LZ4 seems really nice for data with a lot of repetition, but not having any symbolic entropy encoder can really kill the compression for some data. I often default to LZ4, and used to fall back to gzip if that doesn't work well, but I've been really impressed with zstd. It's not as commonly used, but if I'm not concerned about interoperability, I'll be trying to use a lot more zstd.
I have no opinions on the merits of this project but it got me wondering in general. Do big company project repos trivially get thousands of stars from their own employees? Does Github apply any of the strategies that places like Reddit claim to apply to fight astroturfing and such?
I can't answer your questions, but Facebook's React repo has more than 10x the stars than their Zstandard repo has. Many of their other (more obscure / less used) repos have many less stars.
To me this seems totally reasonable and proportional, given the types of developers that might use each library.
Personally, I don't believe there is much to gain by gaming Github's stars (particularly for a compression library), because there is much more to evaluating a repo and its contents than the count of its stars. Granted, that's just my opinion, and it is somewhat tangential to your original point, just my 2p.
[Disclaimer: I've used Zstandard in several projects, and will happily use it again! :) ]
Does anyone consider github stars meaningful? Especially enough to game? I ask this sincerely, I don't think I've never even noticed the star count for a repo.
It is some measure of popularity/maintenance. In choosing something to use, I would first check the activity (as a proxy to rate of bugfixes, for one), and second I will check the number of stars (as a proxy to how many people are watching/using this).
If there are more stars, I would assume that if the owner of the repo dropped off the face of the earth, it's more likely that a new maintainer would step up.
From my personal observation, those two go together and I usually don't take them in isolation. One may be able to game the stars, but they won't be able to game activity (issues, commits, etc.), I'm hoping at least not in a meaningful manner.
Furthermore, I would add the following considerations (in no particular order): documentation existence/quality, possible use of CI, reasonable coverage (if shown/known), use/inclusion in other projects, outward quality/appearance of code (can I read it, could I fix it, could I maintain it), licensing (clear/compatible) — outside of the repo: are there articles about it on the web, are there videos about it on Youtube.
It varies depending on what the repo is, and what I might intend to use it for, as to how in-depth I will actually go with each of these. But it's always worth a quick scan. There is much that can be inferred from examining such things.
Zstandard is waaaay more Yann Collet (sp?) than Facebook. Yann’s work is literally world class, and you’ll only be doing yourself a disfavor by boycotting it. I’d be shocked if he didn’t have the ability to casually stroll into any FAANG and name his price.
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...