> Law is code. Why haven't we fixed this yet?
Properly drafted English-language laws do read as code. It's like how you can describe formal logic in English: "All Americans wear hats. I am an American. Therefore I wear a hat."
I take it you're arguing for something like Lojban, though. I see two problems there. First, it's not the (de facto) national language, which infringes on one's fundamental right to read the laws of the land. Second, and probably more incurably, these types of languages don't remove the most important kinds of ambiguity in laws.
Most disputes over ambiguity in legal text do not arise from syntactic ambiguity. It's not usually confusion over what words the "not" modifies, or which "if" is nested within which. Rather, most problems arise from semantic ambiguity, i.e. the meaning of the individual words. What constitutes a "bank," exactly? What does it mean to do something "in a reasonable period of time?" If you're "bearing" arms, where exactly are you bearing them?
You might argue that the problem then comes down to defining your terms well. That's certainly a noble goal when drafting legal text. But one of the principle goals of legal drafting is to make the text flexible enough to apply to specific circumstances that the authors could not have foreseen. This, of course, is in constant tension with the need to be precise. The more precise you are, the easier it is to interpret the law, but the more likely it is that someone will find a loophole arising from the excessive narrowness of your language.
A language like Lojban won't alleviate these difficulties, unfortunately.