GitHub Hits One Million Hosted Projects | Hacker News Reader

GitHub Hits One Million Hosted Projects

techcrunch.com

105 pointsbkudria19 comments

I'm extremely happy at GitHub's success. It's one of my favorite sites. But it seems a bit strange to count gists as projects. I use them like I used to use pastebin.

I completely agree Github is excellent but this is blown a bit out of proportion.

The Linux kernel and Git projects on Github linked in the article are both just mirror projects for projects hosted elsewhere.

What difference does it make, it's not like 1 Million is anything other than a number anyway. The point is that Github has grown a lot and 1M is just a convenient time to share some love.

I know the Github guys pretty well, and I'm glad their project is doing well.

I guess I'm pointing out it's bad journalism.

The numbers are slightly misleading. 40% of those projects are gists (pastebin like service). While those gists COULD turn into full projects, it's doubtful. Thus the number is more like 600,000 hosted projects.

To be fair, we've only advertised "repositories" on our front page — not "projects" — and gists are all full-fledged git repositories. TechCrunch added the "projects".

"Project" is a pretty ambiguous term, of course; there's gists that are multi-file full-fledged database adapters, there's gists with one-line bug fixes. Both could be more influential than full-fledged "normal" projects... it depends on the circumstances. I think each have their place in programmer society.

A metric like 'active repositories' is probably more meaningful and doesn't involve the woolly distinction between 'projects' and 'gists'. Say, repositories with a checkin (other than repo creation) in the last N days.

With the ease of forking even plain 'repository' is a relatively vague term.

I chose to use "projects" over "repositories" in the headline to make it more understandable to a non-coder. There are some gists are more worthy of being called a full project than many repositories anyway. Hopefully this doesn't take away from the main point which is that github is kicking ass.

It's well deserved. Github made project hosting a pleasant experience. For opensource projects we had Sourceforge and Berlios before, but their approval processes were tedious and seemed random at times. With Github you are set up within a few seconds to share code. Git made forking/cloning easy, Github made forking trivial. Their Jekyll-based Github pages is also great.

Anyone know the ratio of public/private repos?

I wonder if forks are considered projects too...

They are, and they absolutely should be. They're completely separate projects/repositories from those from which they came. The owner of a forked repository is completely free to make whatever changes they please -- from a few bug fixes to taking projects in completely new directions (as I've done with a forked repository of my own). As far as GitHub is concerned, in the context of a forked repository being a first-class citizen, it's a minor detail that the code was originally "copied" (to use the term loosely) from somewhere else.

In comparison, how many hosted repositories do BitBucket have?

Tried Google with no luck.

http://bitbucket.org/repo/all/popular/2008/ is the last page of repos. At 15 items per page, that would give ~30120 repositories.

But you have to take into account that on BitBucket you get one private repository for free, so I guess there's loads more of them than on GitHub. Also, I'm not sure if all repositories are on that list or not - they don't seem to filter it, since there are some repos 0bytes big at the end.

Just found a way to do a similar thing on github: http://github.com/repositories/recent?page=8433 is the last page of repositories with 30 items per page - that gives ~252990 repositories. Not even close to the number cited in the article... I'd really like to know where does the difference come from.

That only gives you public repos. For one, they said that 40% of that 1M is gists (which would make it ~400k, which is more than the number of public repos. Which I guess makes sense since there are probably people that create a new gist almost every single day.) Also, there are lots of private repos that won't show up on the repo-browsing page.

It's true that there are private repos, but with ~600k project repos and 252k public... I find it hard to believe that the rest are private.

If they had ~350k private repos, they'd get between 385k$ and 490k$ per month - I assume that would be big news mentioned somewhere already...

Why is that hard to believe? We haven't kept it a secret that GitHub makes millions of dollars a year.

I think there are lots and lots of companies can't run smoothly without GitHub anymore.. Mine is one of those...

Reply on news.ycombinator.com