[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
reactions to Will's updates
- To: scheme@mc
- Subject: reactions to Will's updates
- From: Bert Halstead <rhh@mit-vax>
- Date: Sun, 21 Apr 85 17:43:51 est
unhash-object sounds like a bad idea. In Will's scenario, if an object
(0 . 0) is passed to hash-object, and, say, 77 is returned, subsequent
calls to (unhash-object 77) must return either (0 . 0) or #!false
(presumably the latter is an out if the (0 . 0) has been garbage-
collected). This implies that 77 can NEVER be reused as the hash
code for some other object passed to hash-object! It also implies
(do we want this?) that hash-object is a one-to-one function -- it
is never permissible for more than one object to map to the same hash
code. This is not necessary if hash-object is just a way to compute
an index into some user-maintained hash table, where objects could
be associated with whatever information is desired.
re numbers: Is anything substantial gained by disallowing explicit
"+" signs on the exponents of numbers in scientific notation? -Bert