And then you quite quickly get to a point when you need to curate an "awesome list of awesome lists" like https://github.com/sindresorhus/awesome or "awesome list of awesome lists of awesome lists" like https://github.com/t3chnoboy/awesome-awesome-awesome ...

> This list is incomplete; you can help by expanding it.

In theory once you get a "awesome list of awesome lists" you can just add subsections to it.

Just so long as there is never a requirement for a subsection of "awesome lists that do not contain themselves".

I agree that it's needed.

The problem with those kind of sites is that the site owners try to game the ranking system.

These lists get far more Github stars than actual software...same problem here.

You keep the curators to 5-6 experts in each field, with a reputation for being straight.

If they do any tempering, expel them.

This requires a rich person to pay them, or a community to motivate them, which gives the super-power to an agreeable, but fair BDFL.

This already exists. The problem is how to find it, since on the surface it looks the same as all the junk lists

Well, people trying it, getting good results, and thus creating word of mouth for it, should be enough.

Well, the list of tools is at least as useful as the most useful tool on the list, and probably reflects the overall usefulness of its category. Which means that if we go enough levels farther, the master list of lists [of lists...] would reflect the overall usefulness of all software on GitHub.

Also maybe GitHub needs a "list of stuff" category, so these lists of things don't muddle search results and conversely can be searched for easily. Curated lists of links are generally quite useful and putting it all in a public vcs with a pull request infrastructure makes it easy to contribute.

I definitely see a usefullness in these lists, I just feel sad that they get a lot more attention than actual code repositiories.

Indeed, Github should find a solution for this. I didn't care much about Github stars but recently I noticed they're being required in various scenarios, for example some CDN require a certain number of Github stars to include a library [1]

[1] https://github.com/cdnjs/cdnjs/blob/master/CONTRIBUTING.md#b...

I don't think Github stars should be considered as anything more than built-in bookmarking. Personally, I only use them that way. With this view, it's obvious why lists get more stars than the projects listed.

> the list of tools is at least as useful as the most useful tool on the list.

Simple disproof: a list of everything is not more useful than the most useful thing, because navigating the list has a cost.

Also, an analogy: would you eat a bowl of candy that had the best piece of candy in the world and also the most poisonous substance (malware)?

I agree. It seems no algorithm could kill "curators" yet.

Well, Github could select an official list per technology, move it to eg. the github.com/explore part and reward the curator with money or software licenses/subscriptions.

What if I'd like something that the arbitrary community manager of GitHub doesn't? Why would a software project management site be the arbiter of good software?

dmoz was that site for a long time after yahoo became it's 2000s incarnation, but then dmoz went away, there are a couple of live archives of it, but still...

It appears that dmoz was resurrected as https://www.curlie.org/

But these lists also already fulfill the need. When there's a category of something people are interested in someone will eventually make an "awesome list of ...". The good lists become popular and become included in various "awesome list of lists" improving their discoverability, while the bad lists fade into obscurity. And since there is no central authority and plenty of alternatives each list can be opinionated instead of including everything under the sun.

These lists follow the same "survival of the fittest" that underlies capitalism and evolution, and while coordinated efforts have many advantages they also have all the disadvantages of monoliths. I think a Yahoo-like site would have a hard time producing a better result.

Free markets mostly work because people vote with their dollars. Democracies mostly work because people vote with their votes. Web democracies fail because votes are trivial to manipulate and gamed by SEO.

But web dictatorships work great because switching costs are low, so if a dictator doesn't deliver people just go somewhere else. While centralized link lists are often democratic, awesome lists form a network of dictatorships.

I made a site that attempts to be a quick, hierarchical overview of the software tool landscape and the bevy of alternatives: https://www.tooltldr.com