RE: angle-bracket notation for keys (was: Selection changes)

["Drew Adams" <drew.adams@oracle.com>]

> > > '<...-insertchar>' is very ugly and... unattractive!
> 
> <FOO> is what makeinfo produces from @key{FOO}.  The purpose of @key
> is to distinguish between a single keypress and typing the keys F, O,
> and O in that order.  In the printed manual, @key produces a picture
> of a key with a label.
> 
> If you want to suggest a change in what @key produces in the on-line
> Info manual, this isn't the right forum.  You need to write to
> bug-texinfo@gnu.org.

I believe this was all discussed in those old threads, and I really do not wish
to reopen the discussion - that wasn't my purpose, as I said.  If others want to
pursue it, fine - I've already summarized what I think about it.  If there's
another tour around the block I'll sit this one out.

To be clear, however, I'm not really concerned about what appears in the printed
manual.  My concern (really, my _only_ concern wrt this question) is the UI and
online doc (Info).  I made that clear before, as well.

And I have nothing to say about the tools we use to create Info or the printed
manual.

My point was only about how Emacs talks about itself when you use it: help,
messages, Info.  The format we use is unnecessarily ugly and noisy.  A better
notation is possible (more than one, no doubt).  That's all.

Whether moving to such a better notation would be difficult in terms of tools
used I cannot say.  I'm not interested in pursuing such tool changes, myself.
If others want to do that, and if bug-texinfo@gnu.org is the place to do it, be
my guest.

To repeat, however, it is not only, or even primarily, about the manual, even
the online one (Info).  The UI is not created by makeinfo, AFAIK.  We've decided
to use angle brackets for named keys throughout the UI.  The Lisp functions that
we've written intentionally return the angle-bracket syntax: `<mouse-2>' vs
`mouse-2'.  That's not the fault of makeinfo.  See the old threads for info
about some of those functions.

The result gets an `A' for consistency, but an `<<F>>' for readability, IMO.
FWIW, I use Emacs 20 quite often (testing libraries etc.), and even though its
use of key notation is inconsistent (`<F1>' and `TAB') it is far better than the
consistent interface of Emacs 22+ - easier on the eyes and brain (parsing).
Fewer unnecessary brackets gives you a welcome break.

You know the tired joke about LISP meaning Lots of Irritating, Superfluous
Parentheses.  Well, they aren't superfluous, but our angle brackets surely are.