Re: Imports / inclusion of s.el into Emacs

[Stefan Monnier]

> Heh, a quarter-serious dose of consistency for a fuzzily defined "core"
> Sure that's worth it? Wouldn't you rather work on namespaces?

I expect that adding lexical scoping to Elisp was significantly easier
than adding namespaces will/would be.  And I'm not nearly as often
frustrated by the lack of namespaces as I was by the lack of lexical
scoping (or as I am by the lack of namespacing).  So my estimation of
the ratio gain/pain is not looking good (note that the "pain" factor
there refers to the pain of the designer/implementor rather than the
pain inflected on the rest of Emacs coders).

But maybe someone else has an idea for how to add namespacing much more
easily than what I expect.

> The two issues intersect!

But it could be in a destructive rather than constructive way: having
a separate "ELISP" package would reduce the impetus to impose a good
naming structure within this package, which is what this thread is
all about.

>> There can be good reasons to have exceptions, exemptions, etc... but
>> they should be understood as that.  From this point of view, I think
>> it's hard to justify an exception for `multibyte-string-p`
> Other than: it's there already: it's ingrained in code and programmers
> minds.

I see exactly 1 occurrence of `multibyte-string-p` in GNU ELPA, so
I expect that the majority of Elisp programmers have never used that
function in their code.  There are some renamings I'd like to see happen
which would definitely go against the "ingrained" habits (including
mine), but I don't think `multibyte-string-p` is affected by this aspect.

>    João "who instantly regrets mentioning the manual cause I just say
> unibyte-string there, too"

Yup: `multybyte-string-p` is not even mentioned in the manual.


        Stefan