[plt-scheme] Re: to define, or to let

From: Anton van Straaten (anton at appsolutions.com)
Date: Sat Apr 10 16:42:31 EDT 2004

Paul Schlie wrote:
> > ... The alleged advantage to programmers of a fixed order of
> > evaluation is that even if they make a mistake and have an
> > order dependency that they didn't intend, the program will
> > still behave consistently.  If that consistent behavior
> > seems to conform to the program's desired behavior, then
> > the order dependency may be allowed to remain, unrecognized.
> > This is the real world consequence of the program having
> > an unambiguous semantics - it's good because you get consistent
> > behavior, but it's not all good: unfortunately, such a situation
> > pretty much qualifies as a latent bug, or at best, fragility
> > in the program. ... evaluation order the program has a
> > well-defined behavior, even if it achieves that behavior
> > partly "by accident".  Semantics is not usually concerned
> > with the intent of the programmer, only with what the
> > programmer actually ended up encoding in the program....
> Sorry Anton, but you really shouldn't try to pass this crap off
> as fact, and end up confusing folks who don't know any better.

Paul, in our offline discussions you've been full of unsupported assertions
too.  If you want to engage me publicly or privately, please do so with
something concrete.


Posted on the users mailing list.