3 ms·

This is an issue that is dear to my heart. I think it is possible to almost always avoid killing a product. When a company doesn't want to invest further they c

This is an issue that is dear to my heart. I think it is possible to almost always avoid killing a product. When a company doesn't want to invest further they can do combinations of the following: * Just open source it. * If code has external IP, isolate them behind interfaces. * If product is service and uses internal infrastructure, you can still isolate those interfaces (assume someone else would build necessary infrastructure) * If core devs have left or aren't interested, hire offshore or freelancers to do this work * Call for maintainers/fans who are invested in the product to participate * Instill culture in the company so every new project is architected with eventual open sourcing in future due to reasons such as shutdown * Put product up for sell to potential companies who are in same space and ask them to take on work for making sure IP and dependencies are abstracted

makecheck 8y ago
My observation at companies has been that there is rarely the required foresight to be able to prune something off into a separate open-source project. Google uses a mono-repo I believe, which already makes it extremely unlikely that they’ll be able to even build a thing outside the company, much less let it thrive. And “cancellation time” is not when new resources will be diverted to fixing broken-only-outside-the-company behavior. I’ve only seen this work a couple different ways. One: if the company spins off something very large (like an entire division) that was already self-sufficient so it is pretty clear how to keep it all functional from the outside. Or, two: if, from the very beginning, the person pitching a project insists on an open-source/flexible model and takes lots of steps to ensure that it is constantly working in that way.

3 ms·

This is an issue that is dear to my heart. I think it is possible to almost always avoid killing a product. When a company doesn't want to invest further they c

by sytelus 8y ago

makecheck 8y ago
My observation at companies has been that there is rarely the required foresight to be able to prune something off into a separate open-source project. Google uses a mono-repo I believe, which already makes it extremely unlikely that they’ll be able to even build a thing outside the company, much less let it thrive. And “cancellation time” is not when new resources will be diverted to fixing broken-only-outside-the-company behavior. I’ve only seen this work a couple different ways. One: if the company spins off something very large (like an entire division) that was already self-sufficient so it is pretty clear how to keep it all functional from the outside. Or, two: if, from the very beginning, the person pitching a project insists on an open-source/flexible model and takes lots of steps to ensure that it is constantly working in that way.