[Checkers] romaybe => ropoly?
jaimeq at MIT.EDU
Sun Apr 6 14:13:52 EDT 2008
I agree that "readonly polymorphism" is better than "readonly maybe."
However, the abbreviation seems odd. When I see ropoly, I want to
pronounce it "rope-oly." That word ropoly used two different forms of
abbreviation: ReadOnly -> ro and POLYmorphism -> poly. I don't see
the need to use all of readonly. For example, it could be abbreviated
to read, in which case the new keyword could be "readpoly." This also
separates all words clearly, since read and poly are read separately
(no pun intended).
This possibility made me realize that "polymorphism over readonly"
isn't the phrase we've been saying, but rather, "polymorphism over
mutability." Therefore, the keywords "mutpoly" or "polymut" are
tempting, but aren't as nice to see as "readpoly" or "polyread."
On Apr 6, 2008, at 1:11 PM, Matt Papi wrote:
> Personally, I think I'd want the "poly" part to stand out over the
> "ro" part, so I'd prefer something like "polyro". (Either way, though,
> I think "poly" is superior to "maybe".)
> - Matt
> On Sat, Apr 5, 2008 at 11:19 AM, Michael Ernst
> <mernst at csail.mit.edu> wrote:
>> Jaime and Telmo-
>> The name "romaybe" was intended to intuitively indicate that the
>> argument might be either mutable or readonly (it's "maybe" readonly).
>> However, I'm not sure that the English word "maybe" clearly
>> indicates this
>> meaning. In papers and documentation, we explain romaybe by
>> that it offers parametric polymorphism over mutability. Thus, I
>> have been
>> thinking that perhaps we should rename the keyword from "romaybe" to
>> "ropoly". (And if a similar mechanism were introduced in, for
>> example, the
>> nonnull checker, it would be "nnpoly" instead of "nnmaybe".) What
>> do you
>> checkers mailing list
>> checkers at lists.csail.mit.edu
More information about the checkers