CircleCI - So we raised a bunch of money
blog.circleci.com
blog.circleci.com
Oh, and the favicon should spin while tests are running.
I'd like the favicon to spin. I seem to recall it causes a memory leak in chrome though - I wonder if that's been fixed.
Also, why did you opt for building your own thing instead of running basically properly hosted Jenkis + plugins? (I know your FAQ mentions Jenkins, but it only explains why I'd want a hosted service, not not-Jenkins.)
Oh, and minor nitpick, after a particular entry your FAQ entries become mis-indented: https://dl.dropbox.com/u/38476311/Screen%20Shot%202013-02-25...
We're not worried at all. TravisCI has focussed on making a great product for open source, and we've focussed on making a great product for teams of full-time web developers. They have different needs, and both of us would need a lot more to be able to move into each others' space.
To give you a few examples, they allow you to test your code against different versions of Ruby at the same time, to show your build status on your public GitHub page, to be notified by IRC, and to send build emails to mailing lists. That stuff is great for open source.
By contrast, our features are all designed for your company's web application. We offer first-class support for Continuous Deployment. Our builds are lightning fast to reduce your feedback loop (a customer clocked us at 1m:30s vs 3m:50s). We also support automatic-parallelization to run your tests across multiple machines. We allow you to SSH into our machines to debug your tests. Our emails are personalized so you only get notifications you care about. We allow you to specify the exact patchset of Ruby (we support other langs, this is just an example) that you need.
There are probably some other differences, but basically, we've spent all our time making CI and CD amazing for teams of full-time web developers, and the caliber of our customers shows that: Stripe, Kickstarter, MongoHQ, and Intercom, to name a few who are well respected here. A product that has been honed for open source just doesn't provide the same productivity.
As for Jenkins, well its just not great. Its not a great experience, it allows you to do anything, but doesn't make it easy to do anything. The other part of this is that we wanted to focus on the platform, and Jenkins would have restricted what we could do. By limiting ourselves to a certain subset, we've been able to make the sort of product we described above - tuned to our customers needs. Would have been hard to do otherwise.
Thanks for pointing that out - I've pushed a fix.
I understand why the Travis features you mention are good for open source, I'm not sure I understand why they're not good for web apps. Notifications (email and IRC) sound great for in-house stuff too; and build statuses seem to make sense on some central web page that gets a lot of hits from your dev team (be it the Github front page, or whatever) as well.
The deployment part makes sense though, thanks!
For notifications, my point was the way of implementation. For open source projects, there tends to be a fairly low-traffic mailing list where all notifications are sent. For businesses, there tend to be way more commits, and developers care a lot less about what happens in branches they aren't working on. So we personalized our emails so only the right people get the right emails.
So its both a case of which features to prioritize, and the way the features are implemented, that make the difference for commercial web development teams :)