arewewebyet icon indicating copy to clipboard operation
arewewebyet copied to clipboard

[RFC] Criteria for crates to be listed

Open orium opened this issue 6 years ago • 9 comments

According to the contribution guidelines old, unmaintained, crates can be added. Also, crates that are not established can be added (e.g. a crate with 500 download is probably no "web yet"). This also implies that they should not be removed.

I believe there is no value in having unmaintained crates, as well as crates that have very minimal use. I suggest changing the CONTRIBUITING.md to reflect that.

orium avatar Sep 11 '18 18:09 orium

We started to discuss this today with @adriantombu and it feels reasonable to introduce minimum criteria for crates to be listed on www.arewewebyet.org

Something minimal that I think would be a good baseline:

  • The crate's repository is not archived
  • The crate is not flagged as unmaintained by cargo audit
  • The crate has at least X recent downloads

Establishing what a reasonable value for X is could be controversial, but we can start with something very low (50? 100?) as a filter for crates that have just been published and are not quite ready to be recommended to people as being "web yet".

I wouldn't use "last commit" as a filter since some crates can be considered "done" and therefore receive no commit activity even though they are widely used.

Wdyt? @ibraheemdev @Turbo87 @marcoow

LukeMathWalker avatar Mar 21 '24 13:03 LukeMathWalker

I'd recommend a slightly higher limit due to https://blog.rust-lang.org/2024/03/11/crates-io-download-changes.html, but other than that this sounds reasonable 👍

Turbo87 avatar Mar 21 '24 19:03 Turbo87

another thought: it probably makes sense to limit the number of entries per category to some degree. e.g. if we already have 15 entries for category X then adding a 16th one would need a stronger argument than adding a second one to category Y which only had a single entry so far.

Turbo87 avatar Mar 21 '24 19:03 Turbo87

another thought: it probably makes sense to limit the number of entries per category to some degree. e.g. if we already have 15 entries for category X then adding a 16th one would need a stronger argument than adding a second one to category Y which only had a single entry so far.

I agree, although I would set a high limit (e.g. 10?).

LukeMathWalker avatar Mar 30 '24 08:03 LukeMathWalker

Hey folks, just chiming in to see if there is anything more to discuss here to achieve an agreement so that we can start cleaning the issues 💪🏼

adriantombu avatar Apr 11 '24 14:04 adriantombu

I think we're good to go if we figure out a "minimum number of recent downloads" number. Based on the recent changes to crates.io, I think 2000 is a good initial threshold. We can adjust if needed.

LukeMathWalker avatar Apr 12 '24 10:04 LukeMathWalker

For reference, I'm around 2k-4k downloads for my published crates that no one uses so I bet we can go even higher 😄

adriantombu avatar Apr 18 '24 07:04 adriantombu

We can raise it to 4k, but I wouldn't go any further as "the first line".

LukeMathWalker avatar Apr 27 '24 13:04 LukeMathWalker

To summerize the discussion so far, here are the main criterias:

  • The crate's repository is not archived
  • The crate is not flagged as unmaintained by cargo audit
  • The crate has at least 4k recent downloads
  • Limit the number of entries per category to 10

Does it look reasonable for everyone?

I'll have several days off this week and the coming one, so if we can agree on those criterias I can have a first pass on the issues.

adriantombu avatar Apr 29 '24 13:04 adriantombu