Hacker News Reader

Sometimes I wonder if it wouldn't help to try to set legal precedents that extend todays extreme intellectual property obligations in software to legal documents.

When you come to think of it. A contract is very similar to a piece of code. The main difference is that it is to be executed by humans instead of machines and enforced by a legal system instead of a machine's design.

Contracts have been deemed non patentable and a high bar has been set for copyrightability compared to other text because of the requirement that contract language be well defined with particular sentences having an agreed upon interpretation based on court precedents.

This means lawyers writing contracts don't have to worry, for each sentence that they write, that it may be infringing on someone else having written a similar one. They also don't have to worry that a combination of clauses have been patented and they can't use it in their contract.

Programmers on the other hand worry about it for every new piece of code they write. The bar is so low for patentability and copyrightability that we pretty much know someone somewhere has patented a large part of the code we write. We just hope nobody with the patent is interested in going after us because we aren't worth the trouble. But there is always that feeling that if ever we become commercially successful, the trolls and wolves will be waiting to try to get a slice of our profits. This is quite an annoying disincentive to write new code or 'innovate' which in this case means just plain doing our job.

I was pondering this further and it would actually be conceivable with a bit of natural language processing and machine learning to build a 'verdict machine', maybe a Watson type of thing.

You would input legal documents, contracts and laws to 'program' the machine, It would read these legal documents and compile them into legal tests that depend on case facts.

Then when you needed to judge a case, you would enter the facts or fact probabilities about the case in the machine. It would then generate a verdict and tell you which legal tests passed or failed to arrive at the verdict.

Even it it was just used as support to help courts untangle cases or for law student as an educational tool, it could be a useful machine.

Here is the kicker: All contracts and laws are actually software code for this machine.

You've certainly got lots of data to train it with. And you can test it by simply testing its predictive power. But instead of selling it to the court system I'd sell it as a service to law firms and/or use it in prediction markets. This would ensure that it would be funded and refined until everyone simply accepted its outcomes as a foregone conclusion.

The fault of most modern politics is that politicians are taking action on issues they have no idea about, and the only words they hear about such problems are those paid for with the largest pocketbooks.

I think the fault runs even deeper than that. The idea that "if lots of ignorant people work together and find common ground to agree on, something intelligent will pop out" seems to be a cornerstone that we have built our society upon. Two ignorant legal teams duking it out is not the recipe for a sensible outcome; the only thing that happens when one of them wins is that they can convince themselves that the victory must mean something profound.

The only hope for the system that I have seen so far is surprisingly intelligent and sensible judges injecting sanity into the system.. but that is hardly something to rely upon.

I don't think anyone ever planned in the long term for the system to evolve into ignorance. I don't remember any hard evidence, but I would figure earlier on in Americas history congressmen might consult intellectuals and experts on issues they were debating in congress, or maybe a judge might pull in an expert. I think practices like that fell out when law became such a profound convoluted discipline unto itself that anyone without a law degree was disqualified from engaging a court at an intellectual level.

From my perspective, the fact that the US constitution was given an initial set of amendments, as well as making amendments unusually difficult for the 'will of the people' to change (and as well as all the other safeguards put in place) suggests that the drafters/signers of the original document were placing more trust in themselves than in the democratic process which would follow. The safeguards I think were viewed as necessary because 'the will of the people' wasn't viewed as something that would, by definition, be good. Basically I don't think they really trusted democracy.

That's just my, probably controversial, take on it though.

They just didn't want controversial ideas to make it into the constitution since they would be so hard to overturn, so they made the amendment process insanely complex.

The track record on that isn't too stellar though, considering prohibition and what not.

I remember quotes about Jefferson thinking the constitution should be rewritten every generation. Not sure about the accuracy there, but I like the idea.

> So often in reading those transcripts I wanted to grab one of the lawyers and scream in his face, "you idiot!! What the hell is wrong with you?!? Are you stupid????"

I can think of a couple of reasons that don't involve lawyers being stupid:

1. If the truth is against a client's interest, a lawyer is probably going to try to confuse the issue as much as possible (unless it seems tactically unwise for other reasons, of course).

2. IIRC, they were stuck with some bad precedent to work with (especially with respect to APIs). That meant they were often arguing issues in a broken frame, rather than from first principles. My understanding is they can't effectively make the first-principles argument until the appeals get to the court that created the precedent in the first place (or higher).

So to be fair here, there were two issues at play:

1. Oracle's patents that Google supposedly violated. (And thank God most of those stupid patents were actually invalidated)

2. Oracle's copyright on its APIs that Google also supposedly violated. The API discussion involved the copyright part of the case, and here the question "What is an API" does become central, especially considering that the judge decided in the end that APIs are not copyrightable - when such a decision is made it'd better be crystal clear what is meant by an "API"

Admittedly, the API example was not totally on point since it was dealing with copyright. However it illuminated the problem with lawyers even under the best of circumstances not understanding software and programming.

I agree that the question of "What is an API" was quite important for good reasons in that case. What was gravely concerning to me was that the lawyers from neither side had a clear idea of the answer, despite the fact that had the best possible set of circumstances for lawyers to get that kind of understanding. Specifically, their clients gave them access to some of the smartest people in the world, they had virtually unlimited time and money, and the lawyers themselves are some of the best in the country and were experienced in complex litigation with technical companies. Extend this kind of conceptual ignorance to the lawyers who are arguing about and making policy on software patents, and it's both terrifying and infuriating.