On 6/23/07, Terrence Brannon <[EMAIL PROTECTED]> wrote:
On 6/23/07, Raul Miller <[EMAIL PROTECTED]> wrote:
> 1.2 "Two ways to compute the square of a number:"
> *~
> ^&2
> *:
> (]+^:[0:)
> I'm sure I could come up with more
yes, I have changed the wording to:
Two ways to compute the square of a number, one of which must be
monadic.
All of the above are monadic, as are definitions like
"Monad" and "Dyad" are syntactic characteristics, and all of the
above are, syntactically monads. [Though, I will grant that all
of the but one contains as a part of its definition a dyadic verb.]
Anyways, if you want a way of squaring a number without using
any monads in the definition, you can start with a definition like
(]>:@]^:[^:[0:)
and then convert it to explicit form, to get rid of the dyads
3 :'>:^:y^:y 0'
Note that this somewhat simplifies the expression, since we
can take advantage of the parser to get our arguments to the
instances of the ^: conjunction.
I want to make sure that *: is used.
For that, I'd ask for a primitive (single token) expression.
function is really a verb. argument is really noun. application and
value are OK terms. Added wording:
Please be aware that function and argument are not commonly used when
discussing J. The terms "verb" and "noun" (a.k.a. array) will be
introduced later.
Ok, but I suspect you should explicitly state why you are limiting the
meaning of "function" so that adverbs and conjunctions are not
included.
> 1.11
> Some of your other terminology also seems odd, to me. For example,
> I figured out that by "formal name" you mean the J primitive (+ for
> example),
Section 1.11 of http://www.jsoftware.com/help/learning/01.htm
describes formal and informal name. The informal name for + is Plus.
Ok -- that's not official of course, but I guess it makes sense in that
context.
> But what does "family size" mean?
verbs in J come in a family size of 6.
(monad, dyad) * (plain, dot, colon) = 6 possibilities.
Concretely:
monadic + +. +: together with
dyadic + +. +:
form a family of 6 related verbs which all do about the same thing.
Oh, I see -- you're talking about lexiography and syntax... though
I think in this context "verb" is the wrong word. Consider, for
example ; or [
And, on the other hand, 3 :string IS a verb.
Anyways, personally I would make the more specific (limiting it to characters
where the "right answer" is correct), or toss it entirely. Or perhaps
rephrase it to get at the concepts from other directions.
> Also, I see other ambiguities similar to your "two ways to square a
> number" question. For example "Add 1 to 7 monadically"
> 1&+ 7
> >: 7
> [EMAIL PROTECTED]
> 1 1&p.7
oh, yes, well this is chapter 1 and Bond has not been introduced yet.
Nor has Atop. So pretty much any new reader will only choose >: 7
...if you can guarantee your reader is not curious enough to have
examined the dictionary. (Which, frankly, seems like a losing
proposition.)
On the other hand, as long as you don't explicitly require a certain
answer, prospective users of your test who do things like
{.\:~7?8
are not a problem.
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
