functions
> and commands separately and avoid making backward incompatible changes.

I don't think that's necessary -- I doubt that `skkdic-convert' is used
by anybody outside of the Emacs build process.  Changing what it does is
fine (but it should be noted in the NEWS file that it has changed).

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Taiju HIGASHI <higashi@taiju.info> writes:
>
>> Sorry, I had forgotten that handa-san had advised me to add functions
>> and commands separately and avoid making backward incompatible changes.
>
> I don't think that's necessary -- I doubt that `skkdic-convert' is used
> by anybody outside of the Emacs build process.  Changing what it does is
> fine (but it should be noted in the NEWS file that it has changed).

That may indeed be true.
I have had a experience using it outside of the Emacs build process, but
I only used it to generate the Japanese dictionary for the vocabulary
with no reduction, after customizing the function, so it seems unlikely
that I will use it in the future.

Can I leave the editing of NEWS to the maintainer? I have no confidence
as I think accurate English text is required.

Thanks,
-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Lars Ingebrigtsen]

Taiju HIGASHI <higashi@taiju.info> writes:

> Can I leave the editing of NEWS to the maintainer? I have no confidence
> as I think accurate English text is required.

Sure, that's fine.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
> Date: Tue, 07 Jun 2022 09:47:31 +0900
> 
> I thought that if reducing vocabulary is an option, some people might
> question what reducing vocabulary means. So, to make it easier to convey
> the intent, I changed the configure option to with-small-ja-dic and also
> changed the description.

Thanks, but I thought we all agreed that by default the build should
NOT reduce the vocabulary, so that use of the option would be needed
only if someone would want the current behavior back?  And this v3
does it the other way around?  Or did I miss something?

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Taiju HIGASHI <higashi@taiju.info>
>> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
>> Date: Tue, 07 Jun 2022 09:47:31 +0900
>>
>> I thought that if reducing vocabulary is an option, some people might
>> question what reducing vocabulary means. So, to make it easier to convey
>> the intent, I changed the configure option to with-small-ja-dic and also
>> changed the description.
>
> Thanks, but I thought we all agreed that by default the build should
> NOT reduce the vocabulary, so that use of the option would be needed
> only if someone would want the current behavior back?  And this v3
> does it the other way around?  Or did I miss something?

Indeed. There was absolutely no need to change the behavior of existing
functions or commands. It was an unnecessary modification.

I'm going to modify the v3 patch to allow the behavior to be changed by
optional arguments, without changing the past behavior of existing
functions and commands.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Taiju HIGASHI <higashi@taiju.info>
>> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
>> Date: Tue, 07 Jun 2022 09:47:31 +0900
>>
>> I thought that if reducing vocabulary is an option, some people might
>> question what reducing vocabulary means. So, to make it easier to convey
>> the intent, I changed the configure option to with-small-ja-dic and also
>> changed the description.
>
> Thanks, but I thought we all agreed that by default the build should
> NOT reduce the vocabulary, so that use of the option would be needed
> only if someone would want the current behavior back?  And this v3
> does it the other way around?  Or did I miss something?

The default behavior of the build process of the v3 patch does not
reduce the vocabulary, but It reduces the vocabulary when the
--with-small-ja-dic option is specified.

However, as you pointed out in another email, I have made the
inappropriate correction in the ja-dic-cnv.el, So I fix it.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

[-- Attachment #1: Type: text/plain, Size: 760 bytes --]

Taiju HIGASHI <higashi@taiju.info> writes:

>> Thanks, but I thought we all agreed that by default the build should
>> NOT reduce the vocabulary, so that use of the option would be needed
>> only if someone would want the current behavior back?  And this v3
>> does it the other way around?  Or did I miss something?
>
> The default behavior of the build process of the v3 patch does not
> reduce the vocabulary, but It reduces the vocabulary when the
> --with-small-ja-dic option is specified.
>
> However, as you pointed out in another email, I have made the
> inappropriate correction in the ja-dic-cnv.el, So I fix it.

I have attached the v4 patch. Sorry for the many times I have taken your
time with this simple code modification.
Please check it out.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: v4 patch --]
[-- Type: text/x-patch, Size: 6909 bytes --]

From b91102b60575904bd2726e146be4b85f33161f5d Mon Sep 17 00:00:00 2001
From: Taiju HIGASHI <higashi@taiju.info>
Date: Tue, 7 Jun 2022 21:46:14 +0900
Subject: [PATCH v4] The vocabulary in ja-dic.el should not be reduced by
 default.

* configure.ac: Add the "with-small-ja-dic" configure option.
* leim/Makefile.in (${leimdir}/ja-dic/ja-dic.el): Change the build
method depending on whether or not the with-small-ja-dic option is
specified.
* lisp/international/ja-dic-cnv.el (skkdic-convert-okuri-nasi): Add
the "no-reduction" optional argument. When it is specified, then
generate a Japanese dictionary with no reduced vocabulary.
(skkdic-convert): Add the "no-reduction" optional argument.
(batch-skkdic-convert): Add the "--no-reduction" command line argument.
---
 configure.ac                     |  5 +++++
 leim/Makefile.in                 |  8 +++++++-
 lisp/international/ja-dic-cnv.el | 26 ++++++++++++++++++--------
 3 files changed, 30 insertions(+), 9 deletions(-)

diff --git a/configure.ac b/configure.ac
index 313a1436b5..3e6eab94f8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -491,6 +491,7 @@ OPTION_DEFAULT_ON([threads],[don't compile with elisp threading support])
 OPTION_DEFAULT_OFF([native-compilation],[compile with Emacs Lisp native compiler support])
 OPTION_DEFAULT_OFF([cygwin32-native-compilation],[use native compilation on 32-bit Cygwin])
 OPTION_DEFAULT_ON([xinput2],[don't use version 2 of the X Input Extension for input])
+OPTION_DEFAULT_OFF([small-ja-dic],[generate a small-sized Japanese dictionary])
 
 AC_ARG_WITH([file-notification],[AS_HELP_STRING([--with-file-notification=LIB],
  [use a file notification library (LIB one of: yes, inotify, kqueue, gfile, w32, no)])],
@@ -6492,6 +6493,7 @@ AS_ECHO(["  Does Emacs use -lXaw3d?                                 ${HAVE_XAW3D
   Which dumping strategy does Emacs use?                  ${with_dumping}
   Does Emacs have native lisp compiler?                   ${HAVE_NATIVE_COMP}
   Does Emacs use version 2 of the the X Input Extension?  ${HAVE_XINPUT2}
+  Should Emacs use a small-sized Japanese dictionary?     ${with_small_ja_dic}
 "])
 
 if test -n "${EMACSDATA}"; then
@@ -6590,6 +6592,9 @@ SUBDIR_MAKEFILES_IN=`echo " ${SUBDIR_MAKEFILES}" | sed -e 's| | $(srcdir)/|g' -e
 
 AC_SUBST(SUBDIR_MAKEFILES_IN)
 
+SMALL_JA_DIC=$with_small_ja_dic
+AC_SUBST(SMALL_JA_DIC)
+
 dnl You might wonder (I did) why epaths.h is generated by running make,
 dnl rather than just letting configure generate it from epaths.in.
 dnl One reason is that the various paths are not fully expanded (see above);
diff --git a/leim/Makefile.in b/leim/Makefile.in
index 3b4216c0b8..29b9f3b2f8 100644
--- a/leim/Makefile.in
+++ b/leim/Makefile.in
@@ -32,6 +32,12 @@ leimdir = ${srcdir}/../lisp/leim
 
 EXEEXT = @EXEEXT@
 
+SMALL_JA_DIC = @SMALL_JA_DIC@
+JA_DIC_NO_REDUCTION_OPTION = --no-reduction
+ifeq ($(SMALL_JA_DIC), yes)
+	JA_DIC_NO_REDUCTION_OPTION =
+endif
+
 -include ${top_builddir}/src/verbose.mk
 
 # Prevent any settings in the user environment causing problems.
@@ -134,7 +140,7 @@ generate-ja-dic: ${leimdir}/ja-dic/ja-dic.el
 ${leimdir}/ja-dic/ja-dic.el: $(srcdir)/SKK-DIC/SKK-JISYO.L
 	$(AM_V_GEN)$(RUN_EMACS) -batch -l ja-dic-cnv \
 	  --eval "(setq max-specpdl-size 5000)" \
-	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" "$<"
+	  -f batch-skkdic-convert -dir "$(leimdir)/ja-dic" $(JA_DIC_NO_REDUCTION_OPTION) "$<"
 
 ${srcdir}/../lisp/language/pinyin.el: ${srcdir}/MISC-DIC/pinyin.map
 	$(AM_V_GEN)${RUN_EMACS} -l titdic-cnv -f pinyin-convert $< $@
diff --git a/lisp/international/ja-dic-cnv.el b/lisp/international/ja-dic-cnv.el
index 7f7c0261dc..f0929410ea 100644
--- a/lisp/international/ja-dic-cnv.el
+++ b/lisp/international/ja-dic-cnv.el
@@ -295,7 +295,7 @@
       (setq skkdic-okuri-nasi-entries-count (length skkdic-okuri-nasi-entries))
       (progress-reporter-done progress))))
 
-(defun skkdic-convert-okuri-nasi (skkbuf buf)
+(defun skkdic-convert-okuri-nasi (skkbuf buf &optional no-reduction)
   (with-current-buffer buf
     (insert ";; Setting okuri-nasi entries.\n"
 	    "(skkdic-set-okuri-nasi\n")
@@ -311,7 +311,9 @@
           (setq count (1+ count))
           (progress-reporter-update progress count)
 	  (if (setq candidates
-		    (skkdic-reduced-candidates skkbuf kana candidates))
+                    (if no-reduction
+                        candidates
+                        (skkdic-reduced-candidates skkbuf kana candidates)))
 	      (progn
 		(insert "\"" kana)
 		(while candidates
@@ -322,10 +324,12 @@
       (progress-reporter-done progress))
     (insert ")\n\n")))
 
-(defun skkdic-convert (filename &optional dirname)
+(defun skkdic-convert (filename &optional dirname no-reduction)
   "Generate Emacs Lisp file from Japanese dictionary file FILENAME.
 The format of the dictionary file should be the same as SKK dictionaries.
-Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
+Saves the output as `ja-dic-filename', in directory DIRNAME (if specified).
+When NO-REDUCTION is t, then not reduce dictionary vocabulary.
+"
   (interactive "FSKK dictionary file: ")
   (let* ((skkbuf (get-buffer-create " *skkdic-unannotated*"))
 	 (buf (get-buffer-create "*skkdic-work*")))
@@ -389,7 +393,7 @@ Saves the output as `ja-dic-filename', in directory DIRNAME (if specified)."
 	(skkdic-collect-okuri-nasi)
 
 	;; Convert okuri-nasi general entries.
-	(skkdic-convert-okuri-nasi skkbuf buf)
+	(skkdic-convert-okuri-nasi skkbuf buf no-reduction)
 
 	;; Postfix
 	(with-current-buffer buf
@@ -427,15 +431,21 @@ To get complete usage, invoke:
 	(message "To convert SKK-JISYO.L into skkdic.el:")
 	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert SKK-JISYO.L")
 	(message "To convert SKK-JISYO.L into DIR/ja-dic.el:")
-	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L"))
-    (let (targetdir filename)
+	(message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert -dir DIR SKK-JISYO.L")
+        (message "To convert SKK-JISYO.L into skkdic.el with not reduce dictionary vocabulary:")
+        (message "  %% emacs -batch -l ja-dic-cnv -f batch-skkdic-convert --no-reduction SKK-JISYO.L"))
+    (let (targetdir filename no-reduction)
       (if (string= (car command-line-args-left) "-dir")
 	  (progn
 	    (setq command-line-args-left (cdr command-line-args-left))
 	    (setq targetdir (expand-file-name (car command-line-args-left)))
 	    (setq command-line-args-left (cdr command-line-args-left))))
+      (if (string= (car command-line-args-left) "--no-reduction")
+          (progn
+	    (setq no-reduction t)
+	    (setq command-line-args-left (cdr command-line-args-left))))
       (setq filename (expand-file-name (car command-line-args-left)))
-      (skkdic-convert filename targetdir)))
+      (skkdic-convert filename targetdir no-reduction)))
   (kill-emacs 0))
 
 
-- 
2.36.1


[-- Attachment #3: Type: text/plain, Size: 10 bytes --]

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Hi Eli,

Is there still a problem with the v4 patch I submitted the other day?

I'm sorry if there are still some misunderstandings.
It is possible that I don't understand, and I would be glad if you could
enlighten me.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
> Date: Thu, 09 Jun 2022 22:10:05 +0900
> 
> Is there still a problem with the v4 patch I submitted the other day?

I didn't yet review that patch.  I will do so soon.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Taiju HIGASHI <higashi@taiju.info>
> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
> Date: Tue, 07 Jun 2022 22:08:12 +0900
> 
> I have attached the v4 patch. Sorry for the many times I have taken your
> time with this simple code modification.
> Please check it out.

Thanks, I've now installed this on the master branch.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> From: Taiju HIGASHI <higashi@taiju.info>
>> Cc: larsi@gnus.org,  handa@gnu.org,  emacs-devel@gnu.org
>> Date: Tue, 07 Jun 2022 22:08:12 +0900
>>
>> I have attached the v4 patch. Sorry for the many times I have taken your
>> time with this simple code modification.
>> Please check it out.
>
> Thanks, I've now installed this on the master branch.

Thank you very much!
And thank you for correcting the text and creating the NEWS entry!

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > Subject: Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

Assuming there is a good reason to have this dictionary in Emacs,
would it perhaps be better to make the size of Japanese vocabulary a
run-time choice for the user?  Instead of abuild-time option, that is.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

Hi,

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > Subject: Re: [PATCH] Add an option to not reduce vocabulary of the Japanese
>
> Assuming there is a good reason to have this dictionary in Emacs,
> would it perhaps be better to make the size of Japanese vocabulary a
> run-time choice for the user?  Instead of abuild-time option, that is.

Thank you for your suggestion.

I figured that if I initially selected a dictionary without a reduced
vocabulary, it was infinitely unlikely that I would change to a
dictionary with a reduced vocabulary, so I just had to select one at the
time of construction.
Conversely, a user who selects a dictionary with a reduced vocabulary is
unlikely to change to a non-reduced vocabulary dictionary in the middle
of the process, as there must be some reason, such as a lack of machine
resources.
The cost of maintaining two different dictionaries is also a concern.

However, this is my personal opinion and other users may have different
opinions. I would like to hear Harada-san's opinion and others to
consider. Thank you very much.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

> However, this is my personal opinion and other users may have different
> opinions. I would like to hear Harada-san's opinion and others to
> consider. Thank you very much.

I had mistakenly written Handa-san's name. Sorry about that.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > I figured that if I initially selected a dictionary without a reduced
  > vocabulary, it was infinitely unlikely that I would change to a
  > dictionary with a reduced vocabulary, so I just had to select one at the
  > time of construction.
  > Conversely, a user who selects a dictionary with a reduced vocabulary is
  > unlikely to change to a non-reduced vocabulary dictionary in the middle
  > of the process, as there must be some reason, such as a lack of machine
  > resources.
  > The cost of maintaining two different dictionaries is also a concern.

There are still some computers that multiple users share.
If there is one installed Emacs on that computer, some users
may want to use the larger vocabulary and some the smaller.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>   > I figured that if I initially selected a dictionary without a reduced
>   > vocabulary, it was infinitely unlikely that I would change to a
>   > dictionary with a reduced vocabulary, so I just had to select one at the
>   > time of construction.
>   > Conversely, a user who selects a dictionary with a reduced vocabulary is
>   > unlikely to change to a non-reduced vocabulary dictionary in the middle
>   > of the process, as there must be some reason, such as a lack of machine
>   > resources.
>   > The cost of maintaining two different dictionaries is also a concern.
>
> There are still some computers that multiple users share.
> If there is one installed Emacs on that computer, some users
> may want to use the larger vocabulary and some the smaller.

I see. I didn't have enough imagination.
As you say, it would be useful to have the option at run-time.
I don't know if I can do it well, but I will give it a try.

-- 
Taiju

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > The Japanese dictionary bundled with Emacs has a small vocabulary.

Why do we include a Japanese dictionary in Emacs?
Free dictionaries are good things, but we don't normally package them
as part of Emacs.

Is this needed for a Japanese input method?


-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> Cc: emacs-devel@gnu.org, higashi@taiju.info
> Date: Fri, 03 Jun 2022 19:52:32 -0400
> 
>   > The Japanese dictionary bundled with Emacs has a small vocabulary.
> 
> Why do we include a Japanese dictionary in Emacs?
> Free dictionaries are good things, but we don't normally package them
> as part of Emacs.
> 
> Is this needed for a Japanese input method?

Yes, it's needed for the Japanese input methods.  That's the only
reason we have that dictionary in Emacs.

Re: [PATCH] Add an option to not reduce vocabulary of the Japanese

[Taiju HIGASHI]

>> From: Richard Stallman <rms@gnu.org>
>> Cc: emacs-devel@gnu.org, higashi@taiju.info
>> Date: Fri, 03 Jun 2022 19:52:32 -0400
>>
>>   > The Japanese dictionary bundled with Emacs has a small vocabulary.
>>
>> Why do we include a Japanese dictionary in Emacs?
>> Free dictionaries are good things, but we don't normally package them
>> as part of Emacs.
>>
>> Is this needed for a Japanese input method?
>
> Yes, it's needed for the Japanese input methods.  That's the only
> reason we have that dictionary in Emacs.

I do not know the historical background. But I am sure that a dictionary
is necessary for Japanese input methods.
Personally, I am very grateful to have a dictionary optimized for use
with Emacs.
Because Japanese input requires frequent Kanji conversion, for which low
latency is very important.

-- 
Taiju