Re: Syntax question

[fjm@ti.com (Fred J. McCall)]

In article <3k7t8o$4pt@hacgate2.hac.com> dkusu@atc-1s.hac.com (David Kusuda) writes:

>In article dim@felix.seas.gwu.edu, mfeldman@seas.gwu.edu (Michael Feldman) writes:

>[snip]

>>The named association is very convenient in teaching about parameters,
>>and makes for nice readable code (IMHO, of course), but in general is
>>really "syntactic sugar." 

>I agree with Michael's points regarding named parameter association.  But the named
>parameter association is also necessary to:

>  a) Specify the parameter list in an arbitrary order--I think this is a weak point
>     as far as readability/maintainability, but still a point.

I don't know that I agree here.  Sometimes things 'read' better from the 
caller point of view if you can reorder the parameters.

>  b) Omit parameters which have default values.  I think that this is very useful,
>     too bad C++ doesn't have the same thing.

I quite agree.  The other thing I wish C++ would 'borrow' from Ada is the 
whole idea of rep specs.  There are applications where it is definitely handy 
to be able to specify exactly how many bits an object should occupy and just 
how its components should be laid out.