Well... http://jsoftware.com/help/primer/tacit_definition.htm defines tacit as "In a tacit definition the arguments are not named and do not appear explicitly in the definition." (It also needs to be updated to use x and y instead of x. and y. when referring to the arguments of an explicit definition. Though I really wish that that whole migration of argument names had been done more gradually... Actually, given that it was purely for the convenience of OO coding, I'm sort of wondering if it was even a good idea in the first place. Oh well... nothing is perfect.)
So when you say 1 : 'a u ]' returns a probably tacit verb, you are not really saying that 1 :'a u ]' is tacit. Instead, you are saying that u is probably tacit, and that a is probably tacit. But that does not mean that 1 :'a u ]' is tacit. But I do agree that the locale and timing difference for resolving the definition of 'a' is worth calling out. And I do appreciate your use of the word modifiers - but it probably deserves a mention where you use it in your essay? As for this: a_t_=: 2 a_base_=: 6 cocurrent 'base' +: a2_t_=: 1 :'u a' 12 +: a2_t_ 4 The issue here, I think, is that we did not use the name a2_t_ to find the definition we used to calculate the value 12. We set a2_t_ to that value, but that's just a side effect. If expected +: a2_t_=: 1 :'u a' 4 that would have some confusing implications. In particular, it would mean that we would expect this: a1_t_=: a2_t_=: a3_t_=: 1 :'u a' a1_t_ a2_t_ a2_t_ a3_t_ a3_t_ 1 :'u a' What we have, instead, is simpler and easier to reason about: a1_t_=: a2_t_=: a3_t_=: 1 :'u a' a1_t_ 1 :'u a' a2 _t_ 1 :'u a' a3_t_ 1 :'u a' In other words, the current design is that if we assign the same value to several names they all get the same value. I am comfortable thinking of that as "not a bug". Thanks, -- Raul On Wed, Sep 17, 2014 at 3:10 PM, 'Pascal Jasmin' via Programming <programm...@jsoftware.com> wrote: > 1 : 'a u ]' is adverb that returns an anonymous and probably tacit verb. I > say "probably" because u could be explicit, but the term tacit still seems > appropriate to describe the overall expression. > > 1 : 'a u y' is an adverb that executes its explicit definition in the > caller's locale. It could be stated as it is not evaluated until it is > passed y, or it is bound with its unsubstituted definition. > > In the first example, a will be evaluated (and substituted with a constant) > in the adverb's locale. In the second, a will be evaluated in the caller's > locale. So, IMO, its worth calling these 2 different types of adverbs. > Something purely tacit such as @] is still like the first type of adverb > > I use the term modifiers, because this distinction applies equally to > conjunctions. > > conj_t_ =: 2 : 'a + v + u ' > conjE_t_ =: 2 : 'v + u +a + y' > > +: conj_t_ +: > 2 + +: + +: > +: conjE_t_ +: > > +: (2 : 'v + u +a + y') +: NB. what is returned no longer has any reference > to t locale. > > I understand the nounconj reference makes for a distraction on a seldom > considered topic, but it seems like a separate enough type of modifier > > a2_t_ =: 1 : 'u a' > +: a2_t_ > 4 NB. returns noun > > +: 1 : 'u a [y' > +: (1 : 'u a [y') NB. returns verb phrase that will be parsed/executed and > use a in base/caller > > the following seems like a bug to me, > > a =: 6 NB. in base > +: a2_t_ =: 1 : 'u a' NB. seems to pre-optimize to get a from base. > Though behaviour is correct if defined on separate line in base. > 12 > > erase 'a' > 1 > +: (a2_t_ =: 1 : 'u a') > > +: a > > > ----- Original Message ----- > From: Raul Miller <rauldmil...@gmail.com> > To: Programming forum <programm...@jsoftware.com> > Cc: > Sent: Wednesday, September 17, 2014 2:00 PM > Subject: Re: [Jprogramming] OOJ and calling a verb from another locale > > Reading that article, I stall when you say > > "Even if they are defined with 1 : or 2 : , the first 2 types of > modifiers should be considered tacit," > > I do not know what you are referring to by the phrase "the first 2 > types of modifiers". Do you mean "the first 2 types of adverbs (or > conjunctions)" or are you referring to nounconj and nounconj2? If the > latter, I think the statement is erroneous. If the former, I'm a bit > dubious about the distinction. > > The concept of tacit is slippery enough that it's probably worth > quoting the definition you are using, and describing what it is about > the context that makes "tacit" a relevant concept whenever we talk > about it. > > Anyways, I got stuck there, and I thought you should know. > > Thanks, > > -- > Raul > > On Wed, Sep 17, 2014 at 12:54 PM, 'Pascal Jasmin' via Programming > <programm...@jsoftware.com> wrote: >> You may find this article helpful: >> http://www.jsoftware.com/jwiki/PascalJasmin/3%20types%20of%20adverbs%20conjunctions%20and%20binding >> >> These are indeed issues you understand by surprise (why is this code not >> doing what I meant it to do) comming from other languages. >> >> >> ----- Original Message ----- >> From: Brian Schott <schott.br...@gmail.com> >> To: Programming forum <programm...@jsoftware.com> >> Cc: >> Sent: Wednesday, September 17, 2014 12:05 PM >> Subject: Re: [Jprogramming] OOJ and calling a verb from another locale >> >> Pascal, >> >> I especially liked the point you made below. I think such a point would be >> very appreciated by new J programmers and may not be well known. >> >> >> >> >> "... In J, the verb (2 + myvar"_) will produce a constant verb based on >> the value of myvar at definition, while (2 + 3 : 'myvar') will obtain the >> latest value of myvar. ..." >> >> >> >> >> -- >> (B=) >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > > >> >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
