Re: C-x C-x (was: C-d deleting region considered harmful)

[Andrew W. Nosenko]

On Tue, Sep 21, 2010 at 01:09, Stefan Monnier <address@hidden> wrote:
> So, I think we agree that we should change the active-region deletion so
> it only happens for `delete' and not for C-d.
> Good.  We're just waiting for someone to cobble up
> a corresponding patch.

Very please to excuse me, but may be it need to stop decoupling
aliases into surprisingly different functionality?

I always used to C-d and "del" key as aliases under Emacs.  And want
to continue the same.  Just because keyboards are different.  Full
sized pc keyboard is one thing with the one distances between keys and
good separated navigation block.  Netbook keyboard is absolutely
another thing with absolutely different key sizes and different
physical keys placement.  I was (and is) very exited by these [as I
think] intentionally provided aliases that greatly improve usability.
And please, don't ruin it!

And for clearness: it's not about whether delete-char should to delete
active region (if any).  Anyway I can to turn such feature on or off
through customize or init.el.  It's about decoupling/breaking of very
useful aliases, which allows to feel comfortable under variety of
highly different keyboards.

>
> Now for the spinoff thread that keeps spinning off other ones: I'd be
> willing to decouple C-x C-x from "(re)activate the region", but then we
> need to find some other way to reactivate the region.
> Suggestions welcome.  Of course there's C-u C-x C-x, but I don't find it
> very appealing (just like C-u C-x C-x doesn't sound too appealing to
> people who currently want to exchange-mark-and-point without activating
> the region).

About stop activating region by C-x C-x.  I'm (of course) unable to
speak for anyone who uses t-m-m, but for me C-x C-x, same as C-x C-x
C-x C-x is intentionally thing for activating region.  And usual goal
is to send this region to the external filter (as in C-x C-x C-u M-|).
 Just depending on goals I want to start review or editing of results
from top or from bottom and therefore use 2 or 4 C-x.  And it is
simple and logical: exchange point and mark and activate region, just
repeated twice if need (thanks to "activate" instead of "toggle"
semantics).  Now it may be transformed to exchange and some when
latter activate.  What if I hit C-x yet another 2 times?  Continue to
activate region?  Start to toggle?  Start to toggle every 2nd (even)
pair of C-x and do nothing on every 1st (odd) pair of C-x?  Something
another?

Again, please excuse me, I have no intention to insult anyone
personally or as group.  I just argue that here are many things.  And
logical consistence is not the least of them.  Just because decreases
learning curve (in short term) and keeps productivity in the long
term.  And it is just from user's point of view, without counting the
Emacs developers convenience and time (any inconsistency produces
exceptional cases, any exceptional case produces special control flow
branch, any (especially irrational) branch is the source for errors or
at least brain resources eater).

-- 
Andrew W. Nosenko <address@hidden>