but the owner of copyright in an anthology of poetry could sue someone who infringed the whole anthology, even though the infringer proves that publication of one of the poems was unauthorized.
Whoa. You're not reading that whole section correctly. Moreover, you're quoting the analysis, not the law itself.
What this sentence from the analysis says is that the owner can sue somebody who published the entire work in an infringing fashion, even if the infringer can prove that one part of the work was not specifically authorized for publication by the infringer, or was infringing copyright.
IOW, if I write a poem and include it in an anthology, I can sue for infringement if somebody uses my work in an infringing way, even if they can prove that some
other poetry was infringing! Two wrongs don't make a right, basically.
It's a basic concept of fairness here- the poet, in that case, may not have been aware that the anthology is not legal, but still has protection over their work.
Now, does that cover stuff put into CA? Well, it covers Trepan, and certain pieces of code put in that were originally published elsewhere, I think. But I think that anything published via SVN is almost certainly
not protected.
Let's go back to the cookbook example, I think it's easier for people to understand.
I've published a cookbook. I am the copyright holder.
Now some guy takes my cookbook, scans the whole thing, and puts it onto the Internet, for free, in an editable PDF- which is clearly infringement.
Next, some
other person comes along, and edits the PDF, adding some new recipes, and tries to pass it off as his own work.
I, the copyright holder, can sue
both of them. That's what that whole section is trying to explain.
Copyright starts with the original holder- in this case, Atari. Cavedog, like any developer for a big publisher, would have had to give their rights to Atari (or whoever owned Cavedog back then, who would have sold those rights as part of Atari's purchase, etc.).
Even "innocent infringers", who thought what they were doing was legal, aren't immune. For example, there are lots and lots of retro-game sites that have had to remove many games, due to DMCA requests. There have been many, many, many instances in which Nintendo has either shut 'sites down, or eradicated the distribution of Super Nintendo ROMs, even though they haven't made the hardware and no new games have been published in years. They'll probably stop someday, but not yet- I read, the other day, a ROM 'sites declaration that certain titles, like the Final Fantasy series weren't being distributed because they'd been threatened by lawyers. The people distributing these works weren't the hackers who got rid of copy protection or copied the ROMs. But they're still infringing.
In short, no, that section's not providing a fig leaf for CA. Just because Caydr was the original copyright violator does not mean that further violations are all right- otherwise, people putting out hacked copies of games would just change a few things and say it was theirs, for example, which is clearly illegal.
I think, honestly, that the more I look into this, the more it looks like I'm right about it, and that the only wiggle-room concerns "mere aggregation". Meh, I have the case law coming, I've just been away from home for over a week, so I haven't had much time to do all of this. Tomorrow, when I'm not getting the new nifty models done for a new unit in P.U.R.E., I'll be happy to try and explore this further, but tbh, I think your line of reasoning is based on hopeful assumptions.
If we can't find a clear way to prove that games for Spring represent aggregations, and that they're to be considered complete software programs that just happen to have a bunch of subroutines written in several languages or scripting methods, then it's we're in a very bad place here, imo- games would either have to be wholly GPL, or not, due to the copyleft issues.
Now, would that prevent other projects from using CA's code legally? I really don't know. I'm not sure how the law would apply here.
That's why I want more clarity here about aggregation. If we've established that data structures aren't being shared, and that we have APIs that serve to provide, among other things, separation between same, and we know that each COB script and each Lua script, unless it specifically calls another one, is probably separate... then it appears to support the idea of an aggregate, at a technical level.
However, from a judge's perspective, it'd be very hard to argue that an entire game is not a "work" but is merely a bunch of separate code that was combined and produced a given result. It'd be like arguing that Windows is just a few major DLLs and EXEs that just happen to produce an operating system- I don't think anybody would buy that, in reality.
I'm really looking for any serious arguments that would strengthen the aggregate argument, though, as I think it's about the only thing that salvages multi-licensing as a serious legal concept. Without multi-licensing, using GPL looks incredibly unattractive to most people. Share-alike is one thing. Having to put an entire game under the GPL, because of a few lines of code somewhere, is why the past license wars have been so awful. We'd be right back to saying that E&E is now GPL, for example, and I would have to yank the one piece of GPL'd code I use besides World Builder (the Healthbars Widget) or take extreme measures to demonstrate that it's just an addon module, and in no way establishes dependency
And using gamecode of any kind would be flat-out impossible. Miserable situation for all concerned, tbh.