OpenHeatMap is an abject failure commercially, but I’ve kept it running because around 40,000 people a year use it to create simple visualizations. It’s free and aimed at non-technical folks, so a lot of them are from non-profits, schools, local activist groups, and other causes that I enjoy helping.
I spend a few hours a week answering emails, and occasionally have to do some server maintenance, but it’s generally a fairly light labor-of-love. The source code has always been up on Github, but a few months ago it was taken down after a copyright complaint. Unfortunately Github don’t seem to have any process for me to contest this decision, and after a few inconclusive exchanges with their support team supplying further information they’ve stopped replying to my messages.
It’s not really hurting me personally, I’m still able to keep up my maintenance on the project from a local copy, but I’ve had to field puzzled emails from people who want to fork or learn from the project, including one team with an intriguing idea for a suite of visualization tools aimed at non-profits. I don’t like disappointing these folks, so I’m putting this together as an explanation of why I’ve not been able to help them.
The actual complaint isn’t completely nuts, but what has left me sad is how Github handled it. About five years ago somebody emailed me a bug report, attaching a few spreadsheets of addresses and names, without any description of what they were. This is a pretty typical use case for the website, I’ll often see what look like short fragments of a phone book as test files. I fixed the bug and added the file to my unit tests. A few months ago, years after I’d originally created the unit test, I received an email from a CTO at a consulting firm, angry that what appeared to be a list of his staff was available if you dug around enough on Github, though it was unlabeled. It appeared one of his employees had sent it to me as a test case. I felt a bit embarrassed – looking back I’d be more careful about what I used as my unit test inputs, at least scrubbing them more vigorously so that only address data remains (though a lot of the tests are about identifying what is address data in a soup of other columns). I went ahead and removed the offending data from my repository and its history, and checked the result in. I also started the process of removing the unit test directory entirely from the public project, though that was a longer task. Unfortunately the complainant found another copy of the file he hadn’t spotted before, and rather than contacting me, got in touch with Github and persuaded them to disable the project.
I was pretty disappointed, but assumed I could get the project back online with the unit tests removed entirely. Unfortunately they don’t seem to have any kind of process for resolving problems like these, and the only point of contact I’ve had is through their main firstname.lastname@example.org email. I tried reaching out on Twitter too, but without any luck. Right now the project’s stuck in limbo, apparently permanently banned, and I’m not sure how to get it online again. I’m a long-time fan of Github, and there’s no other provider that offers such a good environment for sharing source code, so I’m just sad that they don’t seem set up to handle this sort of problem, and I hope this doesn’t affect other projects going forward.
[Update – I’ve had an email back from Github, it sounds like my previous mails may have gone astray, and they’re on the case.]
[Update 2 – OpenHeatMap is now back online, thanks to everyone for their support!]