Fair enough. I'm a bit concerned that often evaluations of the quality/usefulness of something are subjective (e.g. the people who like flymake would argue that flycheck might be "redundant", etc). To me - it's fine to try to solve the same problem in multiple ways if there's no clearly superior way of doing something. I guess a package is never really obsolete until no one uses it and maintains it.
In general I agree that some quality criteria have to be met for a package to be included on NGE, I just hope that those criteria are going to be of the objective kind.
> I'm also curious what "redundant package" even means in this
> context. There are always many ways to achieve something and usually
> there's no clear way to decide if some approach is much better than
> the alternatives. Given how early we are with NonGNU ELPA I think that
> concerns about "obsolete" and "redundant" packages are quite overdone.
I don't think it's time to start pruning NGE, indeed, but we could avoid
adding packages that are on MELPA that are already obviously
obsolete/superseded.
For example, I use the software "anki" quite a lot, and there are four
different packages for it. I have tried all of them, and frankly
speaking only one of those packages is relevant. Therefore, I would
advise against adding the others; as Philip points out, this is more
likely to waste users time than be very helpful.