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


The only problem that could be solved by filling in this hole is to
give people a way to supply fewer than all of the optional positionals
while supplying some of the named arguments. My suggestion doesn't help
with that much.

As I've mentioned before - but as I know I must mention again - the effect
of this might have some relevance to CLOS in which it might be useful (the
CLOS folks have not yet decided) to discriminate on the names of named 

For example, writing this

(defmethod f ((x c1) (y c2) &required-key foo bar) <code1>)
(defmethod f ((x c1) (y c2) &required-key baz ola) <code2>)

is simpler to understand than this

(defmethod f ((x c1) (y c2) 
              &key (foo nil foo-p) (bar nil bar-p) 
	           (baz nil baz-p) (ola nil ola-p))
 (cond ((and foo-p bar-p) <code1>)
       ((and baz-p ola-p) <code2>)))

And when one writes instance initialization code in CLOS, you will be
writing lots of code like this.