| Richard E Maine 2005-02-24, 4:00 pm |
| In article <ogbr11h913mk7glib9ijfub7tspfmt94iq@4ax.com>,
Dan Nagle <dannagle@verizon.net> wrote:
> The TYPELESS proposal solves several problems:
Yes. Personally, I have strongly mixed opinions on the TYPELESS
proposal. I do see that it addresses some problems. In my opinion, the
biggest negative is that it tries to address a few too many problems and
I don't like the way that all the parts mix.
This shows up starkly in what Giles calls (and I agree with him) the
oxymoronic name. Looks to me that enough stuff got tacked on to TYPELESS
that it is a type, much like any other type. Yes, it is different from
the existing types, but then you'd expect a new type to be so. I think
it is going to be really confusing to explain how something is of type
TYPELESS but that this doesn't mean it has no type.
I know that names are in some sense trivial and easy to change. But
then, I've also seen multiple cases of "placeholder names" that ended up
being the actual names just by inertia. I find the name TYPELESS to be
such a misnomer that it makes it hard for me to look past the name.
> it standardizes and rationalizes existing behavior
> of many compilers regarding BOZ constants,
Yes. I agree that's one of the good points. Incompatibilities in
BOZ treatment continue to be a big PITA.
--
Richard Maine | Good judgment comes from experience;
email: my first.last at org.domain | experience comes from bad judgment.
org: nasa, domain: gov | -- Mark Twain
|