RFC PEP candidate: q'<delim>'quoted<delim> ?

[Bengt Richter]

On Thu, 7 Mar 2002 08:09:52 -0600, "Chris Gonnerman" <chris.gonnerman at newcenturycomputers.net> wrote:

>Ugly.  You state that code like this:
>
>def somefunc():
>    astring = """this is the first line
>this is the second"""
>
>... is ugly.  Fine, but IMHO so is the proposed syntax.
>
>Recall always that the charm of Python is it's simplicity.  Even with the
>new language features added in 2.1 and 2.2, a "bear of little brain" can
>remember all the rules more or less at the same time.
>
Well, I think the charm is that one may use a very lean and clean subset of
the full set of features, but that need not imply that the full set must be limited.

Python already has plenty to boggle a "bear of little brain" in meta-classes
etc., so I don't think a feature that can safely be ignored should be criticized
on the basis of its taxing little bears' brains ;-)

>I am against pasting and patching one after another new language construct,
>syntax sweetener, etc. to Python because it detracts from that core
>simplicity
>each and every time.
>
I really agree. But I think the ability to __quote__ arbitrary raw text in source
is a __missing capability__. It's not something you can already do, that I'm just
proposing a sugar coating for. NB: I'm not saying you can't __define__ arbitrary string
content by writing suitable escape sequences etc. That's not the problem. The problem is
safely delimiting arbitrary __existing__ text __within__ a source file, without changing
it and without having to inspect it or make iffy assumptions about its content.

I think, e.g., that "for i in 5: ..." is *more* sugary, because you already have the
ability to do it with "for i in xrange(5): ..."

>This is why I am against incorporating the "$varname" interpolation
>operator.
>At least, IMHO it is not ugly.
>
But things like this you can often build using classes and subclassing. I would rather
extend the building capability (as has been done greatly with 2.2) than add pre-built gizmos.

>There seem to be a never-ending stream of new language feature proposals.
>Praise Guido they don't all get incorporated!
>
Agreed.
>(Don't take this personally.  I've had my share of similar ideas; we all
>think we are like Guido and can improve the language.  Learning to trust his
>language-design skills was hard, but rewarding.)
Well, once this has boiled down to a clear proposal, I'd like to hear his pronouncement.
Thus far, I don't seem to have succeeded in communicating the fact that Python can't
do now what it could using some implementation of this idea. It's not a new idea. It exists
in other contexts, so it's not a personal pet invention. It's a matter of whether Guido
will think the base functionality is desirable. If he does, he'll likely think of better
syntax for it. I would guess he invented triple quotes as an easy and convenient way to
deal with quoting a mix of single quotes. This just takes the next step, recognizing
that another fixed sequence (or two) of characters as delimiters will not do it.

Regards,
Bengt Richter