5 ms·
To expand on your point, this CodePlex post mentions there were 350 projects with commits in the last 30 days. I'd venture to guess that number was much larger
by rorosaurus 10y ago
To expand on your point, this CodePlex post mentions there were 350 projects with commits in the last 30 days. I'd venture to guess that number was much larger for GitHub. The people have voted on which service provides the most value.
- 0x6c6f6c 10y agoMight actually look into doing a quick script and checking this out.. It'd be interesting to look at. Plus I may only be able to scrape public projects, where private repositories may account for some large amount of the real number since those are typically paid for (except for Student accounts, afaik).
- gjtorikian 10y agoNo need to script! You can use Google's BigQuery with GitHub's data. Some documentation is here: https://www.githubarchive.org/ https://www.githubarchive.org/ A `PushEvent` is fired for every open source repository that has `master` updated: ``` SELECT COUNT(*) FROM ( TABLE_DATE_RANGE([githubarchive:day.], TIMESTAMP('2017-03-01'), TIMESTAMP('2017-03-31') )) WHERE type = 'PushEvent' ``` It's over 19 million!
- zokier 10y agoDon't you need something like GROUP BY repo.id in there too?
- xfalcox 10y agoYes. By running: SELECT COUNT(DISTINCT repo.id) FROM [githubarchive:month.201701] WHERE type = 'PushEvent' I got 1896676, so almost 2M distinct repos with commits in last January.
- lucaspiller 10y agoJust a heads up that 'last January' would generally refer to January of the previous year, not the current year. http://english.stackexchange.com/a/43000 http://english.stackexchange.com/a/43000
- tohuw 10y agoThat's nonsense. I would never assume "Last January" means 2016. I've never met anyone who talks the way this arbitrary "rule" implies, either. Human language is not computer language; trying to force standardization of relative terms is rarely useful. We interpret by context.
- ukyrgf 10y ago"Last January" would definitely mean January 2016.
- jasonkester 10y agoNo, he's correct. To a native English speaker, "since January" would mean "in the last 4 months", whereas "since last January" would mean "in the last 16 months" (assuming it's April now). There's no ambiguity there. It's just how the language works.
- andyjohnson0 10y agoI'm a native English speaker (lived all my life in the UK) and I interpret 'last January' as January 2017.
- sdesol 10y agoFor those that don't follow Git hosting, it may not be obvious, but GitHub's mind share, is truly massive. If you want to compete head to head with GitHub, you'll basically, have to wait for them to REALLY screw up. If you want any sort of chance, you'll have to approach things, from a totally different angle. Make code reviews your number one priority. Be the place that people want to put their code in, because they can review code, significantly easier on your site, as opposed to GitHub's. Make management your number one priority. Be the place that can provide managers with code metrics and analytics, that GitHub can't. And so on. To take mind share away from GitHub, you'll have to treat Git hosting, as a secondary thing. You can't be a derivative of GitHub and expect to compete, and succeed. Like Snapchat, Instagram (before acquisition) and others, you can only compete with Facebook, from a completely different angle. If you can provide a novel solution, that cannot be easily duplicated by GitHub and others, getting users to store code with you, will be significantly easier. With GitLab, Gitea, Gogs, etc., Git hosting has turned into a commodity product, which means you'll need to focus on providing value based on the code stored and not the management of the code.
- uiri 10y agoMake code reviews your number one priority. Be the place that people want to put their code in, because they can review code, significantly easier on your site, as opposed to GitHub's. This isn't really a high bar. Doing more than one iteration of code review with GitHub's tooling is a complete mess. In addition to that, both PR comments and Commit comments exist which can be extremely confusing if you do make one when you meant to make the other. There is already other tooling which leverage GitHub's APIs to provide a nicer interface for code review.