Suggestion: make sequence and map interfaces more similar

[Tim Golden]

On 01/04/2016 08:59, Antoon Pardon wrote:
> Op 31-03-16 om 16:12 schreef Mark Lawrence via Python-list:
>> On 31/03/2016 14:27, Random832 wrote:
>>> So can we discuss how a unified method to get a set of all valid
>>> subscripts (and/or subscript-value pairs) on an object would be a useful
>>> thing to have without getting bogged down in theoretical claptrap about
>>> the meaning of the mapping contract?
>>>
>>
>> We can discuss anything here until the cows come home, but it's a
>> complete waste of time if the powers that be over on python-ideas
>> and/or python-dev don't agree.  This was suggested a day or two back
>> but seems to have gone completely over people's heads.
> 
> Just because you are not interested, doesn't mean it's a complete waste of time.
> Discussions like this often enough produce suggestions on how one could handle
> these things within python without the need for the powers that be to agree on
> anything.
> 
> If you are not interested just don't contribute. Others can make up their own
> mind on whether this is a waste of their time or not.

FWIW I'm broadly with Antoon here: wider-ranging discussions can be
interesting and useful. (And informative, especially where people speak
knowledgeably about an area outside my own competence). There *are*
technical forums where anything outside their strict subject matter is
frowned upon or curtailed. I don't think we need to be that rigid here.

However I think there are a couple of lines which can be crossed. In one
case a a poster (perhaps abruptly) says: I think Python should do this;
why doesn't it? The other is where the discussion goes so far into
cloud-cuckoo land that it alienates all but a few devoted adherents to
the thread. Especially where it goes round in circles.

For the latter, I take the view that I know where the delete key is (or
the "ignore thread" button or whatever) and I just skip the thread when
it shows up. Perhaps missing some interesting points in the process if
it comes back down to earth but that's the way it goes.

For the former, I think it's fine if someone is asking in a genuine
spirit of enquiry, ie to learn about the history of a particular
decision or the ramifications of an alternative which might not be
obvious at first glance. (Why do we have both lists and tuples? Why does
Python index from 0? etc). People who know something about it can
explain if they wish.

If it becomes clear that the poster is in fact pushing for a change
(either intelligently thought-out or naively ill-considered) then I
would push them towards Python-ideas sooner than later, because that's
exactly the purpose of *that* mailing list. People on python-ideas want
to go to and fro over the relative merits of proposals. Specifically, I
believe that's the only mailing list which GvR follows apart from
python-dev. Any discussion here would likely have to be repeated over
there anyway, so why not go there earlier on?

Courtesy & respect on this and any list are important, so as long as
someone's not being genuinely rude or abusive, your best plan is to make
your point clearly and politely and then step back. I've been impressed
again and again on Python lists where people maintain a courteous front
in the course of a perhaps quite heated discussion. And more impressed
when people who have lost their cool come back and apologise (without
necessarily backing down from their point, of course).

Feel free to contact the list owner [python-list-owner at python.org] if
you think there's a real contravention of the list etiquette but I'm
personally not inclined to jump too heavily on rambling discussions or
wild-eyed ideas as such.

TJG