You don’t have to love Facebook to love the code it open sources

You don’t have to love Facebook to love the code it open sources

Commentary: Can open source creators be separated from their code? Yes, and they should.

Facebook : Illustration

Image: Chesnot/Getty Images

In recent years there’s been a shift in open source to trying to attach ethics to licenses. Open source developers have tried to use licensing to block use of their software by actors (or, rather, actions) they don’t like. So there’s the Hippocratic License, the anti-996 license and more. I’ve argued against these licenses as essentially unenforceable, creating far more problems than they purport to solve. 

In light of the exceptional open source software that Facebook creates, I’d argue there’s one more reason to eschew such licensing: Great code needn’t correlate with great code authors. Or, otherwise put, you can despise Facebook while loving the code it releases.

SEE: A guide to The Open Source Index and GitHub projects checklist (TechRepublic Premium)

What’s not to love?

More about Open Source

There’s a multiple choice of reasons to dislike Facebook. Maybe it’s the way Facebook handles user privacy. (There’s no shortage of permutations of this one.). Maybe it’s the negative effects Facebook has had on elections and national discourse. Or, in my case, maybe you just don’t like how it tweaks algorithms to ensure you see content that will make you more likely to click ads, not necessarily to become a better, more informed person. 

I really, really dislike Facebook. I got off the platform a few years ago and have never once felt compelled to get back on. Sometimes I have missed the fakery of Instagram (a Facebook app), but a quick glance reminds me that, no, I’m not missing anything by watching the highlight reels of others’ lives.

And yet…Facebook does profound good with the code it releases. Am I allowed to give its open source code a thumbs up?

Creator and created

I hope so because Facebook has created a treasure trove of open source software, and continues to add to the pile. For example, Facebook just released a video dataset (“Casual Conversations”) to help uncover artificial intelligence bias. Yes, you could cynically say that Facebook is hardly the right party to talk about algorithmic bias, but this software is separate from the company. 

In fact, Facebook has been the source of exceptionally powerful open source for a long time. Need a great database? Facebook launched Apache Cassandra in 2008. Building a mobile app? Facebook released React in 2013 (and relicensed it to be more permissive in 2017, in the wake of criticism). Want a great project for deep learning? Facebook created PyTorch in 2017. And then there’s GraphQL, Presto and much more. 

But should you use that code, if you feel about Facebook as I do? Can you trust Facebook?

Here’s the thing: With open source software, you don’t have to trust the creator of the software. Because of the licensing, the code is somewhat self-contained. Worried about what might be inside? Look at the code. Worried about the future of the project? Fork it. In open source, you never need be hostage to the source of the software. 

This brings me back to the so-called ethical licenses. They try to make the creator of the software persist as a presence for the downstream developer of the software, and that strikes me as wrong-headed. One of the virtues of open source is that, in open source, nobody knows that you’re a dog, as it were. All that really matters is the code and the license that frees it from the originator thereof. This means I can use Facebook (or insert any company or individual with whom you may disagree) code without concern, and so can you.

Disclosure: I work for AWS, but the views expressed herein are mine.

Also see

Source of Article