zlib license

[pchaigno]

This pull request adds the Zlib license to choosealicense.com. It was first used by the zlib project (see the end of its README). It's now used by many other projects (e.g., certain syntax definitions for programming languages).

Requirements:

1. Zlib is the SPDX identifier.
2. Zlib is part of the list of OSI approved licenses and GNU's list of free licenses.
3. This GitHub search result displays ~2,830 Zlib licenses. I downloaded 2,534 of them, from which I counted 1,580 repositories (from 1,439 different users).

[mlinksva]

This looks good to me, thanks @pchaigno!

I'll merge tomorrow if nobody else points out any problems in the meantime.

CI is failing due to an actual issue with an external site. I informed them about it at creativecommons/creativecommons.org#323

[mlinksva]

@pchaigno I noticed one thing -- the using example is not ideal. Apologies, I just realized the docs needed updating, done with #439. using should be:

A list of up to 3 notable projects using the license with straightforward LICENSE files which serve as examples newcomers can follow and that can be detected by licensee in the form of project_name: license_file_url

The eponymous example for zlib isn't one that new projects should follow nor one that could be automatically detected. Are there other zlib-licensed projects that would serve as more straightforward examples? See similar discussion re ISC in #377.

[pchaigno]

Unfortunately it's mostly Zlib + small projects. I went through Zlib-licensed projects on GitHub but couldn't find any well known (at least to me :/). Should I remove the using entry then?

[mlinksva]

Yes, just remove using for now. We can add projects if/when good examples found. Thanks.

[waldyrious]

@pchaigno but don't give up just yet :) I'm working on a way to query the recently released Github data on Google's BigQuery, to find the most popular ISC-licensed repos, and once I get reliable results I'll share the process, as it will probably be useful to others like you.

[pchaigno]

Yes, just remove using for now.

Done.

I'm working on a way to query the recently released Github data on Google's BigQuery, to find the most popular ISC-licensed repos

@waldyrious Nice! I have to get into that myself. Could be useful for quantitative searches at Linguist as well.

[mlinksva]

Now last row in http://choosealicense.com/appendix/ and at http://choosealicense.com/licenses/zlib/

[waldyrious]

@pchaigno I've managed to nail down the query as I needed: here's how to get all ISC-licensed repos on github, sorted by number of stars in a given period of time:

SELECT gh_licenses.repo_name, gh_since2015.num_stars
FROM [bigquery-public-data:github_repos.licenses] AS gh_licenses
JOIN (
    SELECT repo.name, COUNT(*) AS num_stars
    FROM TABLE_DATE_RANGE([githubarchive:day.], TIMESTAMP('2015-01-01'), CURRENT_TIMESTAMP())
    WHERE type = "WatchEvent"
    GROUP BY repo.name
) AS gh_since2015
ON (gh_licenses.repo_name == gh_since2015.repo.name)
WHERE license="isc"
ORDER BY num_stars DESC

Here's the result (repos with less than 10 stars in that period hidden)

Searching using the year tables instead of the day tables (2011 to 2015, so this excludes 2016 activity) could capture additional repos that haven't had recent activity:

SELECT gh_licenses.repo_name, gh_year_tables.num_stars
FROM [bigquery-public-data:github_repos.licenses] AS gh_licenses
JOIN (
  SELECT repo_name, COUNT(*) AS num_stars
  FROM TABLE_QUERY([githubarchive:year], 'REGEXP_MATCH(table_id, r"^201\d")')
  WHERE type = "WatchEvent"
  GROUP BY repo_name
) AS gh_year_tables
ON (gh_licenses.repo_name == gh_year_tables.repo_name)
WHERE license="isc"
ORDER BY num_stars DESC

...but neither query works for repos that haven't been starred by anyone. Ideally, there should be a repos query akin to the sample_repos in the recently released data. Then we'd be able to do something like the below:

SELECT gh_licenses.repo_name, gh_repos.watch_count
FROM [bigquery-public-data:github_repos.licenses] AS gh_licenses
JOIN (
  SELECT repo_name, watch_count
  FROM [bigquery-public-data:github_repos.sample_repos]
) AS gh_repos
ON (gh_licenses.repo_name == gh_repos.repo_name)
WHERE license="isc"
ORDER BY watch_count DESC

...which currently returns only 1516 rows (as opposed to the 5432 rows and 4126 rows from the above queries, respectively).

@fhoffa any chance we could get such a dataset (a repos table) anytime soon?