Anyone remember kuro5hin?
I do. It was great.
The Lobsters of Old, no?
Eh… sort of the Old Lobsters of old. I mean, the jcs Lobsters, before this current era of gentrification. Yeah, it was a splendid place. and doomed
Hard disagree, the only associations I still have are “somehow pretentious” and “not technical enough”.
It’s one of the websites that got recommended all the time but I didn’t manage to even consume content there, much less participate. Maybe just a personal mis-fit though.
Yelp is just a crowd-sourced petty extortion scheme. Do we really need any more of that?
Publish them if you want… or not. But do your own tool evaluations. It’s slower but so much more meaningful. It’s just not reasonable to expect any particular reviewer (even if honest and unbiased; big if!) to share all your objective and subjective criteria. When you realize that, then you’re left shopping around for reviewers. Aggregating them just muddies the waters. It’s a waste of effort. If you don’t trust your own critical judgement, why would you trust anyone else’s?
It’s an analogy, so it requires a little bit of imaginative steelmanning.
Since nobody’s getting paid by the download for libraries, there’s a lot less incentive to put a bunch of effort into paying people to write reviews. I think reviews for open source libraries might work a lot better than Amazon or Yelp.
In my overall experience, what the FOSS community lacks in crass commercialism it makes up for many times over in religious fervor. “Yelp for Github” sounds like the kind of elevator pitch that might get somebody six months of runway, a few years ago, if they were stuck in the right elevator. But I’d expect it to be an absolute shitshow in practice. Mere idle speculation on my part, of course.
It’s not very high quality, but occasionally I’ve found useful open source tools on https://alternativeto.net/. It’s definitely more focused on end-user apps instead of libraries or infrastructure, but it’s better than nothing.
Funny that after the title and initial comparison with review sites, the word ‘review’ never occurs again. Has nobody ever tried to create a site where people write reviews of packages they’ve used?
That’s what I’d like to know!
The modern web does a horrible job at spam prevention, SEO abuse, etc, and I would absolutely hate to see something like Yelp come to the open source world.
In the Ruby ecosystem, I’ve found The Ruby Toolbox quite useful. It predefines categories of libraries, e.g. Web Authentication, so you can easily get a list of all projects that help with a particular problem. Within each category, it exposes various metrics about the popularity and activity of each project, and it summarizes the metrics with flags like “No release in over a year”. The best project is not always the highest-rated in its category, but it’s a good starting point.
Very cool. Thinking a little more, it makes sense for this kind of thing to be separate for each eco-system. Nobody’s going to be comparing libraries across say Python or Ruby.
I especially like the metrics “issue closure rate” and “pull request acceptance rate”.
You’re probably right, but I find myself wanting to do cross-language comparisons from time to time. Knowing that “this is like requests in language X” would be high praise. Similarly, when learning a new language, I’ll eventually ask “how does this language do Quickcheck (https://hypothesis.works/articles/quickcheck-in-every-language/)?”
A major reason to pick languages for a project is knowing what works well that you will need. E.g. machine learning libraries, time handling libraries, http libraries… In this, cross language comparisons would help newcomers who don’t yet know what works best. If I were starting a project that needed to talk SOAP, I would want to know where the best libraries were.
I think various topic-specific “awesome lists” fill some of the better parts of the need described by the article fairly well.
Freshmeat used to be ok for this kind of thing, too.
The amount of effort it would take to keep spam and other forms of abuse out of anything that looks much like yelp itself would be non-trivial. I’m having a hard time imagining a model that would make that much effort worthwhile.
I generally judge open-source tools by their documentation and history.
I’ve done a lot of evaluating open source, distributed databases over the last couple of years, and something like this would have been very useful. There’s just only so much signal I can get out of running and testing something for a few weeks.