>> What I do not know, is how to construct conjunctions tacitly; >> so, I will be awaiting to see what Dan might have up his sleeve. > >As I said, I'm not yet convinced this is possible. The only hint I have is >my tacit (nee "anonymous") evoke adverb can produce a conjunction. Yes, I got the part that you were not sure yet; that is why I wrote "might." I also knew that an adverb can return a conjunction (e.g., (<'@') `:6 ) that is why I wrote "construct" and I still do not see how a conjunction could be constructed tacitly. >m v Ag3 would be equivalent to the DoJ definition of [email protected] (note the >absence of the ` ). I feel this would be a "purer" approach. I see what you mean now and I agree (a la Oleg's strand notation implementation). >[1] Because operators have the highest precedence in J, no other part of >speech can take them as arguments. This leads me to believe that any >adverb that produces a conjunction must have that conjunction quoted & >passed in as an argument (noun or verb), in one form or another. Right, operators cannot be taken directly as arguments. You can pass them indirectly as strings, as you said, but you could also pass them indirectly as (their) atomic representations which I feel blend (together with other arguments) more naturally into a gerund to be processed by an adverb, for example, via a tacit verb that can, in principle, generate another atomic representation of any arbitrary complex tacit adverb (which is a function of the gerund) to be evoked in the last step. The gerund, as you suggested, could also be provided in strand notation as well (Oleg's version relied on a global noun to accumulate the elements but, I think, this might be avoided by embedding the accumulation in the resulting adverb at each step).
________________________________ From: Dan Bron <[email protected]> To: Programming forum <[email protected]> Sent: Sun, November 22, 2009 5:28:13 PM Subject: Re: [Jprogramming] Wrong agenda? Pepe wrote: > We can model, if I am not mistaken, the expected behavior of [email protected] >explicitly as: Yep. My first model was a bit more convoluted than yours, because it also supported [email protected] (in an attempt to use the interpreter [j] to implement @. according to the DoJ [J]). The second model was similar, except that the conjunction, given two arguments, produced a verb, instead of waiting for the argument(s) to the derived verb as well (and so was more faithful to @. than the first model). > What I do not know, is how to construct conjunctions tacitly; > so, I will be awaiting to see what Dan might have up his sleeve. As I said, I'm not yet convinced this is possible. The only hint I have is my tacit (nee "anonymous") evoke adverb can produce a conjunction. And, as the name implies, the adverb is defined completely tacitly. However, I'm pretty sure the conjunction produced has to be quoted in the first place, which may render the phrase non-tacit. The question is almost equivalent to "with conj=.'@' then is 'conj'~ tacit?", except that the intermediate assignment isn't required. My instinct is to answer no, and furthermore that the answer may even be irrelevant, because the requirement to quote the conjunction defeats the purpose of tacit code to an extent (e.g. +1 :'@' would be a shorter way to express the above). So, unless I can find a sneaky way to produce a true tacit conjunction in J (which is probably impossible [1]), I would not have approached Viktor's question this way. > it is possible to write tacitly and adverb Ag2 so that (m`v) Ag2 would > be equivalent to the expected [email protected] but it would not be straightforward Yep. It was pointers of this sort that I alluded to in my original message. Though I was thinking more along the lines of an Ag3 so that m v Ag3 would be equivalent to the DoJ definition of [email protected] (note the absence of the ` ). I feel this would be a "purer" approach. -Dan [1] Because operators have the highest precedence in J, no other part of speech can take them as arguments. This leads me to believe that any adverb that produces a conjunction must have that conjunction quoted & passed in as an argument (noun or verb), in one form or another. Barring that, I think we'd need a special train that produced a conjunction, and all trains of that sort were removed in J5. Hence my intuition that a "truly tacit conjunction" is impossible in J5+. But I would love to be proven wrong. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
