Re: Moving packages out of core to ELPA

[JD Smith <jdtsmith@gmail.com>]

> On Feb 17, 2024, at 8:42 PM, Po Lu <luangruo@yahoo.com> wrote:
> 
> JD Smith <jdtsmith@gmail.com> writes:
> 
>> There was a recent discussion about which lisp/progmodes packages
>> belong in core.  A sentiment was expressed that useful languages with
>> non-negligible user bases should probably go in, and others should be
>> in ELPA.
>> 
>> I want to bring up a related point: it should be possible to retire
>> packages from core, once their relevance drops below a critical
>> threshold [1].
> 
> Every instance of this is a mistake.  

That would imply that every instance of bringing a package into core is not a mistake, which suggests superhuman forecasting and decision-making.  But even when including a package is beneficial, situations change over time (>20 yrs, in this case).  It seems only prudent to reevaluate whether a given package still provides sufficient benefit given the non-zero costs associated with it.

> Why is it so essential that such
> packages be removed from core?  What practical advantage does that hold?

Many which I mentioned in my initial message.  The most salient:

- Reduces maintenance burdens, freeing time for packages that have more pressing issues. I have heard from emacs maintainers who have spent significant time trying to understand and fix bugs in IDLWAVE code that is likely unused (even by me).
- Removes "tripping hazards" for users who inadvertently activate the mode for unrelated files and are confused (this is not hypothetical: I've had numerous reports).
- Cuts down on "extra noise" in, e.g., the top level Info help.

> A Lisp file is considered part of Emacs, whether it be in core or in
> ELPA.  They are expected to meet like standards, and bugs (in the
> absence of a maintainer) are the responsibility of the same Emacs
> developers who respond to bugs that concern Emacs in general, i.e., like
> developers.  

Is this really so, in practice?  I have packages in ELPA which are effectively untouched except by me, other than on first ingestion.  And they draw updates from a repo I maintain myself.  Maybe I've misconstrued the situation, but my understanding has been that core packages receive far more attention from maintainers.  And rightly so, IMO: everyone has them installed, after all.  

Also, if ELPA and core are truly equivalent, I cannot then understand the common strategy of "let it mature on ELPA for a few years, then potentially migrate to core."

>>   It deserves support in Emacs.  Just not, IMO, in core.
> 
> Why not?  Why does _anything_, to speak nothing of a package already in
> core, not "deserve" support in core, while deserving support in ELPA?

Because in ELPA, users must proactively opt-in to the use of the package.  For such users — those who have actively sought it out — in stark contrast to the vast majority of Emacs users, the benefits dramatically outweigh the costs.