Re: [PATCH] Improve Malayalam language transliteration

[Eli Zaretskii]

> From: James Thomas <address@hidden>
> Date: Sun, 26 Apr 2020 21:51:06 +0530
> 
> >> -(if nil
> >> -    (quail-define-package "malayalam-itrans" "Malayalam" "MlmIT" t 
> >> "Malayalam ITRANS"))
> >> -(quail-define-indian-trans-package
> >> - indian-mlm-itrans-v5-hash "malayalam-itrans" "Malayalam" "MlmIT"
> >> - "Malayalam transliteration by ITRANS method.")
> >
> > I'm not sure I understand: does this delete the ITRANS method?  Would
> > it make sense to leave it, and just add the new one?  And please also
> > note that lisp/language/indian.el references the malayalam-itrans
> > method; should that be changed as well?
> 
> Yes this does delete and replace it.  I guess we could keep the ITRANS
> method also, but it's practically unusable for Malayalam (given its
> uniqueness and quirks) and not even defined for all its characters - for e.g.,
> the common characters ള (La), ഴ (zha), റ (rra) are missing. As I
> understand it, it was developed for Devanagari and was easily adapted
> to similar Indian languages. The Mozhi scheme is a way of expanding it
> to support Malayalam.

That's okay, but we don't like deleting existing capabilities without
some period during which it's deprecated.  So maybe say in NEWS that
ITRANS is deprecated, and add a comment in quail/indian.el that it
should be removed in some future release.

> Sorry I missed the reference in lisp/language/indian.el. It seems to be
> just defining the default input method for the language. Can be changed
> easily if decided.

If you think this new method is much better, it's okay to make it the
default, I think.  Just mention that in NEWS.

Thanks.

P.S. And please keep the list address on the CC, so that this
discussion is recorded.