[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Accessing dynamic slots using "slot-value"
- To: Gregor.pa@Xerox.COM
- Subject: Re: Accessing dynamic slots using "slot-value"
- From: macgreg@vaxa.isi.edu (Robert MacGregor)
- Date: 31 Mar 87 15:49 PST (Tuesday)
- Cc: Gregor.pa@Xerox.COM, commonloops.pa@Xerox.COM, macgreg@vaxa.isi.edu, RPG@SAIL.STANFORD.EDU, DLW@ALDERAAN.SCRC.Symbolics.COM, kempf%hplabsc@hplabs.HP.COM
- In-reply-to: Your message of 31 Mar 87 10:38 PST. <870331-104035-2584@Xerox>
Gregor,
You hit the slot issue exactly. Our application requires that
"slot-value" (or else a kludged variant of it) work with our envisioned
use of dynamic slots. We don't like kludges any better than the next
person. However, since our architecture expressly forbids anything
resembling "undeclared" slots, the problem with "with"s that you brought
up isn't an issue. Although there may be hundreds of possible slots
for some of our knowledge base objects, all such slots will be
pre-declared.
It would be quite inefficient for us to use a special function to access
dynamic slots, since an "optional" slot which has a :dynamic allocation
high up in a class lattice may become "mandatory" and switch to an
:instance allocation farther down in the class lattice. We want the
choice of storage allocation to be transparent to the slot access calls.
Putting :dynamic slots into a meta-class written on top of CLOS would be
more than satisfactory.