[E-Lang] Proposed Naming Convention Changes
Wed, 22 Aug 2001 14:50:55 -0700
At 01:05 PM 8/22/01 -0700, Bill Frantz wrote:
>At 10:35 AM -0700 8/22/01, Terry Stanley wrote:
>>The problem with remote (rmt) is that the reference could be local. The
>>use of far (farCar) in the Walnut conventions had the same problem --
>>since the car could be local, the use of far differed from what was
>>defined in the taxonomy.
>Perhaps it is my misunderstanding of the goal, but I think of the
>convention as saying how I must treat the name, not what the actual value
>bound to the name is. If the name is fooRmt, it indicates to me that I
>must treat it as remote, even if the current value bound to it is local.
>Or I really am confused.
No, you are not confused. This conflict is why it's difficult to find good
terms. On the one hand, we are trying to communicate how the value must be
treated. On the other hand, we're trying to avoid saying what the value is,
if what we seem to be saying isn't true.
>>>How about using "remote", abbreviated to "rmt" instead of pass. "Q" (for
>>>"Queue") might also work.
I think I like "Q". A "carQ" enables you to queue requests to be delivered
to a car. An eventual send ("<-") on a carQ does such queuing. Since an
eventual send can be done on any live reference, a variable named "carQ" can
hold any of them.