ST> F i. `john
You are wrongly assuming that one could write symbol literals as
they are formatted in the default output (marked with backticks).
This is true in K, but not so in J; there simply is no way to
denote symbol literals in J.
What your attempt above actually (but unintended) does is employ
the ` primitive for gerund constructions, and that in an illegal
way, ending up with an error message.
ST> NB. so lets try to define ` as a verb. Is there a limit on which
ST> characters can form a verb?
All J primitives are denoted by ascii special characters which may
be followed by dots and colons; add a few alphabetics/words which are
ALWAYS end in a dot or colon. You cannot redefine these primitives
of J.
Easy to distinguish from those is the namespace for user/library-defined
words are the alpha_numeric identifiers, all re-definable whenever you
want to.
There are no ways to play tricks with this clear border.
ST> .... (s: ' john') NB. space infront of 'john' is important
And that one still results in a singleton vector... you may want to
add a {. to get a scalar symbol.
Martin
PS: Turning once more back to K: I am too lazy too hunt for a
reference but I seem to remember from the '94 K presentation in
Antwerp that it was a design decision right from the start to have
output data reusable as input via cut'n'paste with no ill effects.
Therefore, K not only groks `foo as symbo literal, it will also
mark singletons on output with a leading ravel comma "," -- loud
and clear. (This directly relates to another thread here.)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
