On Thu, Sep 6, 2018 at 12:03 PM Henry Rich <[email protected]> wrote: > > I agree with your analysis to that extent that I think it shows a > discrepancy between the Dictionary and the implementation, and also some > ambiguity and error in the Dictionary. > > I think the current implementation is what we want it to be. Therefore I > think the Dictionary should change. > > My changes would be as follows: > > 1. A gerund is an atom or list, every atom of which is the AR of a verb.
I think this conflicts with the current implementation if m is a gerund in [email protected] I think I presented some examples which demonstrate this in http://jsoftware.com/pipermail/programming/2018-September/051825.html Then again... maybe what you are saying is that the left argument to @. is not necessarily a gerund??? > 2. In [email protected], m`:n, m}, ^:n, and all partitioning modifiers, the operand > that can hold ARs must be a gerund. Again, ... I think this conflicts with the current implementation. > 3. In [email protected], the result of the execution of v must be an atom (if you > omit this rule this could be a nonce error, but tomorrow will never come) This seems like an issue which should be easy to fix (since n in [email protected] does not have to be an atom). > 4. In [email protected], the atoms of m can be any valid ARs and the result can be > any of the principal parts of speech. Er... wait... so m in this case is not expected to be a gerund? I am going to need some time to see if I can digest that point of view. That said, in the mean time, I would imagine that the left argument to the Evoke Gerund conjunction (`:) must be a gerund, and that context certainly has allowed trains which included non-verb ARs. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
