Interesting Links for 11-04-2016

[andrewducker]

Africa's great opportunity for reform (tags: africa government Education tax economy ) The Largest Ever Analysis of Film Dialogue by Gender: 2,000 scripts, 25,000 actors, 4 million lines (tags: gender data research visualisation movies ) Bitcoin Start-Up Gets an Electronic Money License in Britain (possibly cheaper way to convert pounds to

Comments

[simont]

But if someone then generates that same image algorithmically can _they_ copyright it, even though it already existed?

I believe so, and even more, they don't even have to have regenerated it themself. If you generate a big slew of images algorithmically and then I go through them and select ones that I like, then the copyright is mine, because all you did was algorithmic generation (which doesn't count) and I'm the one who provided the 'creative input' in the form of filtering (which does).

I think what it boils down to is that whether or not an image came out of some algorithm somewhere at any point doesn't affect anything about copyrightability.

But then, perhaps a key difference between copyright and patent law is relevant here: patent law is concerned with prior existence, whereas copyright law is only concerned with provenance. If you and I both independently write identical stories, then in principle, neither of us has infringed the copyright on the other (though, in practice, convincing a court that that's really what happened is about as unlikely as the hypothetical situation is to begin with). But if you and I both file identical patent applications, then one of us gets it and the other doesn't. So perhaps that would make a difference, in principle.

On the third hand, let's not forget that the law isn't a computer - there's always scope for the institutions actually applying the law to reinterpret it through the lens of common sense, even in the US. It may well be that the first test case for this will just get a precedent established by the appropriate supreme court that says 'don't be bloody silly, that's not what "existed" meant in this context'.

[danieldwilliam]

There's something in the provenance idea about the standards of creativity required for patents or copyright.

The standard for copyright in the UK are quite low. The right protects the very particular expression of an idea rather than the idea itself. You and I can both write a comic about an anthropmorphic mouse. I get Mickey Mouse and you get Maus and both of us get copyright. If I translate or annotate your out of copyright work I get copyright in the complete edition but not in the original out of copyright work.

Patents have a higher standard of novelty. Even in the US where things are laxer I think. You have to produce a genuinely new product or process. If I take your invention of a conveyor belt for moving apples past a series of pickers and packers and apply it to oranges I've probably not been creative enough to get a patent.

[simont]

My understanding of the US situation has always been that in principle there's a fairly high standard of novelty, but in practice the USPTO is terribly sloppy about checking it.

So you might perfectly well be able to get patents granted that are about as silly as the example you give there. They wouldn't hold up for 30 seconds if they ever got as far as court, and so (if you're the sort of unscrupulous person who does this sort of thing) you don't actually start serious lawsuits over them, but instead you use them to bulk out large patent portfolios which you then use to threaten people into doing what you want without going to court. Perhaps most of your patents are bullshit, but your victim is going to be as risk-averse as the next corporation and won't want to call your bluff.

[danieldwilliam]

Yeah - the US Patent authorities don't appear to be doing a great job of making patents work for the wider economic good.