This might be specific to software, based on what the GNU project has to say about it[0], I think you should do it like this:

  Copyright 2014, 2016, 2017-2020

(Meaning the work was completed in 2014, a change was made to it in 2016 and every year from 2017 to 2020 was a "copyrightable year".)

[0]: https://www.gnu.org/licenses/gpl-howto.html#:~:text=The%20co...

I feel there is a little product idea in this haha. Like copyright generator based on git pushes/deployments.

I know someone who submits PRs to large open source projects at the beginning of the year to update forgotten copyright values in their codebases in order to get listed as a contributor on github. It's a bit tongue in cheek but on first glance their open source contributions look pretty impressive...

Can just put [first publish]-[now] and save yourself the trouble.

That depends on whether each year until [now] had "copyrightable" improvements. For large open source projects, the copyright notice is placed on each file separately, and it is common to see some files not changing (or changing negligibly) over long periods of time.

Sounds much more like a feature, not a whole product IMO.

That's a hard sale to make around source control. I've written that script a half dozen times over the years, it's a few hours of work and the vast majority of the effort is in tailoring it to the particulars of the project.

Maybe it'd work as a plugin to a content management system.

When the content changes, that new content is under a new, and current copyright. So most of the time using the current year is legitimate.

Of course, it’s probably just automated, since updating the copyright with changes once a year is probably too big of a pain.

Yeah. In PHP, a lot of commercial sites just have date('Y'); after the copyright symbol, this just always gives the current year.

So i guess really one should so 2002-2020

A lot of commercial software such as Windows and Photoshop do this, and it must have been approved by legal. So I suppose this is currently considered the best way to do it.

A range like 2002-2020 conveys two ideas: that the work has been around for a while, and that it is still maintained. Both are important for corporate folks who want an assurance of stability.

I always laugh at the copyright dates on this [1] Australian government website. It claims 2003-2020 but I'm pretty sure it hasn't changed since 2003.

[1] https://www.edge.asic.gov.au/004/compportal/get/ServicesLogi...

Not to be pedantic, but wouldn't changing the year to 2020 itself count as a copyrightable change?

single words / short phrases aren't copyrightable so probably not

The whole work with the change would still be under copyright

Every year at the start of January, a load of slightly too clever for their own good web devs post code snippets all over the Internet showing how you can automate updating your site footer. They are all doing it wrong, for the reasons others have already mentioned, and every year better informed web devs reply to try to correct them. The date(s) shown should be the date(s) from which your copyright runs, and nothing else. A brief conversation with any IP lawyer will confirm this for you. Unlike most matters of law, this one is remarkably consistent from place to place, because the basis is an international agreement that basically every country in the world you're ever going to care about is party to.

There is a small but non-zero risk that by using a later year than the correct one you will actually weaken any legal claim you might need to make. The other party might turn up in court and say, "Look, nathan_f77 only claimed copyright from 2020 on this work, and it's on my site with a date five years earlier!". So automatically updating the date without leaving the original dates in place is a particularly bad idea.

In reality, the main practical purpose served by such notices these days is deterrence: it makes clear that someone specific does claim the copyright, for the benefit of the kind of person who doesn't really know how copyright works and might otherwise assume that because something is on the Internet then it's OK for them to take and redistribute it. They'd lose a legal action anyway, but those are expensive and time-consuming and there may be little to gain from one if the damage caused is minimal, so avoiding it by staking a very obvious claim might be helpful in some situations.

I like to do © LAUNCH_YEAR-CURRENT_YEAR. Showing the launch year is reassuring once you've been around a while. Showing the current year is reassuring in the sense that you're still maintaining the service/content/etc.

It is because the majority do not understand what it means, and just enter a current date because thats what they think it means.

Can not remember who told me about putting the first published date but it was an expert at the time ( we had a lot of copyrighted material).

Every time I go on some website and see that it's ©2016, I am convinced they've done nothing with it ever since.

It's probably a small risk, but I'd worry that just using the current year could bite me in the ass if someone accuses my site of infringing their copyright and we end up in court.

If mine just says ©2020 and theirs says ©2012, my argument that I created mine before they created theirs is weaker than if mine says ©2008.

At the least it means my lawyers will have to take time to explain to the jury why mine saying ©2020 doesn't contradict my argument. They may need to spend time presenting archives to bolster this, and then may have to spend time explaining how archive.org or similar works, and establishing that those can be trusted.

Each side in a civil trial is allotted a fixed time for arguments, rebuttals, and summaries, so any time my lawyers have to spend in court dealing with my ©2020 is time not available for most important things.

I've always assumed that most page's just use this code:

  &copy; <?php echo date("Y"); ?>

oh ya thats very interesting, i feel like the majority of websites just use the current year

My laziness paid off this time!