On Sun, Jan 10, 2010 at 11:50 PM, Dan Bron <[email protected]> wrote: > As I noted in an earlier message [3], we could make any number of > retrospective justifications (knowing how the interpreter acts and building > a justification for this out of the DoJ), but this amounts to apologia. > What we want is deductive (as opposed to inductive)/prescriptive/predictive > rules, such that if we did not already have an implementation of J, we could > build one.
Actually... The dictionary is a reference work but that does not mean that all text in the dictionary must stand alone as reference material. I think you are always going to have chicken/egg problems for something like this, where advanced concepts build on other advanced concepts. Specifically, the rules in II.E detail how parsing works. You might be able to get around this particular issue by resorting to technical jargon used to describe computer grammars. But then you leave your audience cold if they do not have that kind of background. You could work around this problem by including the necessary background but that adds verbosity. And, really, all we are trying to do is make sure we have a properly defined parser, when we have the parsing algorithm detailed on the same page. I can see the value in having a tutorial work which covers this ground in greater detail and from several different perspectives. But I would be quite a bit more conservative about proposals involving the dictionary. Anyways, a verb phrase is a phrase which produces a verb. But there are grammatical rules about such phrases (detailed in II.E) which have to do with when they can be used. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
