[bitc-dev] Primary type arithmetic

[Eric Northup]

On Fri, 2005-02-04 at 14:35, Jonathan S. Shapiro wrote:
> On Fri, 2005-02-04 at 14:21 -0500, Mark S. Miller wrote:
> > Jonathan S. Shapiro wrote:
> > > [...] In honor
> > > of their proposer, I am contemplating (markm+, markm-, etc.), but I will
> > > get over it.  [...]
> > 
> > "safe+", "safe-", etc would have one fewer letter. I don't know about you  ;), 
> > but I also find these names to be clearer as well.
> 
> But in the tradition of academia it is considered important to attribute
> ideas to their source... Also blame...

Shouldn't it be [miller05]+, [miller05]- then?

I don't like "safe+", "safe-", etc. because the question is not safety
of the operation, the question is a matching of programmers' expectation
and runtime behavior.  All of the proposed operators can be safe if used
correctly.

Incidentally, it just occurred to me that it can be difficult to
statically bound the memory use of a program that uses
arbitrary-precision integers.  And that arbitrary-precision +,-,*,...
operators needs to be able to throw an out-of-memory exception :-/

--Eric