Any use cases of `iso-transl-language-list'?

Any use cases of `iso-transl-language-list'?

[Tor Kringeland]

I was looking at `iso-transl.el' and came over
`iso-transl-language-alist'.  By calling `iso-transl-set-language' and
selecting one of the languages you can input certain characters using
one less key-stroke with `C-x 8'.  For example, by enabling Portuguese

  C-x 8 c

produces "ç", which can also be done with `C-x 8 , c'.  Except for
Esperanto and German, I do not see why this feature is necessary,
though.  In the first place, it would be much easier to use an input
method instead if you want to write in one of those language.  Moreover,
support for French, Portuguese and Spanish are incomplete, and could not
be completed using the same pattern.  /E.g./ in Portuguese you have both
"é" and "ê" so there's no good answer for what `C-x 8 e' should do if
you had enabled Portuguese; and one key-storke more (`C-x 8 ' e' or `C-x 8 ^ e')
would produce either character without having to choose
Portuguese as the input method.

This feature seems to have been implemented around 1995, and I would
guess a lot has changed since then with input methods in Emacs.  It does
no harm, of course, but is there really a use case for this anymore?  Is
it something that could be deprecated (at least for Portuguese, Spanish
and French, for which I cannot see any real use case with the current
implementation) in favor of input methods or using the regular `C-x 8'
map, which for each character only requires one key-stroke more?

Re: Any use cases of `iso-transl-language-list'?

[Eli Zaretskii]

> From: Tor Kringeland <tor.a.s.kringeland@ntnu.no>
> Date: Wed, 01 Dec 2021 21:40:45 +0100
> 
> This feature seems to have been implemented around 1995, and I would
> guess a lot has changed since then with input methods in Emacs.  It does
> no harm, of course, but is there really a use case for this anymore?  Is
> it something that could be deprecated (at least for Portuguese, Spanish
> and French, for which I cannot see any real use case with the current
> implementation) in favor of input methods or using the regular `C-x 8'
> map, which for each character only requires one key-stroke more?

If this does no harm, then what would we gain from deprecating it?
Some people might be using that, for whatever reasons, so the risk to
break someone's use patterns is there.  Are there any gains to justify
that?

Thanks.

Re: Any use cases of `iso-transl-language-list'?

[Po Lu]

Tor Kringeland <tor.a.s.kringeland@ntnu.no> writes:

> This feature seems to have been implemented around 1995, and I would
> guess a lot has changed since then with input methods in Emacs.  It does
> no harm, of course, but is there really a use case for this anymore?  Is
> it something that could be deprecated (at least for Portuguese, Spanish
> and French, for which I cannot see any real use case with the current
> implementation) in favor of input methods or using the regular `C-x 8'
> map, which for each character only requires one key-stroke more?

Please do not remove this feature.  It is useful for typing German and
Esperanto text, at the very least.

Re: Any use cases of `iso-transl-language-list'?

[Stefan Kangas]

Eli Zaretskii <eliz@gnu.org> writes:

> If this does no harm, then what would we gain from deprecating it?
> Some people might be using that, for whatever reasons, so the risk to
> break someone's use patterns is there.  Are there any gains to justify
> that?

I tend to agree.  I looked at it, and it seemed much less useful to me
than input methods, but at the same time maybe you don't want to always
have an input method enabled (for example, if you only use these
languages rarely).

It would be good to give `iso-transl-set-language' a docstring though.

Re: Any use cases of `iso-transl-language-list'?

[Tor Kringeland]

Po Lu <luangruo@yahoo.com> writes:

> Please do not remove this feature.  It is useful for typing German and
> Esperanto text, at the very least.

I concur.  For Esperanto and German this is a complete feature, and you
could write anything in those languages with these maps and it has
seniority.

However, for Spanish, Portuguese and French this feature is very
incomplete, and I'd argue that it'd be better to either remove this
feature for these languages or somehow extend it to offer full support
(which would be difficult without having the same number of key-strokes
as the regular `C-x 8' map).  As it is now, it might be misleading in
that it implies support for the character support of said languages,
while in reality it's only one character for each language that saves
one key-stroke.  Also, for Spanish "¡" and "¿" can be input the same way
without activating this map.

IMO seniority is not sufficient justification for keeping "support" for
these languages around, at least in the current incomplete state.

Could we at least mention input methods in the docstrings here, and
perhaps the issues around greater support?

Re: Any use cases of `iso-transl-language-list'?

[Stefan Kangas]

Tor Kringeland <tor.a.s.kringeland@ntnu.no> writes:

> However, for Spanish, Portuguese and French this feature is very
> incomplete, and I'd argue that it'd be better to either remove this
> feature for these languages or somehow extend it to offer full support
> (which would be difficult without having the same number of key-strokes
> as the regular `C-x 8' map).

When typing Spanish, the characters I have trouble inputting are exactly
¡, ¿, ñ and Ñ, but not the accented vowels (I use a prefix character ´
for them).  So it seems "complete" in the sense that it covers exactly
the characters I'd use.  Maybe your experience is different.

> Could we at least mention input methods in the docstrings here, and
> perhaps the issues around greater support?

Documenting this sounds like a good idea.  Could you write up a patch?

Re: Any use cases of `iso-transl-language-list'?

[Robert Pluim]

>>>>> On Thu, 02 Dec 2021 16:14:17 +0100, Tor Kringeland <tor.a.s.kringeland@ntnu.no> said:

    Tor> Po Lu <luangruo@yahoo.com> writes:
    >> Please do not remove this feature.  It is useful for typing German and
    >> Esperanto text, at the very least.

    Tor> I concur.  For Esperanto and German this is a complete feature, and you
    Tor> could write anything in those languages with these maps and it has
    Tor> seniority.

    Tor> However, for Spanish, Portuguese and French this feature is very
    Tor> incomplete, and I'd argue that it'd be better to either remove this
    Tor> feature for these languages or somehow extend it to offer full support
    Tor> (which would be difficult without having the same number of key-strokes
    Tor> as the regular `C-x 8' map).  As it is now, it might be misleading in
    Tor> that it implies support for the character support of said languages,
    Tor> while in reality it's only one character for each language that saves
    Tor> one key-stroke.  Also, for Spanish "¡" and "¿" can be input the same way
    Tor> without activating this map.

For eg French I donʼt see how to disambiguate eg é and è with a single
key without trampling on something else (assuming a US-ASCII style
keyboard). And those are already available on C-x 8 ' e and C-x 8 ` e,
so I donʼt know what you think is misleading.

    Tor> IMO seniority is not sufficient justification for keeping "support" for
    Tor> these languages around, at least in the current incomplete state.

As others have said: theyʼre not causing any harm.

    Tor> Could we at least mention input methods in the docstrings here, and
    Tor> perhaps the issues around greater support?

Input methods, yes. You'll have to expand on what you think these
'issues around greater support' are.

Robert
-- 

Re: Any use cases of `iso-transl-language-list'?

[Tor Kringeland]

Stefan Kangas <stefankangas@gmail.com> writes:

> When typing Spanish, the characters I have trouble inputting are exactly
> ¡, ¿, ñ and Ñ, but not the accented vowels (I use a prefix character ´
> for them).  So it seems "complete" in the sense that it covers exactly
> the characters I'd use.  Maybe your experience is different.

I see your point.  I use an US keyboard layout which doesn't have
modifiers for accents, so I've just gotten used to using `C-x 8' or
inputs methods, where producing "ñ" or "é" require the same amount of
key-strokes.  Though /e.g./ Norwegian keyboards have modifiers "´" and
"`" which one could also use for /e.g./ writing Italian.  There's also a
"^" modifier.  So this is more complete than I had thought, in that it
covers these more tricky characters.

I still think using the default `C-x 8' or input methods is a better
method in general, but in its current state this seems to have a use
case for said languages, so leaving it as is and documenting is better.

> Documenting this sounds like a good idea.  Could you write up a patch?

Yes

Re: Any use cases of `iso-transl-language-list'?

[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 have a feeling that the French, Spanish and Portuguese options were
only sketches of what was intended, never actually finished.
I think this feature is not useful any more and we might as well
mark it obsolete.

-- 
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)