Re: [ELPA] New package: ERC

[Corwin Brust <corwin@bru.st>]

Hi Emanuel!

On Mon, Sep 20, 2021 at 7:16 PM Emanuel Berg via General discussion
about ERC <emacs-erc@gnu.org> wrote:
>
> Amin Bandali wrote:
>
> > It's about time we added ERC to GNU ELPA. :)
>
> I'm obviously out of the loop here but isn't ERC shipped
> with/in vanilla Emacs already?
>
> So why would one put it in ELPA (no disrespect)?
>

I think there are several reasons why it makes sense to place a core
package also in ELPA.  In this case, I see a bunch of benefits; I find
these three most exciting:

1. Getting new ERC versions from ELPA means easier, more rapid access
to fixes and new features.

Currently, our options for updating ERC are
a) building Emacs from the main development branch (not everyone finds
this easy),
b) merging elisp from that branch (when that works, the ERC that will
come with Emacs 28 will have need of at least Emacs 27, IIUC), or
c) waiting for new Emacs releases to be cut, publishing the
accumulated ERC changes since the last Emacs release.

My sense is that this last is most common, at present though I won't
claim I have any data to support that theory ;)

2. Releasing ERC more frequently can improve ERC's quality.

Currently, we release ERC once per Emacs version.  As ERC starts to
release more frequently than Emacs people will be able to experiment
with and provide feedback and bug reports on each.  That additional
testing should (eventually) lead to more stability from the ERC that
ships with each Emacs release.

3. Shorter development cycles will make hacking on ERC more fun.

Development is more exciting the more people who, and the more rapidly
people do see our changes.

By releasing ERC more often (vs Emacs releases), we make it easier for
ERC to be a "proving ground".  This can help increase the desire of
others to try contributing to ERC and thereby gain feedback, skill,
and confidence, and get encouraged to contribute to other ways, such
as developing contributions to other parts of Emacs.  More frequent
releases can also make ERC more useful for experimenting with new
approaches/techniques that we (aspiring) ERC hackers consider may be
useful in other parts of Emacs (if they work, people like them, they
don't turn out to be difficult to maintain, ...).

Cheers!
-- 
Corwin
corwin@bru.st