Re: Packages quality

[Tim Cross]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> What I'm getting at is that users would benefit from extra info about
> the packages, e.g. notions of popularity and health, some lists of
> related packages including alternatives.
>

I think this would be very useful. The Arch GNU Linux distro has a
similar concept with their AUR repository. When you are looking to
install a new package, there are often multiple candidates which would
meet the user's requirements. The AUR provides information on how many
installs there are of a package and a basic user 'score' representing
user recommendation/satisfaction (as well as alternative/similar
packages in the repository).  

> And it'd be good for us to make efforts at consolidating packages
> (i.e. reach out and help package maintainers integrate their new package
> with the older one they thought was crap, as happens too often).
>

Yes, if resources are available. Might be useful if there was also some
automated package analysis to help pinpoint packages which may need
attention (for example, prior to a new Emacs release). Even something
basic, such as checking packages for use of functions/variables which
have been flagged as obsolete or compilations with warnings which exceed
some cut off count or linting warning counts, might be useful. While such 
heuristics would only
provide indicators, they might help prioritise which packages need
attention first or which ones might have more problems after a new Emacs
release etc.