freenode/#sicl - IRC Chatlog
Search
4:06:27
beach
shka_: It may be an interesting article, but it seems peripheral to what we are doing at the moment.
4:06:57
beach
And, by the way, I have worked with Neil Jones some in the past. Also gave some talks to his research group.
6:33:54
beach
Neil Jones and his group did some excellent work on partial evaluation in the past. Both theoretical and practical as I recall.
7:49:20
beach
Certainly when you restrict it to a particular domain. Not that many people have worked on partial evaluation.
14:31:54
Colleen
Clhs: function ensure-generic-function http://www.lispworks.com/documentation/HyperSpec/Body/f_ensure.htm
14:33:40
Bike
honestly i think if you're looking in the standard or MOP for a coherent picture of how method combinations work you'll be disappointed
14:35:14
beach
So, in the expansion of DEFGENERIC, if the :METHOD-COMBINATION option is not given, I can omit it as an argument to ENSURE-GENERIC-FUNCTION, right?
14:36:22
Bike
ensure-generic-function is kind of a combination of make-instance and reinitialize-instance, and both of those have specified behavior for unprovided initargs. so it all seems fine
14:37:20
beach
As I recall, the defaulting behavior of this particular keyword argument is not specified.
14:38:37
Bike
sure, but i mean like, with the reinitialize-instance-like behavior you'd expect that non-specification would mean keeping the existing combination, and such