Re: Some developement questions

[hw]

Eli Zaretskii <address@hidden> writes:

>> From: hw <address@hidden>
>> Cc: address@hidden,  address@hidden,  address@hidden,  address@hidden
>> Date: Mon, 10 Sep 2018 21:48:06 +0200
>> 
>> > The Alt key does work, you probably have your keyboard misconfigured.
>> 
>> I found out less than a week ago that it now finally works.  It did not
>> work for the first 25 years or so, beginning on an Atari ST.
>> 
>> Technically, it may have started working earlier and I didn't know
>> because hadn't tried it for a long time because it hasn't been working
>> for decades.
>
> It worked for me for the last 30 years.

It didn't do that for me.

> [...]
>> > Not Alt itself, some combinations that begin with Alt.  Like Alt-TAB,
>> > for example.  There are very few such combinations, and Emacs avoids
>> > binding important functions to them.  So I think explaining that in
>> > the tutorial would not be TRT, as the issue is quite obscure.
>> 
>> I don't know, isn't that hard to say?  If a WM (or something else) were
>> to use Alt+v for something, it might not work, or work intermittently,
>> in Emacs.
>
> We don't need to solve hypothetical problems, only those that happen
> in practice.  In practice, only a small number of ALT combinations are
> usurped by window managers, and they tend to be the same combinations
> in all WMs, for good practical reasons.  We shouldn't expect that set
> to grow or to usurp frequently used combinations, because that would
> cause user outcry.

You have no way of knowing what key combinations a user may have defined
with the WM or with other software that uses or provides binding things
to keys.  You do know that there can be race conditions between Emacs
and other software in the usage of key bindings.

What is the advantage of waiting until a user has problem, like the Alt
key not working in Emacs, and reports it --- or assumes that's just the
way it is and that the ESC key should be used instead --- compared to
put a small paragraph into the tutorial so users can be aware of
possible race conditions right away and may be able to fix them without
asking for help?

Emacs is exceptional in how much the developers care about what the
users say.  With other projects, even if there might still be user
outcries, nobody cares.