[Python-Dev] Set methods for mapping views

[Steven D'Aprano]

On Tue, 27 Oct 2009 08:54:52 am Nick Coghlan wrote:
> Raymond Hettinger wrote:
> > [GvR]
> >
> >> I still wish we could go back in time and unify sets and dicts, if
> >> only to find out how that experiment would turn out.
> >
> > I'm curious about the outcome of another experiment along those
> > lines. Is anyone seeing uptake for the set methods on mapping views
> > in Py3.x?
> >
> > I haven't seen any code using it, nor any bug reports or
> > documentation requests, nor any code in the ASPN cookbook, nor
> > mention of it on the newsgroup or python-help.
> >
> > Has anyone here seen any hints about how this is faring in the
> > wild?
>
> If anyone is looking for further explanation as to why Guido's
> moratorium on core language changes is a good idea, allowing a chance
> for answers to questions like Raymond's above a chance to evolve
> naturally is what I see as the most important rationale.

I don't understand that rationale. Let's take a concrete example. The 
new `yield from` syntax was accepted but now will be delayed by the 
moratorium. How would the addition of `yield from` delay or prevent 
people using set methods on mapping views?

If a proposed feature directly competes with a feature in 3.x, then it 
might delay usage of the 3.x feature -- but if that were the case, the 
proposal would almost certainly be rejected on the basis that 3.x 
already has a feature to do that very thing.


-- 
Steven D'Aprano