[JimDabell]

Don’t. The copyright year is not the current year. The copyright year is the year it was created / first published.

If you keep changing the year to the current year, you’re essentially lying about when it was created and artificially extending the duration of your copyright by one year per year.

If you put the wrong year, it renders the copyright notice invalid in the USA. This isn’t a big deal (creative works are copyrighted by default), but if you’re putting the notice there at all, presumably you care about it actually being legally valid.

Most people should just remove it. It doesn’t matter.

[spiffytech]

If you're putting the copyright date on a website, shouldn't it be the year of the most recent content/code change? I'd expect it to be similar to publishing a second edition of a book, which would get a new copyright date from the first edition. Older revisions of the website (e.g., in the Wayback Machine) would be subject to older dates, but the date on the page should be date of the version of the site the user is accessing, not the date of the first version of the website.

[nelgaard]

I agree. And apart from the legal questions, it is annoying to search for someting, get a website with a copyright 2021 notice and no info that it was actually written in 2016.

[LinAGKar]

>artificially extending the duration of your copyright by one year per year.

Are you? Isn't the expiration based on when the author dies, so the publication date doesn't matter?

[laurent92]

What about source files? Are they required to have OSS headers with copyright + licenses?

[Thiez]

No. When you just find some source code without a license (either as header or included as a separate fire) you must assume there is no license, so "all rights reserved" applies. So no header is required to protect the code.

What you may not do is just assume that the code you found is free to use when no license is included.