There is also a nice price delta between the two. Stash is $9000 (currently) for 500 users/yr. GitHub Enterprise is $125,000 for 500 users/yr. That might win over some
Disclaimer: I work in the game industry. In the last few years I've worked on projects with Valve, EA, Blizzard, Bungie, Infinity Ward, Treyarch, id, NCSoft, etc... the list goes on and on. They all use Perforce.
Perforce is double the cost of GitHub Enterprise, at $350,000 for 500 users/yr, including their (amazing) support for when the server wont restart in the middle of the night after routine maintenance.
Regardless of how sweet their support staff is, P4 has relatively painful branching, so all of these companies who have multiple teams developing on separate branches (fairly common in companies who outsource large chunks of work), an entire engineer is dedicated to concocting epic change lists for a week at a time every time you want to merge.
GitHub Enterprise could make that issue a bit more bearable for engineers (binary files get checked in regularly, but rarely are two people branching the same binary that would need merging later) and save a nice chunk of money.
Developers would probably enjoy using a DVCS and apparently don't mind loosing an entire guy who is doing merges all day, while artists, designers, and other not-so-technical folks who also contribute to the repo would have a hard time migrating from a workflow that is ingrained from internship onward. Tech-heavy companies and those who develop solely on Mac or iOS are the rare exception.
A big problem is that Git has no concept of what Perforce calls "exclusive checkout" where only one person can claim and file and work on it. That's a deal breaker for most. Communication is hard. Make it apparent that someone has claimed a file.
Retraining users from Perforce to GitHub (however enlightened that may be) has a real cost. I would estimate thousands of dollars per employee in lost productivity. Gotta make way, way easier to start working in the git workflow, especially on Windows. All game console development happens on Windows.
Fixing the issues above could crack the entertainment industry nut (and probably many others) and likely reap some nice rewards for other groups who can't afford to switch.
> A big problem is that Git has no concept of what Perforce calls "exclusive checkout" where only one person can claim and file and work on it. That's a deal breaker for most. Communication is hard. Make it apparent that someone has claimed a file.
Is this a problem particularly in the games industry for some reason? Because in "regular" development, people always think it'll be a problem before they try it. Turns out not to be, especially with the great merge support you get from git.
I could see it being an issue for binaries though. If anything, I'd expect people to balk at having to clone the entire history of large binary files, which don't typically delta-compress very well.
Realistically, I'd be surprised if that difference was more than a rounding error for any organization with 500 active developers. $21/month/user is essentially nothing.
In most the big orgs I have been with, price was not the deciding factor, but rather a host of other considerations. That being said if two product where identical in feature's and ease of use, you can guarantee that a significant cheaper product will get further consideration, price is a factor it's just not usually the top factor in big orgs. In a previous life I was an executive at some fairly large companies and personally had 20-30 mil budgets and I can say that at least in the orgs I was with, something 100k+ would get some scrutiny along with some questions as to whether there was a equivalent cheaper product. In many org's purchases over a certain threshold have to be vetted. I know the ones I was with where well south of 100k.
It's $250/person/year. Yes, it's probably a rounding error when you're talking about folks who, with salary and benefits, are almost certainly costing the company $100k+/person/year.
For the software, yes, but then you have to host it and maintain the server it's on. Sure that's not going to be much, but it's still something to consider.
