WHO CARES. Hit this author over the head with Bitbucket or whatever and he's not going to disagree with you; in all likelihood he'll just say "anything that gets your code out there so that people can interact with it and build on it is going to build currency for your project".
We really, really don't have to get in a big Git vs. Mercurial fight over this. They're all fine. It does not matter. In fact, Patrick will tell you not to use any of them and host your own code, if you care about SEO.
> In fact, Patrick will tell you not to use any of them and host your own code, if you care about SEO.
Won't it be a non-issue if your repository isn't public which would be the case for most of the startups. Except for a couple of non-critical utilities, companies generally don't open source their code on their way up, and when they have reached the critical mass and are open sourcing big projects, they generally would have sufficient SEO weight to outweigh the repo effect.
I put up a single page with some very simple examples of doing computer vision in Python on my website three years ago. It's gotten enough links to give my site some authority in Google's eyes. I don't think it would be very difficult for a software company to find some code that others would find useful and post it on their website. And in fact, it would be a lot easier than generating other content, considering that software companies write software all day.
Admittedly, I'm not a source control or git expert by any means. However, in a lot of work environments, it is a contest. Your company decides at the outset that they will commit to use a set of technologies for the next project and that's that.
I am personally looking at it from the outskirts as a guy trying to build a startup. When I look around for a library or plug-in I need, I end up on Github. When I look at cloud hosting providers like Heroku and AppHarbor (even though they've since added support for other source control tools), I see support for git.
I'm a machine learning guy, so my goal is to find the tools that minimize my time developing software and maximize my time discovering new AI algorithms. From that perspective, it is a competition: I do not have time nor care to tinker with all the different source control methods. I just want what works, and what works is git.
While I agree with this statement very much, and I know various different source control programs, I still from where I am standing feel like Git is something that new startups and people in generally work with.
It is one of the first features touted when a newer service comes to town (sometimes alongside Mercurial, but generally support for Mercurial comes after git).
I personally really enjoy git, I love what it has to offer, and honestly it has a better community aspect to it, Github is simply amazing, with Gitorious a close second and yes BitBucket exists, but it is not even close. It doesn't have the same community feel, pull requests are practically non-existent, the online tools are not that great (the wiki being a mercurial repo is actually pretty nice), and overall it just seems to have stagnated since being bought by Atlassian.
As for the corporate world, if you want in house solutions either has some choices. You can go with straight Git/Hg without a web front-end, you can use their built-in web front-ends or you can get something like Kiln and Gitorious (being open source is kinda nice really, low barrier to entry).
Where Git still lacks is on the Windows side of things, someone really should come along and write a Git compliant client that is more windows centric and runs correctly rather than having to have MingW installed and a whole range of other tools and utilities. The version also lacks behind somewhat so sometimes you may run into a bug that you can't replicate using a newer version (had a nasty merge that caused that issue, worked fine on my Mac, no so much on Windows).
Mercurial since it is entirely written in Python will work great on Windows. The TortoiseHg and TortoiseGit will always leave something to be desired but at least they exist and can make it somewhat easier, other than the fact that Git's terminology doesn't match up with Subversion or CVS (what most devs have used before).
Exactly! Just use what you like the best. There is no contest, there is no winner.
Saying git is the "clear winner" over hg or svn or what have you is like saying that ruby is a "clear winner" over python or c++ or whatever. It makes no sense.
It is? How so, exactly? It has a small lead over other systems as far as number of users goes, but i'd hardly call it a "clear winner".