03:17:53 eihrul joined #tunes 03:47:30 [QUIT] eihrul quit: [x]chat 08:56:27 smklsmkl joined #tunes 08:58:39 [QUIT] smklsmkl quit: No windows for this server 09:00:02 [QUIT] downix quit: changing servers 09:05:25 smklsmkl joined #tunes 09:34:54 ult joined #tunes 09:39:32 [QUIT] ult quit: Leaving 10:43:44 [QUIT] MysticOne quit: Leaving 11:01:22 [QUIT] smklsmkl quit: Ping timeout for smklsmkl[ppp14.dial-in.verkkotieto.fi] 11:03:02 smklsmkl joined #tunes 11:19:58 [QUIT] downix quit: Ping timeout for downix[d-ma-boston-31.ici.net] 11:20:55 Kyle_L joined #tunes 11:30:55 who's here? 11:34:07 smklsmkl is. 11:35:09 I was thinking about generics in Modula 3 (templates in C++ or polymorphic types in Haskel) 11:35:58 They appear to be types, parameterize by types. 11:36:03 (2nd order types) 11:36:29 Nothing really interesting until I thought about field definitions. 11:37:09 yeah, haskell and ML types are second order i think 11:37:14 To be clear I will say fields are slots in objects. 11:37:37 Fields are templates too! 11:37:53 They are parameterized by the type they accept as domain 11:39:08 object types are sort of recurive 11:39:56 i'm not sure if a method can be separated from an object 11:40:05 The recursivness you speak of has not come clear to me yet. 11:40:17 What do you mean by method? 11:40:48 I think of functions as objects. 11:41:06 (but not nessesarily objects as functions) 11:41:13 a function of an object 11:41:28 (member function in c++) 11:41:58 I would imagine that if you refer to a method of an object, then that reference is an object, and so is what you refer to. 11:42:51 A method appears to me as a map between a message type and a specific function based of object class. 11:43:20 Method = (function, class, messagetype) tuples 11:44:34 I like to think of a message as intent, function as result, and a method as the mapping (not nessesarily accurate if humans are involved) 11:44:35 anyway, a type of an object would be like: class b { int c (int i) { ... } } has type b = < c : 'a -> int -> int > as 'a 11:45:24 MMmmmm. what lang was that? 11:45:53 the code is c++, but the type is ocaml style 11:47:38 so the type of the methos depends on the type of the class and vice versa 11:47:40 Maybe you can expand that for me please. 11:47:46 s/methos/method 11:54:32 specifically I am not sure what b = < c means. 11:55:18 b is the type of the object, between < and > are it's methods 11:55:31 td joined #tunes 11:55:39 c is a method with type 'a -> int -> int 11:55:58 How is the typing recursive? 11:56:11 "as" makes 'a abbreviation of < c : 'a -> int -> int > 11:56:51 the type of the method depends on the type of the class and vice versa 11:57:33 doesn't that make it recursive? 11:58:22 So I am clear. You are saying that the method definition is dependent on the class definition, and the class definition dependent on the method definition? 11:58:55 yes ... 11:59:34 That appears to be unnessesarily confusing things. 12:00:04 why? 12:00:19 I see no advantage to a recursive dependency than just have the def of class depend on def of methods, period. 12:01:32 well, in OO methods usually depend on the classes 12:02:31 Forgive my naivete, but in what way? 12:02:43 (recursion can always be removed) 12:03:23 with 'self' or 'this' variables, (most of the time implicit) 12:03:44 actually that is depending on object 12:04:06 but there is not much difference 12:05:02 You are saying that the action taken by an executed method depends on the object it acts on. That is different than the defintion of the function depending on the class of the object. 12:06:29 Recursion is useful for reducing programmer workload. It is not useful to understanding (not to me anyway). 12:07:42 anyway, the type of a method depends on the type of the object or class it is defined relative to 12:10:48 Maybe I am unclear of the term "type of a method". I am interpreting it as its defintion. Am I right? 12:11:22 [QUIT] kc5tja quit: THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK 12:12:33 Suppose we have two classes, A and B with methods, cA and cB defined differently on each respectivly (handling the same message/parameters) then you are saying they 12:13:11 then you are saying that cA and cB are the same (c) but is a paramterized function based on class? 12:13:25 cA=c(A) cB=c(B) ? 12:14:01 they have the same structural type 12:14:11 classes are not types 12:15:19 (IMO) 12:15:34 OK, then "structural type" is unclear. Does that mean "have same interface?" 12:16:07 yes 12:17:12 I will have to think about that unusual view of classes/methods for a while. 12:17:41 Thanks. 12:19:31 btw I have read that classes are not types, but the distinction always eludes me. May you give me an example? 12:22:49 well, in structural object typing, in your previous example there are two classes, but both of them have the same type 12:23:42 class basically tells you how to construct objects, types tells how you are allowed to use them 12:24:18 Oh! OK. I have always considered similar defintions to be different types. 12:24:55 they have different types in java and c++ 12:26:00 but the type systems of java or c++ aren't very reliable :/ 12:26:49 I have been calling types interfaces. 12:30:24 OO typing is a very confusing area ... all kinds of problems with co- and contravariance 12:31:43 smkl: thanks for your help. 12:32:24 this means that if you inherit from a class, the type of the subclass might not be a subtype of the type of the parents class ie. it might cause a run time error if you try to use an instance from child class like instances of it's parent 13:02:32 [QUIT] rares quit: Ping timeout for rares[dialin-10.hartford.bestweb.net] 13:33:52 bineng joined #tunes 15:42:00 [NICK] lar1 changed nick to: lar-movie 15:45:28 [NICK] kc-crunch changed nick to: kc5tja 17:41:17 [QUIT] gREMLiNs quit: BitchX WWW Site -- http://we.got.net/~brian/ 17:54:34 [QUIT] Kyle_L quit: Ping timeout for Kyle_L[cr168790-a.nmkt1.on.wave.home.com] 18:20:36 [NICK] kc-errands changed nick to: kc5tja 18:26:40 morton joined #tunes 18:26:50 morton left #tunes 18:30:22 [QUIT] I440r quit: [x]chat 18:30:25 I440r joined #tunes 18:37:28 [QUIT] kc5tja quit: THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK 18:45:38 rares joined #tunes 18:45:47 water joined #tunes 18:45:52 hey water 18:46:00 hey 18:46:33 hi water 18:46:44 water has been reading lots of fpsood in spare moments today 18:47:25 so you recommend that book? 18:47:35 fpsood? 18:47:39 hm 18:47:55 functional prog'ing sood? 18:47:58 "a functional pattern system for OO design" 18:48:16 it depends on your background and goals 18:48:39 for example, i've read nearly all of the bibliography already 18:48:43 seemed interesting 18:48:49 oh 18:49:31 it basically explains various ways of doing fp things in oo languages... one big disadvantage (imo) is that they use eiffel for all the examples 18:50:11 yeah 18:51:12 i'm having a hard time getting into programming again 18:51:17 hm i keep missing people who rarely show up 18:51:24 like bineng 18:51:45 heh. yeah he left while back 18:52:51 rares: what's this you just gave me? 18:53:19 the fs manager i'm designing (for eons now :/ ) 18:53:27 heh 18:53:30 just a couple of ideas 18:53:35 sounds like slate :) 18:53:55 you read it already ? 18:53:57 ) 18:54:02 yeah 18:54:28 how long in this case is eons? 18:55:42 1 year of what ifs ( transition from windows to linux threw me off balance) then on and off and now I'm quite clear on it and it's taken a few eeks 18:56:02 ok 18:56:18 but it's moving faster now 18:56:51 I'm close to where it's just dirty work 18:57:24 My first release will be a definiton for Mandrake 7.0 18:57:31 i assume you could just write a driver for it for linux (from looking at the pseudo-requirements here) 18:57:40 heh 18:57:49 water i'm close to that 18:58:12 I'd love to be as transparent as possible 18:59:14 incompatibility will cease to have meaning as far as fses are concerned 18:59:26 maybe 19:00:05 reflection is fundamental 19:00:07 ) 19:00:07 it depends on how your reflective architecture works i think 19:00:31 you hardly specify anything about it so far 19:00:48 that's next 19:01:12 in fact I should skip the definitions for now and design that next 19:01:24 perhaps 19:01:44 Infinity you're ass is mine 19:01:46 maybe you should look a little at how self has mirror primitives 19:02:03 there is no mirror :) 19:02:11 P 19:02:26 it's some powm I read 19:02:35 you need mirrors for reflection... it's very similar to quoting 19:02:41 one pupil goes up to be tested 19:03:00 talks about the mirror of the mind 19:03:13 the clarity of the wind of thought 19:03:18 blah blah 19:03:41 then an old man comes to the ceremony and he says: 19:03:49 there is no mirror 19:04:18 he gets the seat or trophy or whatever 19:05:11 the problem i've had with most papers on the subject is that related notes are pages away from each other 19:06:03 well, just for reference, check the self programmer's reference (around page 130 iirc) 19:07:04 oops 19:07:13 pg. 40 19:07:26 1+3 is close enough 19:07:38 will slate know what close enough means? 19:07:43 heh i was thinking of another place i bookmarked 19:08:01 what do you mean by "close enough"? 19:09:15 like when it only needs to find a general object 19:09:27 if A contains B 19:09:57 and B can do Z 19:09:59 i think i should plead the fifth and just say "it can be added" :) 19:10:18 it might just be nough to tell A to do Z 19:10:29 oic 19:11:01 you could alter the meta-object of A to do searching like that, yes 19:11:25 that's gnu fs will do 19:11:36 I'm going to have a visibility setting 19:11:42 but it's not something a system should allow in general without using a different code architecture 19:12:10 that way it frequently reads the network neighborhood ahead but only so far 19:12:17 (iow to make containership mean something else like graph-connectivity in a network) 19:12:23 right 19:12:44 this is one of the reasons eihrul and i are so into meta-objects 19:12:44 so far = save bandwidth cpu resource 19:13:10 i'm into meta objects because IRL i'm way out disorganized 19:13:26 and i'm PARANOID about it 19:13:28 eihrul joined #tunes 19:13:33 heh 19:13:36 heh 19:13:40 speak of the devil 19:13:48 re eih 19:15:52 rares wonders about creating a threaded irc system 19:16:05 that would be a challenege 19:16:35 people don't use cool systems that are flexible like smalltalk :) 19:16:50 well i'm not a people 19:17:43 i also want to hijack the whois system 19:17:48 I've had it with NSI 19:17:57 but designwise 19:18:02 anyway 19:18:09 not actual malice 19:18:14 anyway is right 19:18:17 (this could go for days, you know) 19:18:30 yep so I'm glad i'm cutting it off 19:19:45 see 19:22:42 Batterij joined #tunes 19:23:04 <--- XeF4 19:23:09 ah 19:23:18 re 19:23:23 rere 19:23:28 mircy be on you 19:28:29 Tril joined #tunes 19:28:29 [MODE] ChanServ set mode: +o Tril 19:28:59 hey tril 19:29:30 hello..what's this about arrow.cx on bespin? 19:29:55 well arrow.cx is currently only hosted from corey's eirene server 19:30:37 or rather is currently *not* hosted anywhere 19:30:57 If he's gone, how will you get the files to move them over? 19:31:01 which bothers me for obvious reasons 19:31:37 i can't, afaik, unless there's a diktuon mirror of which i'm unaware 19:32:04 at the very least, my confidence in him is rather shaky now, to say the least 19:32:13 two leasts, even :) 19:33:17 abi joined #tunes 19:33:46 well, if hcf ain't here, I'll run his bot :) 19:33:52 thanks 19:34:01 i definitely missed abi 19:34:42 hm 19:35:02 so anyway you should know i'm working on the arrow squeak code again 19:35:37 and i'm developing the architecture using some books i recently have acquired 19:37:11 all right 19:37:41 I'm planning to write some lisp code 19:38:23 I'm thinking I might program some other things too 19:45:57 ok 19:46:56 eihrul's still the main compiler fellow for tunes as far as i'm concerned 19:47:26 most of the work i'm doing is to research ideas 19:47:46 we certainly need that! 19:47:57 heh 19:48:10 my research book-buying habit is getting pretty expensive 19:48:42 it's a good thing i have very few month2month expenses 19:49:06 water: ditto, but I seem not to buy other things like.. food 19:49:14 heh 19:50:43 abi: fpsood is A Functional Pattern System for Object-Oriented Design at http://www-agce.informatik.uni-kl.de/~kuehne/fps/ 19:50:51 I _should_ and I do _not_ feel well after eating ramen noodles for ~2weeks.. 19:54:07 you can't get nutrition off ramen alone :) 19:56:26 eihrul: no, but it's convenient since a coworker leaves n+1 cases of ramen in the kitchen and I am living at work until next week (->#offtopic) 19:58:07 Kyle_L joined #tunes 19:58:46 hey kyle 19:59:04 hello water, how are you? 19:59:10 databases suck!! 19:59:17 ) 19:59:33 does (->#offtopic) get actually heeded by log bots? 19:59:43 no 20:00:06 why databases suck 20:00:27 is that a question? 20:00:37 Why do databases suck? 20:00:57 heh 20:01:15 Is that a comment about Diktuon? 20:01:28 not specifically 20:01:43 since i do db stuff at work now 20:01:44 well I think databases suck 20:01:53 oh right. 20:02:00 or badly conceived databases 20:02:10 I don't know any Access. I use PostgreSQL 20:02:16 i spend days doing things that smalltalk would let me do in minutes 20:02:20 come on Oracle for gEdit? 20:02:46 [QUIT] eihrul quit: [x]chat 20:02:59 he'll be back in an hour or so 20:03:32 i think slate design has definitely settled down 20:04:17 everything's been reduced to a MOP design that gets more and more specific each time we discuss it 20:04:24 cool 20:04:33 or rather, a MOP system 20:04:59 MOP? 20:05:08 and i'm currently doing some research on MOPs to allow us to do what forthers call meta-compiling :) 20:05:14 abi mop? 20:05:14 mop is Meta-Object Protocol 20:05:19 thanks 20:06:12 in fact, it seems the easiest language to implement this way is a Joy-like language 20:06:24 or maybe just a combinator language 20:07:07 in laymen's terms, this means actually reading in other languages' source files as Slate programs :) 20:08:02 a must if you are to take advantage of all that has been written so far, even if it does not contain all the specification you like. 20:08:24 not really 20:08:35 but it helps 20:08:46 Where are you going to get the word editors and the low level drivers? 20:08:54 the problem is that languages have some hidden dirtiness 20:09:03 Yes 20:09:15 But, like you said, it helps. 20:09:39 implementing the meta-compilers for C might not actually be worth the effort 20:09:56 but who knows, it's worth a try 20:10:38 at any rate, i have to concentrate on this arrow/slate experimental code 20:21:08 heh i think i've finally found something that would get me to use CVS :) 20:21:32 a Squeak CVS client with a graphical interactive interface :) 20:21:33 what could that be? 20:21:46 still being developed :/ 20:22:43 although editing the tunes pages with it would still not be a piece of cake 20:23:30 water: do you also need HTML WYSIWYG with that? I bet squeak doesnt have one of those. 20:23:41 (i'm not trying to be condescending ... cvs is annoying to learn) 20:24:05 squeak actually can handle a bit of html wysiwyg 20:24:24 are there any games in squeak? 20:24:26 only the basic stuff, though 20:24:29 yep 20:25:14 tetris, chinese checkers, a card game engine, samegame, hangman, ... 20:26:02 network multiplayer? 20:26:08 hm 20:26:24 someone is working on a ychess(?) interface 20:26:34 not too important if the games have it.. but I assume networking can be easily done. 20:26:37 otherwise you have to pull some tricks to get that to work 20:27:18 I haven't looked at the SQUEAK VM spec - would it be too difficult to write a gcc machine description for? 20:27:45 s/QUEAK/queak/ 20:27:54 the blue book spec is online 20:27:57 gcc uses machine descriptions? That is more advanced than I thought they would have. 20:28:29 they're a bitch to write though 20:28:35 and the squeak vm is written in smalltalk and translated to ansi C before being compiled for the platform 20:29:07 waterSqueak joined #tunes 20:29:21 yeah it handles networking pretty well 20:29:43 PWS and comanche are web server software for Squeak 20:30:46 waterSqueak left #tunes 20:36:12 hm hm hm 20:36:30 the tunes choice construct 20:38:45 [QUIT] Kyle_L quit: Leaving 20:40:27 Tril: is my military e-mail address still allowed to post to the Tunes mlist? 20:41:47 well... it would be... but the new mlist software doesn't support it yet. I'm trying to get Fare to write it. So for now I'd have to approve posts from non subscribers (which I dont mind doing and I do check my e-mail quite often) 20:42:11 um i'd say that Fare is too busy with other things atm 20:42:29 well, he asked if he could help so I told him how :) 20:42:33 unless it's a trivial script 20:42:39 oh 20:43:25 the old mlist software had security holes so we can't use it. 20:43:41 when will the new mail archives get onto the web? 20:44:19 i dont know, but I do know who I need to ask to find out :) 20:44:24 shall I send him a message? 20:44:29 yes 20:44:31 who? 20:45:01 i was going to mail the Debian package maintainer of the mailing list software, Sympa. The current package version doesn't have WWSympa support in it, which is what you want. 20:45:26 The LATEST version of Sympa has it, but there's no package. So we just need to ask if he will package it. 20:45:55 ok 20:48:00 do you want to meet irl again to discuss ideas before you get back to school? 20:48:40 eihrul joined #tunes 20:48:47 re 20:49:23 water: me? 20:49:26 Tril: you 20:49:27 oh right 20:49:37 i had to think about the reference "irl" 20:50:11 sure, we can 20:50:28 ok 20:50:37 just pick a day on a weekend sometime 20:50:46 I don't necessarily have anything to bring up but I'm willing 20:51:04 i just need to discuss ideas in person with someone 20:51:14 maybe I can find a time to visit with you and my friend who is at Microsoft 20:51:20 consecutively 20:51:23 ok 20:51:33 he's interning there this summer and I wanted to get a tour :) 20:51:50 hm 20:52:02 can you get me in, too? :) 20:52:54 internship is ok but why would anyone work there? 20:53:19 $$$ 20:53:22 let's see. He's a computer science student. He failed the interview last year. He learned more, and passed an interview this year. So, that's all I know about how to help you get there. 20:53:35 both in salary, stock options, and project funding 20:53:41 s/both// 20:53:46 $$$ from a company that can't produce a product without killing off the wholke lot 20:53:55 He is working there to get experience in the field. Working as a programmer at a computer company. 20:54:08 sorry but that's your future son's business you're fucking with :) 20:54:08 oh i don't want to *work* there 20:54:16 oh, you wanted to come 20:54:26 yeah just to look around 20:54:34 rares left #tunes 20:54:37 do you want to steal some secrets from their research division? :) 20:54:38 i haven't made my way over there yet 20:54:43 lol 20:54:57 for instance the Intentional programming project maybe? 20:55:01 those guys e-mailed me last year BTW. 20:55:11 oh yeah, they're obviously way ahead of Project Oxygen in terms of .NET research ;) 20:55:28 intentional programming is NOT that great 20:55:39 They saw a reference to IP in the tunes mailing list logs (that I posted) and wanted to know what relationship I had to IP 20:55:50 it's the middle-manager's way of looking at C/C++ code 20:56:05 yeah you've mentioned that before 20:56:08 I think they wanted to make sure I wasnt stealing their idea? 20:56:18 how can you steal "ideas"? 20:56:19 most likely 20:56:40 eihrul likes the "ideas are discovered" vs. the "ideas are created" paradigm. 20:56:52 so do i 20:57:07 but then i think of ideas as people more than i do of humans :) 20:57:29 there is already a wishlist bug for sympa asking for WWsympa to be packaged 20:57:31 23 days old 20:57:49 I'll try to add a "Me Too" on it 20:58:04 hm 20:58:20 heh yet again slate discussion avoids #tunes :) 20:58:47 what? 20:58:55 hmm 20:58:58 stuff about meta-compiling 20:59:05 did you just accuse me of changing the topic of discussion ? 20:59:11 not at all 20:59:15 we should base tunes on hmm (hesitant meta-manipulation) 20:59:34 abi: hmm? 20:59:35 no idea, tril 20:59:35 slate discussion just does this on its own of late 20:59:51 Batterij: :P 21:00:09 Batterij: meta-compiling is an issue from forth experience 21:00:26 right now, forth is the only language you can do it in 21:00:26 Good one. HMMMMMM = .. Meta meta meta meta ... manipulation 21:00:47 heh 21:01:08 Computer Science: 1. n. the practice of assigning concepts to acronyms 21:01:24 heh 21:06:24 water considers writing up the lazy object pattern on the pattern wiki 21:08:53 [NICK] lar-movie changed nick to: lar1 21:09:38 Tril! :) 21:09:48 Tunes mlist oldie of the day: 21:09:50 http://lists.tunes.org/cgi-bin/wilma_glimpse/tunes?query=epsilon&errors=0&maxfiles=50&maxlines=10&.cgifields=lineonly&.cgifields=filelist&.cgifields=case&.cgifields=partial&.cgifields=restricttofiles 21:10:18 where s/epsilon/choice function/ 21:12:25 lar1: hello 21:19:11 water: I found a WWSympa distribution, I'll try setting it up very soon 21:19:32 ok 21:20:01 the choice construct is essentially a relation to the clone concept 21:20:36 where choice does not pre-suppose an existing computed instance 21:22:21 it also can be seen as a co-inductive result of a function.... it specifies a result (with vacuous constraints) and accepts its existence until a function fulfills the requirements 21:22:36 hm that didn't come out right 21:25:04 basically it amounts to "there exists x" which can be composed in various ways with constraints on "x" 21:26:02 and "x" is considered consistent to its type until proven otherwise, due to constraints introduced 21:26:55 which essentially means that the system allows any proof/function for the object that works until it finds a contradiction 21:27:22 even if the proof/function is non-computable (like a person's intuition) 21:27:31 in that way, it's co-inductive 21:27:59 yes, that explanation works 21:28:11 "Access without authorization denied!" -- ACM 21:28:58 i'll sponsor an account for you 21:29:15 since you'll get student rates 21:32:42 [QUIT] Alyssa quit: Alyssa 21:41:24 oh cool! more drivel from lmaxson on the tunes list! :) 21:43:36 damn, he really is an old fool 21:44:39 i wonder what it'lll take the guy to shut up 21:44:56 abi, tunes list archive or somesuch? 21:44:56 i don't know, batterij 21:44:59 i can't discuss anything serious on the mlist yet 21:45:21 the tunes mlist is not getting archived publically currently 21:45:35 http://www.tunes.org/collaboration.html 21:46:01 k.. I am on a serial terminal and did not bother to start screen before IRCing 21:46:30 this guy really infuriates me 21:47:02 he explicitly contradicts the tunes goals and insists we are wrong about how to go about our (HIS) goals 21:48:07 i know a fellow at work who is as stupid and stubborn as lmaxson 21:48:17 billh joined #tunes 21:48:19 everyone hates him 21:48:20 Hello. 21:48:23 hey bill 21:48:30 what's up ? 21:48:44 lmaxson pollutes the tunes mlist with his ignorance 21:48:56 billh was talking to a 20 year old kid that claims to have a FFS+softupdates+snapshot replacement. 21:49:09 ffs? 21:49:16 billh: amiga ffs?!! 21:49:30 billh: why do you mention that crap here? 21:51:51 FreeBSD FFS. 21:52:16 I mention it because it's wierd that he'd make that claim. 21:52:29 offtopic, dude 21:52:37 like everything else you bring up here 21:52:59 we have #{} or #offtopic for other stuff 21:57:27 [QUIT] Batterij quit: Leaving 22:02:31 water returns to reading fpsood 22:19:33 [QUIT] water quit: The Tao went that-a-way! 22:19:41 water joined #tunes 22:19:53 oops 22:20:02 wrong button? :) 22:20:18 yep 22:23:01 [QUIT] billh quit: BitchX: your way, right away 22:29:46 [QUIT] td quit: Ping timeout for td[ilm25-51-108.ec.rr.com] 22:31:45 downix joined #tunes 22:32:04 hey dx 22:32:18 hey, I need LISP help 22:32:31 we're discussing, could LISP be used for device drivers? 22:32:48 sure 22:33:46 also, how large would a LISP interpreter have to be? 22:34:39 I've gotten into a project and I'm seeing if LISP would work for it 22:35:32 a tiny embedded lisp could be pretty small 22:35:40