search

phoboslab / qoi

The “Quite OK Image Format” for fast, lossless image compression
MIT License
6.92k stars 331 forks source link

Boost license? #135

Closed vinniefalco closed 2 years ago

vinniefalco commented 2 years ago

The MIT License requires attribution in binaries, can this be re-published under the Boost Software License? https://www.boost.org/LICENSE_1_0.txt

Lokathor commented 2 years ago

Alternately, the Zlib license, which is used by Zlib itself and also SDL2, among other things

phoboslab commented 2 years ago

What's wrong with requiring attribution in binaries?

vinniefalco commented 2 years ago

What's wrong with requiring attribution in binaries?

Good explanation here https://pdimov.github.io/blog/2020/09/06/why-use-the-boost-license/

Calinou commented 2 years ago

On the other hand, you may consider that the required attribution in binaries is a good thing from an end user perspective. I think it's important for software to keep track of its dependencies in a public way. This way, anyone can know if a specific piece of software is affected by a security vulnerability – not just its developers who may no longer be maintaining the software in question.

vinniefalco commented 2 years ago

I think it's important for software to keep track of its dependencies in a public way

Nothing wrong with that, but it could be opt-in instead of mandated. The problem with the mandate, is that the dependent library will never be written in the first place (because the MIT license is not permissive).

Lokathor commented 2 years ago

Wait wait... what? The MIT license is basically one of the two default permissive licenses, the other being the Apache-2.0 license.

How do you figure that MIT isn't permissive enough? You can do anything you want as long as you stick one extra note onto what you make, that's as permissive as it generally gets.

vinniefalco commented 2 years ago

that's as permissive as it generally gets.

Clearly not since the Boost Software License (and other licenses) is more permissive.

Lokathor commented 2 years ago

Yes, and I wish that more libraries used Zlib.

However, my point is about how common different licenses are, and MIT is certainly far far more common than Boost.

shuckster commented 2 years ago

Good explanation here https://pdimov.github.io/blog/2020/09/06/why-use-the-boost-license/

If I were cynical, the content of this article might appear to be a veiled complaint about the boring and tedious work of reciprocally thanking the authors of the free software they take advantage of. It's essentially a protest about exercising the software equivalent of "good manners".

I'm not that cynical (although clearly cynical enough to think it) but wanting to avoid boring and tedious things isn't always a good excuse not to do them. Why so in this particular case?

vinniefalco commented 2 years ago

Why so in this particular case?

Again, because some companies and organizations (for example Boost) cannot incorporate software which requires attribution in binaries. No one is saying not to give credit to the authors, only that this credit should be limited in scope to the source code and not the binaries (since that creates additional burdens which some organizations cannot support).

phoboslab commented 2 years ago

because some companies and organizations (for example Boost) cannot incorporate software which requires attribution in binaries.

Then they need to write their own implementation or chose one that does not have this requirement. If someone want's to use my code, I want the attribution. Seems reasonable to me.

vinniefalco commented 2 years ago

If someone want's to use my code, I want the attribution

I think what you mean is, you want attribution in the executable as well not just the source code. I'm OK with people using my code without attribution in executables, because my happiness comes from helping other people not in seeing my name. And I can help more people by not requiring attribution in binaries, since my library can access a wider audience. Are you interested in your libraries reaching a wider audience and helping more people?

shuckster commented 2 years ago

What does happiness have to do with anything? People find that in very different ways. It kinda sounds preachy to bring it up. As for "accessing a wider audience", is it not the case that the measure of software success is a many tendrilled beast, influenced by more and bigger things than just attribution?

I mean, if a piece of software is popular/useful enough, it doesn't really matter what the license is, right? The users will demand the feature, so the developers will have to suck-it-up (in a very minor way in my own judgement) and just add the attribution. A mere one-time courtesy, not some ongoing maintenance nightmare. Certainly it's not in the same league as integrating the feature in the first place.

Also, isn't attribution less about "seeing your name" and more about leaving bread-crumbs that can lead others back to the source of a helpful thing? If so, in what way does removing it "help more people"?

Apologies if this sounds attacking. Perhaps changing the license to Boost would facilitate everything you promise. But being a moral huckster about it ("You want to help more people, right? And demanding your name next to your work makes your intentions questionable!") well, it doesn't seem very well mannered to me. MIT is pretty ubiquitous, and I think people know what to with it by now.

Anyway, all with apologies to the repo-owner. I don't quite know why I'm jumping on this...

vinniefalco commented 2 years ago

But being a moral huckster about it...

I wonder if you actually read this: https://www.reddit.com/r/cpp/comments/gr18ig/faster_integer_parsing/frxbit4/

(linked from https://pdimov.github.io/blog/2020/09/06/why-use-the-boost-license/)

pseregiet commented 2 years ago

People have the audacity to complain about having to insert some license text for receiving free quality code ? Apparently so...