02:13:36 [QUIT] ace quit: . 04:30:25 abi joined #tunes 04:34:54 karltk joined #tunes 05:16:29 [QUIT] ree quit: Ping timeout for ree[twisted.goodnet.com] 05:24:51 [QUIT] serge_ quit: [x]chat 06:21:25 arigo joined #tunes 06:27:46 eihrul joined #tunes 06:29:01 [QUIT] Downix quit: Ping timeout for Downix[dc-57-37.ici.net] 06:31:43 Hi Eihrul ! 06:32:15 I am still looking for info about how Slate's MO specify the graph reduction rules... Any pointer ? 06:35:26 unfortunately... 06:35:44 slate does not specify that at all yet :) 06:35:54 Ahh... 06:35:57 as i inferred in the mail, that's one of the highly variable aspects of slate 06:36:04 just about any system would do 06:36:16 so is left open ATM :) 06:36:21 So how could Slate's current state be summarized : 06:36:32 A generic linear graph reduction model ? 06:36:37 no 06:36:40 With rules specified somehow inside the graph ? 06:36:49 graph reduction is an analogy 06:36:53 Ok 06:36:56 it isn't really slate :) 06:37:36 graph reduction in and of itself isn't necessarily great IMO, its just a model :) 06:38:33 Ok; but is the description of Slate at slate.tunes.org still up-to-date ? More or less ? 06:38:49 not afaik 06:39:08 not really sure on the status of slate anymore 06:39:17 i have lots of annoying, stupid school stuff 06:39:23 water has to deal with navy 06:39:25 etc. 06:41:16 How do you see Slate like now ? What important concepts have been pointed out in recent discussions ? 06:43:44 well, what are you asking for here? :) 06:45:11 The old Slate docs I found do not reflect the discussions that were just mentionned in the mailing list archives; 06:45:36 then I am asking, how Slate has evolved... 06:46:44 well, actually, i suppose the online documentation is ok enough for now :) 06:48:44 I was thinking e.g. about the Pigdin framework. 06:48:57 Sorry Pidgin 06:50:09 well, Pidgin was just going to be Selfish with the MOP 06:51:48 nothing very revolutionary or special 06:52:05 but actually... 06:52:05 Ok, thanks. Just a first try I guess. 06:52:15 some things we were discussing wrt Slate that aren't covered were: 06:52:42 we had changed the syntax of slate to be "concatenative", i.e. not hiercarchical but just a stream of tokens like Forth 06:53:10 we had discussed the prospect of giving access to more parts of the environment (like the evaluation stack) to the MOP messages 06:53:44 we also discussed Slate as a meta-language with a framework for syntactical extension that could be used to implement other languages over Slate 06:54:16 also, Brian was looking into developing a type system for Slate or some such 06:55:16 [QUIT] karltk quit: new kernel 06:55:24 Thanks for the update ! 06:57:53 I'm also trying to see how my project "Bazar" relates to all this. 06:58:23 I would now define it as a graph reduction system with a single rule specified inside each "operator" node. 06:58:44 i'd say what you have is basically a lisp 06:58:47 This is somewhat elegant but I now feel it is much too restrictive 06:59:09 Yes, it looks like Lisp, I agree... 06:59:17 like i said, graph reduction is a model for looking at stuff, i don't really think it always necessarily deserves to be a language in its own right :) 06:59:41 Well, I rather do... 07:00:09 but what makes Bazar better than a lisp? 07:00:16 I think the model is impressive in its ability to have concepts such as execution and parallelism emerge automatically 07:00:40 eihrul shrugs. 07:00:41 The difference between Bazar and Lisp is essentially that Bazar uses graph reduction and not execution. 07:00:49 parallelism isn't inherent to graphs at all 07:00:52 i think you're munging issues 07:01:01 Also, not mentionned in the web pages, it's typed. 07:01:17 how can you call eagar evaluation graph reduction? 07:01:43 because you "graph reduction" model seems to be exactly lisp's evaluation model and that of many other eagar languages for that matter 07:02:56 Am I totally off if I believe that "7" and "(+ 3 4)" are fundamentally different for Lisp ? 07:03:23 In the sense that "(+ 3 4)" need to be applied a special operation, evaluation, before it is "7" ? 07:04:40 Fare joined #tunes 07:06:05 Hi Faré. 07:06:51 [QUIT] gREMLiNs quit: I am one who is many 07:07:48 eihrul: you can answer "yes" :-) 07:07:49 arigo yes 07:07:51 7 is atomic 07:08:09 (+ 3 4) is in no way 7 -- it's a s-exp 07:08:21 yop 07:08:33 well, (+ 3 4) evaluates to 7 07:08:55 but it isn't the same as 7, if you consider it at source level 07:09:05 It's a matter of point of view 07:09:07 at machine level 07:09:12 Fare no 07:09:13 and that matter IS important 07:09:16 Yes, I know; I was trying to see what's the difference in a graph reduction model. In such a model, (+ 3 4) is just 7; no evaluation needed. 07:09:18 '(+ 3 4) != '7 07:09:37 arigo: _if_ you consider the expression "up to reduction", indeed. 07:10:02 but even then, you mightn't want to consider it that way. 07:10:06 (+ 3 4) is a list that when evaluated returns 7, it is not 7. 07:10:08 Yes, sure; I am trying to push for languages in which everything is "up to reduction" anyway (still influenced by graph reduction systems) 07:10:32 ult grumbles 07:10:43 arigo: reflection "up to reduction" is not always a good idea. 07:10:55 Fare are you experienced with Java's reflection? 07:11:00 it can't be deterministic 07:11:02 Fare: Ah ! I guess that's an important point. 07:12:10 and considering that people may want to consider different reduction systems on the "same" terms, implicit reduction models are not as simple as you might believe. 07:12:18 (when you want to do real stuff with it) 07:12:27 ult: not much, sorry. 07:14:40 Fare: for example if some data needs to be considered from various points of view ? Or also at a more "programmatic" level ? 07:19:24 arigo: yup 07:20:09 one way to consider things: tag graph data with a head that specifies the evaluation context 07:20:35 headless data is raw, or subject only to "universal" implicit evaluation rules. 07:21:08 (the latter option raises the question of what rules are useful as universal, if any) 07:22:03 of course there will ALWAYS be implicit rules, those that define the "current" system. 07:24:14 So you would rather evaluate data inside a some-how-specified "evaluation context"; 07:25:55 Do you still wish to have a homo-iconic representation of data at some "primitive" level, independently of the evaluation context ? 07:27:56 Like "everything is a car-cdr cell", and the "tag" on this data specifies how it should be considered ? 09:08:32 Guest1749 joined #tunes 09:08:39 hello fare 09:09:11 Is there someone here ? 09:10:44 no 09:14:45 [QUIT] gREMLiNs quit: I am one who is many 09:52:14 eihrul joined #tunes 10:01:54 [QUIT] JALH quit: Killed (NickServ (GHOST command used by JALH_)) 10:36:37 karltk joined #tunes 12:32:11 [QUIT] ult quit: Ping timeout for ult[149.149.201.30] 12:38:53 [QUIT] ree quit: Read error to ree[twisted.goodnet.com]: EOF from client 12:38:56 [QUIT] gREMLiNs quit: I am one who is many 14:19:19 ult joined #tunes 14:34:27 Fare joined #tunes 15:22:35 BC_V joined #tunes 15:52:18 BC_V left #tunes 17:47:59 lar1 joined #tunes 21:49:28 arigo joined #tunes 22:53:17 [QUIT] Melinda quit: Life's a whore and I'm broke. 23:39:11 ult joined #tunes 23:39:26 anyone here? 23:39:39 I 23:39:50 (Don't count) 23:40:42 hey lar do you know the area of a regular hexagon offhand? 23:41:36 No, sorry 23:41:41 [QUIT] ult quit: Leaving 00:00:52 ult joined #tunes