[nanny]

>Sure, because it's MIT license, you can relicense, but is replacing the original copyright notice OK? Even if it's OK, why did they do that?

The file headers don't matter to the MIT/Expat license, what matters is that the original copy of the license is included in the redistribution, which it is (it's at the bottom of lib/zen/std/LICENSE). Replacing the file headers makes sense in this context because the derivative work is now (correctly) copyrighted by connectFree. Zen looks to be in total compliance with the original Zig MIT/Expat licensing terms.

[eternalban]

Note to self: don't ever again use MIT license for my OSS projects. [ascii art fan here..]

Apache says something about "retaining category A header licenses" but not sure what category A means.

[nanny]

Just know what you're getting into. Choosing the MIT/Expat license for your project is a conscious decision that explicitly allows people to do things like connectFree did with Zen. Whether you consider that a good or bad thing is on you. Although, I do highly recommend GPLv3, AGPL, or LGPL. You can read more about all kinds of software licenses here: https://www.gnu.org/licenses/licenses.html

As for the Apache, I don't know anything about that and I can't find anything about headers in the Apache v2.

[eternalban]

> Just know what you're getting into.

This is the whole issue. It's all so complicated. Thinking about it, I suppose it should have been obvious they can just yank comments from an altered source. Somehow I had the wrong idea that original top-level headers were left untouched. IIRC I may have even seen appended top-level lic notices in the wild. Thanks to your informative comments here, I now know otherwise. The thought occurs that OSS licenses are very much mired in du jour technology of writing code. If a future language supports metadata for code, such as history, I would think none (?) of the current batch of OSS licenses would protect the metadata.

I've been warming up to GPL flavors for a while as well now. I'm even careless enough (I seem to suffer from a tendency to "wrong think") to toss around in my head a re-evaluation of merits of non-OSS licenses as worker in this field. An unfortunate thought that keeps cropping up is that "big business and big brother the ultimate benefitiaries of OSS".

[nanny]

Re: metadata: I suspect that it may indeed be already incidentally covered by existing licenses. I replied elsewhere with a link to this wikipedia page: https://en.m.wikipedia.org/wiki/Structure,_sequence_and_orga...

The SSO of a piece of software is called a "nonliteral element". Just speculation on my part here, but I'd bet that version history might fall under this "nonliteral element" part of copyright law.

[shawnz]

Let's say the MIT license demanded that you retain the copyright notice in all source files in addition to just the license file.

What benefits would that give you over the current situation? Anyone who makes a copy already has to include the license text, why does it matter where they include it as long as it's in a conspicuous place?

The real decision point here should be: Do you want people to be able to use your work for commercial purposes, for free, without recontributing their changes? If not, then don't use the MIT license

[eternalban]

You're entirely correct.

Just now I remembered that back in my squandered youth I used to consider code to be a form of creative expression. Code is not my only creative outlet, but it remains an important one. I should not be using OSS at all. I don't even code for the user if I am ever honest about it: I code for the thing itself. Somehow, it seems, I got caught in a dominant paradigm and lost a sense of my own self and values regarding my creative work.