[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Reading the Readtable

    Date: Saturday, 9 January 1982, 14:20-EST
    From: Daniel L. Weinreb <dlw at MIT-AI>

    One problem with the existing feature is that there is no way to ask a
    readtable what is in it.  I admit that I don't know why you would want
    this, but you might.

I know that this is a "problem" with the existing "feature".  To be
able to return ssomething from (STATUS SYNTAX ...) that is USEFUL
(i.e. you can save the number someplace and then us it later in a
DIFFERENT readtable) is IMPOSSIBLE in general without re-implementing
the reader from the ground up and re-introducing various magic
meaningful "syntax bits" ala MacLisp.  I repeat what I said before:
"There is NOTHING MEANINGFUL I could return from (STATUS SYNTAX ...)"

    Also, Alan said:
        I agree that having a symbolic name for a usefull syntax is a good
        idea, but I cannot possibly think up names for all of them.  Remember
        that A is different from E is different from S etc.  Do you have a
        specific syntax that you think deserves a name?

    in reply to Henry's:
        Ideally, it would returns keywords as acceptable to 

    So maybe Henry is just asking for something that returns the "description"
    without returning that information which differs between A and S.

I don't think I understand at all what it is that you think you are
asking for here.  If I am allowed to ignore the difference between "A"
and "S", then how about the difference between "A" and "(", can I
ignore that too?  Why don't I just always return NIL then?  Do you
propose that:

(get-syntax-description #/A) => ALPHABETIC


(get-syntax-description #/S) => ALPHABETIC

??  What happens if you were actually interested in the fact that "S"
can be used in small flonums?  How am I supposed to decide this?  What
about the difference between "+" and "-", do we judge that to be

    I'd also be interested in knowing what the application is, though; maybe
    there's a better solution that we should be implementing.

Thats what I said.  Before we get off on flaming about all kind of
difficult/impossible ideas that I get to implement in my spare time,
why don't we find out what the application is.  Not for a "better
solution", but for a "POSSIBLE solution".