What's your opinion on offering bounties for features?

otacke's picture

There are quite a lot of feature requests, and often the answer is something like "Good idea. Maybe someone is going to make it happen and create a pull request." There is nothig wrong with this answer. Your resources are limited, and this is an open source project for a good reason. I wonder, if setting bounties for h5p features might be a good idea.

On platforms such as Bountysource you can describe what feature you want for a particular software product. Afterwards, people who'd also like to see this feature can post bounties, and developers can hunt for that bounty.

Of course, there are pros and cons ;-) Coders might get an extra incentive for creating h5p libraries or fixing bugs, so new people might get interested and the development of h5p might become faster and more diverse. On the other hand, posting bounties could crowd out intrinsic motivation and make h5p more like a business than a community project. People might even become reluctant to contribute if there's no money attached to it. Also, new features should be implemented to the main h5p branch or there might be uncontrolled growth of potentially incompatible forks. This would probably raise questions on when a bounty should be paid. After completion? After integration? That might take time. What, if it's not integrated at all? And, let's not forget that Joubel offers do do exactly that: implement features for payment. It might cut some of their revenue.

There are probably a lot more aspects that could be discussed. I am interested in your opinion on offering bounties for features. What's your experience? What do you think?

falcon's picture

It is an interesting question. I've seen some examples of organizations offering a fixed amount to add H5P features to existing content types or create new. They haven't asked for these to be contributed back and we haven't heard from any freelancers winning these jobs either.

From Joubel's perspective the important thing is to move H5P forward as quickly as possible with as high quality as possible. We hope to see several H5P consultancy companies being built up around H5P soon. If organizations fund third parties to create features to existing content types created by Joubel, we always hope to be involved to ensure that they're not conflicting with our plans and visions for these content types. Getting a pull request when the project is done without any involvement initially is much better than nothing, but if the new code isn't in line with our plans we might end up in a dilemma. Refusing pull requests is never good.

It would be great if we soon will see several consultancy companies share their own H5P content types and applications on H5P.org.

Maybe we should create our own system for bounties for H5P features so that we could make sure that as much as possible of the funded upgrades ended up being shared with the entire community?

otacke's picture

I think the idea of building consultancy companies around H5P is interesting. I wonder what others think about it.