Re: Changes in latin-ltx.el

[Dave Love]

>>>>> Kai Großjohann writes:

 > If I were to use a TeX-like notation for symbols, I'd use it for
 > writing mathematical formulas.

Maybe you would do that solely, but please allow others to be more
general.  We didn't create that input method simply for maths -- as
its historical name might imply.  Wouldn't you ever write `\ss'
because you couldn't enter `ß' directly?

 > So why not accomodate both kinds of users?  

I am the kind of user who sometimes wants to type mathematical
symbols, which is why I added them.  I already suggested using \...sym
for the compatibility character, à la SGML entities.

 > Make one input method for mathematical symbols and another one for
 > Greek/Hebrew/... text.

You're welcome to, but I want a generic method which I stand some
chance of being able to use.

 > (I wonder if there are also different variants of the same Greek
 > letter, one for mathematical formulas and one for Greek text.)

Didn't I already mention mu and micro?

 > Maybe this "if you have two alternatives, choose both" thing is
 > getting old, but maybe it's the right thing in this case.

There's no point in having alternatives if the user doesn't know which
one to choose.  This is a basic usability issue.  If you want to input
characters which require non-obvious, possibly-ambiguous input
sequences, you probably don't want a normal Quail method (which
currently doesn't even offer normal completion).  You might want to
implement a tabular input method, except that you can't currently do
so with the obvious menuing approach -- that's the main use I'd have
for Unicode-capable menus.