Re: Moving japanese category to localised

[Alistair Crooks <agc@pkgsrc.org>]

On Thu, Feb 21, 2002 at 12:19:42PM +0900, Masao Uebayashi wrote:
> I, as I said in tech-pkg, agree with Alistair that "japanese" category
> should be deprecated.  After the discussion, I considered a bit
> different way of japanese category reorganization.  My idea is to
> classify "japanese" packages into other categories subject to the rule
> that "classified subject to its use" which is used for all other
> packages.
> 
> My classification looks like:
> 
> 	Package		New category
> 	------------------------------------------
> 	a2ps		textproc
> 	ack		textproc
> 	canna-lib	inputmethod
> 	canna-server	inputmethod
> 	canuum		imputmethod
> 	chasen		textproc
> 	chasen-base	textproc
> 	freewnn-lib	inputmethod
> 	freewnn-server	inputmethod
> 	gawk		lang
> 	grep		textproc or misc
> 	groff		textproc
> 	ipadic		textproc or dict
> 	jhd		misc
> 	kakasi		textproc
> 	kbanner		textproc
> 	kinput2		inputmethod
> 	less		misc
> 	man		misc
> 	mh		mail
> 	micq		chat
> 	mknmz-wwwoffle	www
> 	namazu1		textproc
> 	namazu2		textproc
> 	nkf		textproc
> 	p5-Jcode	textproc
> 	p5-Text-ChaSen	textproc
> 	p5-Text-Kakasi	textproc
> 	p5-jcode	textproc
> 	p5-nkf		textproc
> 	qkc		textproc
> 	sed		lang
> 	sj3-lib		inputmethod
> 	sj3-server	inputmethod
> 	skkfep		inputmethod
> 	skkinput	inputmethod
> 	skkserv		inputmethod
> 	texfamily	print
> 	texfamily-share	print
> 	vflib		print
> 	vflib-lib	print
> 	vflib-utils	print
> 	vfxdvik		print
> 	xjdic		dict
> 	xjman		dict or doc
> 
> It appeared to me that the only categories to be newly created are
> "inputmethod" and "dict".  "inputmethod" is for packages to input
> Japanese, and "dict" for data of Japanese characters and words used in
> many situations by "inputmethod" and "textproc" packages.
> 
> These two are very similar to the opinion someone (rxg?) argued in the
> previous discussion, and also apparently correspond to "localised"
> category you mentioned.

My thanks to Uebayashi-san for (a) his translation of my previous
mail, and (b) his insight and suggestions in this mail. My apologies
for taking so long to reply, but I have been considering things
carefully.  I do like the "inputmethod" category name, and think
it would be a valuable addition to pkgsrc, as it's also applicable
to a number of other localised systems.

My thanks also to Takahiro Kambe for his suggestions of better places
to put packages.

I have a couple of issues: I don't like the "dict" category - for all
the other dictionary-supporting packages for non-US English, such as
ispell-*, we put the supplementary dictionaries in the same category
as the main package, and I think that works well. Also, I don't
particularly like placing search engines in the "database" category.
For me, a database is something specific - usually a relational database,
or access mechanism for the same - and not as generic as some people
like to view it. I think the best places for namazu and friends is 
something like textproc.

Regards,
Alistair