[Zope3-dev] Re: Merging schema & interfaces (was: "Sub interface")

[Casey Duncan]

On Tue, 02 Mar 2004 11:54:24 -0500
Gary Poster <gary at modernsongs.com> wrote:

> 
> 
> Casey Duncan wrote:
> 
> > Combinatorial...complexity...must...resist...
> > 
> > all-the-worlds-a-functional-call-ly yr's,
> 
> ;-)  What I wrote are conveniences.  What I wrote could probably be 
> slightly changed to support a variant that accepts a lambda as an 
> argument.
> 
> I think that some basic invariants are useful as individual functions,
> 
> since it aids readability to the people I've shown the interfaces to.

Maybe I'm just wary of throngs of "convenience" functions that somebody
need to know about when a simple comparison is all they want. In most
cases I think what I will want will be different enough from the
provided functions that they won't be that useful.
 
> I'd argue that there is a line to be drawn along a continuum between 
> "everything's a lambda" and "we have a huge library of barely-used, 
> simple invariants", rather than a "simple helper invariants are bad" 
> rule.  YMMV, of course. :-)

I guess I'm in the "simple helpers are bad" camp in this case because it
actually hurts readability IMO and I have to add yet more imports just
to do a simple one operation comparison.

A single general helper for dealing with the message ids is fine by me,
however (not that anyone asked).

-Casey