> You're thinking of objects as references and references as akin to
> pointers, which makes sense because that's how they're implemented in Perl
> 5. If you think of objects as their own entities, however, or think of
> references as something other than pointers (in particular, something that
> doesn't require explicit dereferencing), then using . to access object
> members is entirely compatible with C.
> I tried to make this point before, but I don't think people understood
> what I was getting at.
I did. :) Moreover, the idea of a "dereference operator" dumbfounds lots
of folks. "What's an object got to do with a reference, much less a
pointer?" A p5 object is very confusing to others for this reason, and so
is the syntax.
While I agree with some that we should keep Perl folk happy before making
the world happy, this is one syntax quirk which really doesn't help Perl
IMO. When method invocation has such a generally-accepted syntax
everywhere else, it can only be positive to adopt a similar syntax.