On 3 November 2013 04:38, Leo Liu <sdl.web@gmail.com> wrote:

> On 2013-11-02 23:27 +0800, Bozhidar Batsov wrote:
> > I've attached a patch for your consideration.
>
> Personally I think such one-liners won't be of much use. The designer of
> the hash table API would have considered such use cases and decided to
> rule them out for good reasons.
>

I think this is the wrong kind of reasoning. The fact that something exists
(or doesn't exist) doesn't mean necessary that it was carefully thought
through. In the end of the day everyone makes design mistakes from time to
time. I guess the original reasoning was to provide as minimalistic API as
possible (or more likely - the API was modelled after Common Lisp, which
doesn't feature those functions as well). APIs, however, should evolve and
designers should always take account the way the API is actually used by
its clients.

Common Lisp doesn't have those functions, either, but they're included in
the Alexandria library, that most Common Lisp projects use. The Common Lisp
core is beyond extension for various unfortunate reasons, but Emacs is
not...

As I've originally said - I can't recall another programming language that
doesn't have such methods in its standard library.


>
> Maybe start out as your personal project and see how people react to it
> first? WDYT?
>

I've seen these functions inlined quite often into Emacs packages -
obviously people use them. Don't think that a research paper is needed to
warrant the demand for their inclusion. :-) Plus, as Stefan said - other
non-essential utilities might be moved to the same library.


>
> Leo
>