"the law cares"
...and the law is wrong. It is wrong, because it is no longer serving the interests of the people governed by it. The entire basis for the US legal system is that laws serve the interest of the people, not that laws are handed down from above or that the law itself has any moral or ethical justification. If society's interests are no longer served by a law -- and it is hard to argue that in the case of academic publishing, copyrights actually serve society's interests -- then the law should repealed.
"Do prosecutors and judges have to agree with that statement? What happens when they don't?"
They become the targets of congressional investigations, which is what is happening to Carmen Ortiz right now. It is horrifying that it took the death of a man with a bright future to bring this issue to the public's attention, but now that everyone is looking it is good to see Congress doing something that helps people for a change. Minute as the response from Congress has been, it is at least a step in the right direction.
"prosecutors and judges have to enforce the law as it is"
No they don't. Prosecutors have to serve the interests of the public and are expected to decline prosecution in cases where they are asked to prosecute someone whose actions were potentially illegal but both benefited the public and were neither harmful nor dangerous. Judges are expected to strike down laws that violate the constitution, and to give light or suspended sentences to people who are guilty of violating the law but whose imprisonment or punishment would do nothing to benefit society. The idea that prosecutors and judges have some kind of obligation to enforce laws that are clearly misguided is on the level of the fallacy that juries have a duty to convict if the evidence shows a defendant is guilty (jury nullification is an important check on laws that the public disagrees with).
"does it make scanning and digitzing old paper journals free also? Does it make it so that data entry techs typing in metadata describing old paper journals don't have to paid?"
That is an organizational problem, not a technology or legal issue. Universities could do this via their library system, or we could set up a system where people who voluntarily scan articles or write metadata get preferential treatment when they search or download (think of BitTorrent trackers that give preferential treatment to users with higher seed ratios). For that matter, why should we have to rely on a single organization or system? Maybe you are better than I am when it comes to entering metadata; maybe other people would rather trust your metadata over mine. Maybe we could set up something like a web of trust.
It is not like this could not be done; it is more like the law stops it in order to protect an obsolete business model.
"How are we going to do that exactly?"
Here's a suggestion: exclude scientific articles from copyright restrictions. If you want a compromise, how about this: exclude those articles for which the research was paid for with tax dollars from copyright. As a first step, that would not be so bad. Let the copyright battle be exclusive to entertainment and software, and leave human knowledge out of it.
We are already questioning the CFAA as a result of this case. Why not question copyrights as well?
"So, copyright laws will be enforced when it comes to stealing proprietary source code, but not for old journals?"
That would at least be a step in the right direction.
"Is it that copyright law shouldn't be enforced because you've decided journal publishers are rapacious parasites (I think they are too actually)? Or that copyright law should only get enforced for entities that contribute something to human knowledge?"
It is that copyright law was meant to ensure our access to the knowledge that journal publishers publish, but that in today's world it is actually making that knowledge harder to access. The copyright system was never meant to ensure that the authors of scholarly articles would be paid; it was meant to ensure that the publishing system would be monetized. At the time, the publishing industry was the best way to make enough copies of scientific journals to satisfy society's need, and the publishing system is expensive and must be paid for. Everything other than the distribution of articles is irrelevant to copyrights, because everything else is done by volunteers (writing, reviewing, and editing) and would continue to be done regardless of money.
We no longer need that system, because we have a distribution system that is substantially more efficient and less costly. Copyright is no longer promoting the best system for distributing scholarly articles; it is now preventing the use of that system and promoting an anachronistic system that by today's standards is shockingly inefficient. It seems reasonable to conclude that copyrights should simply not be applied to academic publishing, at least as copyrights currently exist, and that a new legal framework that promotes the use of the Internet should be enacted.
Sure, there are hurdles. Publishers serve as archives, they pay people to insert metadata, and sometimes they even scan older articles for us. I fail to see how these activities could not be accomplished with the publishing industry. Universities already have more than enough computer capacity to act as archives (I know a professor who took the time to scan his entire library, thousands of textbooks -- it fits on a couple of external hard drives sitting on his desk, with room to spare for storing all patents ever granted in all countries). Metadata is something that could be done as a community, perhaps via some sort of system as I had mentioned above, or it could be done by the universities themselves as a one-time cost for old articles (metadata for new articles could be inserted automatically when those articles are published). Scanners are not hard to find -- my school's library makes many available to patrons, I have one on my desk, and my department has many as well (and of course, I know someone who scanned thousands of books on his own); again, we could create a system that incentivizes scanning as a community effort, or rely on universities to do it. If we actually cared about the original goal of copyrights as related to academic publishing, we would be investing some time in finding better ways to do things now that we have the technology to do so.
You could summarize the argument like this: the law is standing in the way of the sort of progress that society is desperately in need of; therefore, changing the law is imperative.