[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
hashing
- To: rhh@MIT-VAX
- Subject: hashing
- From: Jonathan A Rees <JAR@MIT-MC>
- Date: Mon, 22 Apr 85 13:40:27 EST
- cc: SCHEME@MIT-MC
- In-reply-to: Msg of Sun 21 Apr 85 17:43:51 est from Bert Halstead <rhh at mit-vax>
Date: Sun, 21 Apr 85 17:43:51 est
From: Bert Halstead <rhh at mit-vax>
... This implies that 77 can NEVER be reused as the hash
code for some other object passed to hash-object!
This is an intentional feature of the proposal; if recycling is
permitted then the facility takes on a totally different character and
has more liimted application.
Feature:
Non-re-use means that OBJECT-UNHASH, which is desirable for many
applications of these things (like post-GC automatic file closing, weak
tables, and populations), is possible and meaningful.
Antifeatures:
(1) Non-re-use means that OBJECT-HASH is different from Maclisp MAKNUM,
and is therefore, presumably, more expensive.
(2) Non-re-use means that in very-long-lived systems, hash numbers might
go into bignums.
---
The hash/unhash facility is already optional, and I would suggest that
if for some reason you don't like its details you can just not implement
it.
If there is much disagreement now about the facility, then it should be
removed from the report. This removal should be done carefully, since
we agreed on it last fall, and agreements are precious; a different hash
facility probably won't make it into the report.
Jonathan