[Python-Dev] Re: Call for defense of @decorators

[Fernando Perez]

Gustavo Niemeyer wrote:

> - The @ character is used in at least two tools (Leo, IPython),
>   and this is being considered as something bad, but not a
>   show stopper;

Just to record my comments here as the ipython author.  

I've held off on calling on the ipython lists for suggestions on what to
replace @ with, waiting for the dust to settle on this matter.

If @ ends up being accepted, ipython can obviously adapt.  I'll replace it with
alternative syntax, be it @@, %, or some other kind of special-casing trick.

It would not make me happy, and it worries me that ipython explicitly uses (for
good reasons, I like to think), the three characters explicitly avoided by
python:

@ -> ipython's 'magic' control system, which is fully user-extensible.

? -> ipython's object introspection system, with ?? showing additional info.

$ -> in ipython's shell extensions, a syntax for quick shell output capture,
either into strings ($astr=ls *.py) or lists ($$alist=ls *.py).

I worry that soon I'll end up rewriting other parts of ipython as soon as ? and
$ also become python characters.

But I accept that fact, since nowhere did the docs ever _promise_ that these
characters would remain off-limits forever.  I knew there was a risk involved,
and if that happens, I'll figure out a way to solve it.  But I'd much rather
not.

Best,

Fernando.