SQLite creator Richard argues external pull requests transfer up to 25 years of maintenance liability to project creators

@mitchellh I think they have official 25 years support contracts with some entities.

@mitchellh You left out the best part, he calls it "free as in free puppies" lmao

Creator of Sqlite on pull requests: "You say, oh, it's free. No. It's not free. What you're doing is asking me ... to maintain it for you, to to document it for you, to test it for you, to maintain it for you for the next 25 years. That's not free." Yep.

Wise words from a wiser man than me. I've told people for the past decade and I have recent posts on here saying the same: the merge button is the easy part. Its the decade+ (Richard says 25 years) that follows where you've accepted the transfer of maintenance thats hard.

the cost is real, but contrast that with the Linux model where I grew up... all of that cost was acknowledged by everybody. And yet Linux had 1000+ contributors constantly.

Getting code merged to Linux was *very hard* - because of that cost. But possible, the mechanisms well-defined, and that took Linux from a niche to conquer the whole world.

@mitchellh Was just reading about this in Code That Fits In Your Head. Great point.

@mitchellh When your project is big enough, you have a the privilege being able to say no to things... For a more niche project people making demands on the maintainer can just gang up and take over and destroy the project.

@mitchellh Also, this is how he starts all files in the sqlite source tree (img related):

In place of a legal notice, here is a blessing: - May you do good and not evil. - May you find forgiveness for yourself and forgive others. - May you share freely, never taking more than you give.

@mitchellh Plus you are the stedward of the project. The one with the vision. A PR is a _proposal_, you are the one making the call all things considered.

Not even "free" would be a ticket for merge.

@mitchellh This is exactly where you see who actually deploys software with users versus developing into the void (vast majority of engineers).

Maintenance, integration, etc. is often far more expensive than cranking out the code.

part of that yes, that both models can work. But I posit that SQLite did not in fact conquer the world in the way Linux did. I refer less to number of installations (I don't think we even know how many Linux hosts are there), and more to the fact that Linux powers a very diverse set of hardware - from small devices to supercomputers.

That is very, very hard to achieve without an open community. Linux did it because it had pressure from many stakeholders stretching it in new directions.

Saying "the mechanisms were well-defined" paints a pretty rosy picture. Linus Torvalds is legendary for ruthlessly flaming bad ideas to protect the codebase from maintenance drag. Both SQLite and Linux conquered their niche with ruthless quality control. SQLite locks the door, Linux guards the gate.

@mitchellh Every line of code is an asset for present but a liability for future

@glcst @mitchellh sqlite has also conquered the world, so is your point simply that different development & support models exist?

@mitchellh @migueldeicaza This is the same deal with AI writing code. AI: “I wrote code for you!” Human: “Great! Now it becomes part of the system to be maintained…”

@mitchellh For those greedy people, anything they can freely used, they always regard it as they can freely control.

@mitchellh I think the perspective of SQLite is very different from a lot of other projects, if you are building an actual tool or product then cost of external contributions was always too high. But if you are building for the community to help others grow and learn, then tradeoff is wrong

@mitchellh The expensive part is the invisible tail after merge.

A PR can look like a gift on the diff page and become a tiny support contract once users depend on it.

This is the same reason agent-generated code feels dangerous: the first patch is cheap, the next 25 years are not.

@mitchellh 25 years is the unit. the PR review is rounding error

@mitchellh i feel so bad for the maintainers of all the open source projects i've contributed too.

@mitchellh first release is 25y ago so 25y more sounds logical haha