4 ms·
Ah, this is a bit of a false one. (A)GPL's pretty great for things that look like applications (or alternatively, pretty much anything GNU's ever made), but for
by CHY872 4y ago
Ah, this is a bit of a false one. (A)GPL's pretty great for things that look like applications (or alternatively, pretty much anything GNU's ever made), but for anything looking like a library that is designed to be both free as in speech and free as in beer (Qt is the former but not the latter - want commercial usage? Pay us for a differently licensed edition!) it's a complete non-starter because it says that any usage of a library is a derivative work and infects. This is always going to be poisonous to any commercial use, and PyTorch and Tensorflow only gained the traction they have because they could be used commercially - they wouldn't have been successful if GPL'd.
Even LGPL is super challenging here due to requiring dynamic linking. This is fine in Java (unless you use Graal AOT compilation) and Python, impossible in Go, and at best discouraged in other compiled languages like Rust and C++.
I think there's likely a copyleft middle ground - something like the Mozilla public license seems about right, where it copylefts improvements without infecting users. I don't know why Google chose Apache 2.0 over MPL or the like - for something like Android it makes sense, for something like Guava it's not clearcut. But, when people talk copyleft, they're rarely, rarely talking MPL.
I think it's unlikely that Google opensourced Tensorflow with a view to getting some free dev hours, they're doing it to push their OSS agenda, to wield influence over how everyone else does work.
- dgacmu 4y agoThis is entirely correct. Open sourcing tensorflow was about getting the world to use Google's stuff so that that stuff became more valuable / remained valuable / was possible to hire people with experience in it, not about getting people to work on it for free. The ratio of external to internal contributions is extremely small. Google saw what happened with other of their technologies like map reduce (not open sourced, the world reinvented in Hadoop and the progress started happening there instead of in MR) and wanted to avoid becoming the bit player in a space they pioneered in. It's a reasonable strategy. (The risk is that you're giving away something that could be a competitive advantage.) (worked on tensorflow at Google for four years.)
- CHY872 4y agoK8s another example. Kind of amazing that Google effectively owns the now-standard deployment solution off of 'we don't really use K8s but want you to build your applications like us and K8s is a bit like something we do use'. At least Tensorflow's actually what Google runs internally (to best of my knowledge)!
- sombragris 4y agoAnd what about Fuchsia? The primary value I see for it is that it looks like the un-Linux, especially in the license space. Google really hates copyleft and the GPL.
- HideousKojima 4y agoI mean if Fuchsia ever starts getting used for production at some point by anyone, anywhere. Seems like more of a way to do systems research by Google than something designed for real world use.
- sombragris 4y agoIf that was the case (an academic exercise in systems research, nothing for the real world), then arguably a strong copyleft license such as GPL would have been even better.
- dtech 4y agoNice to see MPL mentioned. I always find it strange MPL isn't used more, it's such a reasonable middle ground between AGPL and MIT/BSD.