Suggestion: Broaden scope to all packages

[overlookmotel]

The scope of this initiative thus far is "key packages" in the ecosystem.

I would argue that the huge diversity of packages available for node is in itself key to the ecosystem. Therefore, I feel it'd be ideal if this initiative could widen its scope to trying to put in place systems/assistance which could be applicable to all packages.

Obviously, it'd be an impossible task to make active interventions on thousands of packages, but what I'm trying to suggest is that it'd be useful to bear in mind a wider scope when discussing how to tackle this problem.

An example:

react-loadable is a module which gets 250k weekly downloads, and has been unmaintained for some time including unfixed bugs. 250k evidences a large user base, and judging by the number of open PRs, many motivated would-be contributors - enough I expect to keep the package maintained if they were able to.

But I don't imagine that it's big enough to be considered "key" in the way an express or mocha would.

Only a fraction of node's user-base will use react-loadable. But I would hazard a guess that pretty much every node user depends on some other mid-level package like it. So the experience of using node overall depends not just "key" packages, but a large range of packages.

Perhaps this is already obvious, but I thought I'd make it explicit.

I have a couple of specific suggestions which I'll raise as separate issues as soon as I get a chance, but just wanted to drop this in now in case it's a useful perspective.

Also, just to say, I think this is a great initiative. The burden of maintenance, and the mess of duplicated, abandoned, and half-done packages littering npm that it causes, is the biggest pain point with node. I really hope the community can find some solutions.

[Qard]

I think, at least initially, the group needs to stay focused on a limited subset of things.

For maximum success, however, I think this group should gradually move toward efforts to not maintain things themselves but rather to provide the tools for module authors to pass on maintainership of modules they no longer have the time/energy for, with minimal effort and worry about if it will be in good hands. Providing the tooling and the community enables a much wider audience of contributors, as evidenced by the package counts on npm itself--they fixed the module creation problem by throwing tooling and community at it, but we haven't yet fixed the module maintence problem.

[mhdawson]

I'll echo @Qard's comment that we do need to maintain some focus to start. Having said that I do think it is good to keep in mind as we work through our initial steps how we might broaden the impact or re-use what we do more broadly later on.

[boneskull]

Maybe we work directly with maintainers of "key" packages, but we need to also create tools, guides, and recommendations which any package maintainer could apply (and adapt if needed).

That said, if there's something we propose which just doesn't make sense for "non-key" packages (a "known bug?") we must be explicit about who our intended audience is for that particular initiative.

[mhdawson]

but we need to also create tools, guides, and recommendations which any package maintainer could apply (and adapt if needed).

Absolutely.

[jonchurch]

I believe through tools like https://github.com/pkgjs/wiby https://github.com/pkgjs/detect-node-support https://github.com/pkgjs/support and https://github.com/pkgjs/create the scope has widened to include supporting not just maintainers of key packages but thinking about all maintainers.

I've labeled this stale?, and am inviting the group to revisit.