3:06:56Bikedoes anyone know how slime parses arglists? if i do (defmethod test (&key ((:foo bar))) bar), (swank-backend:arglist 'test) => (&KEY ((:FOO BAR))), but it displays in the minibuffer as &key foo
5:42:26beachI always try to solve some practical problem. To do that, I have to invent some theory or some technique that needs to be proven to work. Then I use that theory or technique to create an implementation. Finally, I measure performance (often compared to existing techniques), and show that the new technique is better than existing ones.
5:44:05beachThis formula works well for Common Lisp, because very little serious fundamental work has been done since the AI winter, and the reality of processors has changed radically since, so new techniques can be invented.
5:45:24SaganManThat's good beach. Those new techniques and theries of yours are patented?
5:46:29beachNo. I am fundamentally against software patents. And I was part of the movement to get the EU parliament to reject software patents by a crushing majority.
5:51:26beach... so instead they are published in the proceedings of ELS and ILC.
13:05:51ColleenCommand Syntax: roll &OPTIONAL (SIZE 6) (TIMES 1)
Documentation: Roll some dice. Note that this is not provided with the intention of providing gambling means.
13:18:42ShinmeraNote that you can ask Colleen for the time as it should be for /users/, not /places/, but in order to do so the person needs a profile.
14:32:26Xofincidentally, on the question of "why would anyone want to subclass method combination anyway?", I'm not sure! But I do wonder whether a more declarative, less arbitrary-code thing might be able to allow compute-effective-method to do some kind of smart caching
14:32:51Xofor rather, the class of the method combination could inform compute-effective-method whether smart caching was allowed or not
14:34:17Bikecompute-effective-method already turns a second value of extra information, it could include a cachep specification