The first thing that comes to mind is why just US law? The general theory would (or should) be applicable to any country or entity that has laws and rules and departments and constitutions and so on. The rest of this comment isn't specifically what you've suggested, but it's very related and has been playing on my mind for a while.
I've often wondered what it might turn out like if a group of expert software engineers were unleashed on a legal web of laws and bills and so on, just to see how much they could simplify it and shorten it to make it easy to follow. You could even mark up laws and make them machine-understandable to allow computers to do the bulk of legal compilations before they were reviewed by a person at the end of the chain... which would save billions of dollars on lawyers for one thing.
To be honest and at the risk of sounding really arrogant, I think that in terms of information management and specification efficiency and usefulness, computer science is probably leagues ahead of anything that's currently being used in the legal profession. By this I don't just mean ordinary organisation of information, I mean the extra layers involved like references and the constants and the procedure and function calls and so on.
If you were to write law properly using that sort of theory then you could be sure, for example, that if a department changed its name then all that would need doing was changing the base value of that name. All the laws and bills and legislation that referred to it would update automatically, and there wouldn't need to be a plethora of patchwork updates to existing legislation in every corner of the system.
One of the big problems with this though is that lots and lots of people need to understand how it works. It's not just the lawyers, it's the people who make the laws and the people who write them down and the people who interpret them in court. Obviously it's important that an ordinary person be capable of understanding how to interpret the system if they're on trial for something.
For the last few years I've been on the commitee for a small organisation, and one of the things that's come up was revising the current constitution. This is a big job, because any change needs a majority vote of everyone at an AGM, and this can only happen once a year and still needs lots of very careful updating after that. The document has always been a very monolithic, with half of made up of information and rules that don't really need to be in a constitution, in my opinion.
For example, one of the proposed changes was to do with contacting members. Someone suggested updating it so we could also formerly contact people via fax and email as well as by telephone. I tried suggesting that the actual essence of this was that the communication was reliable and verifiable, and since we could be stuck with this for 40 years and communication methods are changing so often, maybe the constitution should have an external reference to values that could be changed more easily. Unfortunately it didn't go down well at all, especially with the retired guy who's spent most of his life being a legal arbitur and considers himself an expert on anything legal. I just dropped the matter after that fulfilled with the knowledge that 40 years from now we'll still be able to contact people by facsimile, but probably not by a videophone unless someone gets their act together and spends another three months of their life organising to update it.
Another big problem is that designing things like this really well is simply hard. For one thing, you need people who are really good at what they're doing. If computer science is anything to go by, how many software projects are there out there that can rightfully claim not to have a few hacks in them? And then there's the problems that come with not fully understanding how the issue might change at the beginning. Being caught out later can require substantial rewrites and restructuring. Unless everything is done very carefully by very organised people, I can easily see it turning into another big ball of mud, just like an ordinary software project.
But then in theory if done right, it would make it much easier to use a computer to derive existing legislation from the groups of rules and constants. Because it's derived from rules, the end legislation would theoretically be much smaller and simpler -- unlike the present when there's libraries full of legislation overriding earlier legislation and case examples and so on, most of which require the services of a qualified lawyer to interpret reliably, and even then the depth of interpretation could depend on how much you have to spend.
90% of the government buerocracy that goes into administration, maintenance and interpretation of laws and rules might promptly disappear in a puff of logical efficiency... and I guess this is why I don't expect it will ever happen. Lawyers make their living on interpreting and manipulating what I personally see as a very inefficient and muddy system, designed to be incredibly specific and yet is increasingly vague and debatable. For now at least, I'm quite happy to let the legal profession develop the theory it wants on its own. I don't mind preaching to the converted, but I think trying to argue it to most lawyers would be a waste of time -- especially not being a lawyer myself.
jesterzog Fight the light