02:09:01 karltk joined #tunes
02:27:43 Kyle_L joined #tunes
03:00:48 karltk_ joined #tunes
03:01:38 [QUIT] karltk quit: fastlane.openprojects.net lackey.openprojects.net
03:14:16 Fare joined #tunes
03:58:44 [QUIT] Fare quit: Leaving
04:18:56 Fare joined #tunes
05:02:28 [QUIT] Fare quit: Read error to Fare[APastourelles-101-1-1-83.abo.wanadoo.fr]: Connection reset by peer
05:05:46 [QUIT] ink|sleep quit: Ping timeout for ink|sleep[ppp-206-170-24-96.sntc01.pacbell.net]
05:15:46 [NICK] ult[sleep] changed nick to: ult[class]
05:18:40 Fare joined #tunes
05:33:07 [NICK] corey changed nick to: cor[class]
05:52:00 _101 joined #tunes
05:52:02 <_101> hi
05:52:19 <_101> i'm new to the project and find it VERY intresting
05:52:34 <_101> but recently there's not much moving on the tunes.org site...
05:53:02 <_101> what is the current state, especially of Slate?
06:58:03 <smkl> _101: water is still designing it
06:58:47 <_101> i know, but i would like to read some of the ideas and such... just reading the mail archives :)
06:59:23 <_101> the docs at ~water are too general and my english is not trained for such
07:00:15 <_101> i must see how things work to fully understand and reply/ask about such docs
07:00:57 <smkl> _101: there is not that kind fo documentation yet
07:01:38 <smkl> there was some new documentation, but i think it is nuked, and it wasn't what you'd want anyway
07:01:59 <_101> did you see Lx and Mozart?
07:02:01 <_101> LX
07:02:09 <_101> a programming language and a dev enviroment
07:02:35 <_101> the guy working on it seems to have the same ideas as some subset of tunes
07:02:48 <_101> and it's GNU, you should get togather
07:03:01 <ult[class]> Why do people always come in here telling us what we should do?
07:03:40 <_101> )
07:03:53 <_101> it would be good... if you like it better
07:04:07 <_101> it's just my english
07:04:11 <_101> plain and simple
07:04:27 <ult[class]> Yeah, right.
07:04:31 [NICK] ult[class] changed nick to: ult
07:06:12 <_101> mozart-dev.sourceforge.net
07:06:30 <_101> there's LX also with a feature list
07:07:04 <_101> LX has several good fatures, but it's a compiled language like c++
07:09:22 <_101> must go now, cu
07:09:26 _101 left #tunes
07:40:47 [NICK] ult changed nick to: ult[class]
08:32:25 [QUIT] Fare quit: Leaving
08:58:36 hcf joined #tunes
09:01:14 <corey> hcf, do you want a diktuon2 account?
09:01:35 <hcf> ok
09:01:48 <corey> msg me u/p
09:02:03 <corey> diktuon.tunes.org/diktuon2
09:05:22 <corey> i dont like the tree as it stands
09:05:46 <hcf> bearing bad tasting fruit?
09:07:22 <corey> that and i think slate should be a subnode under hll or 'hll candidates'
09:07:43 <hcf> why not symlink'd to 'hll candidates'?
09:08:47 <corey> hmm
09:09:25 <corey> the tunes pages are resisting conversion to nodes. :)
09:09:36 <hcf> how so?
09:09:56 <corey> hmm
09:09:59 <corey> hard to explain
09:10:33 <corey> its hard to determine whether information should be a subnode in its own right or a paragraph in the node
09:10:38 <corey> well
09:10:41 [QUIT] JALH quit: Read error to JALH[user36-102.jakinternet.co.uk]: EOF from client
09:10:41 <corey> it should be easy
09:10:52 <corey> but the text is not written with that in mind
09:11:06 <hcf> so the auth of the text should do it
09:11:11 <corey> fare? :)
09:11:27 <hcf> or the text should be 'scrapped'
09:11:41 <corey> i.e. rewritten/reorganized
09:11:44 <hcf> not rm'd, just lept as reference
09:11:51 <hcf> s/lept/kept/
09:12:26 <hcf> r u keeping the ~8k limit?
09:12:29 <corey> its not me
09:12:34 <corey> its pgsql
09:12:40 <hcf> i know
09:12:47 <hcf> just didnt know if u had found a way around it
09:12:59 <hcf> its good, should even be smaller ;)
09:13:05 <corey> )
09:13:06 <corey> heh
09:13:11 <hcf> to force modularity
09:13:30 <corey> and reuse
09:14:01 <corey> there's a fare-container-rant somewhere in review
09:14:09 <corey> i think im going to use it as a todo of sorts
09:15:00 <hcf> wheres ur current todo?
09:15:15 <corey> root_diktuon_todo
09:15:35 <corey> http://www.tunes.org/Review/index.html#TODO
09:15:40 <corey> near that is the rant
09:15:56 <hcf> how do u get to root_diktuon_todo theres seems no link
09:16:23 <corey> http://diktuon.tunes.org/diktuon2/display.php?id=0
09:16:32 <corey> then click on the appropiate node
09:16:44 <corey> how does this look in lynx... :)
09:17:04 <hcf> good enuf for this point
09:17:47 <hcf> how the hell does one get to id=0 normally?
09:18:00 <corey> up
09:18:07 <corey> its name is root
09:18:28 <hcf> oic
09:18:46 <corey> i think its neat.
09:19:07 <corey> i thought maybe it might take too many db queries
09:19:10 <corey> but it seems fast
09:19:44 <corey> im using php's xml parser
09:20:19 <corey> makes it easier to parse the nodes and to extend
09:20:31 <hcf> the nav is odd
09:20:35 <corey> odd?
09:21:20 <hcf> it confuses w/ the traditional top,bottom,prev,next style
09:21:32 <hcf> but next doesnt work right
09:21:51 <hcf> i said that wrong
09:22:17 <hcf> ok, i expect next from root_diktuon to take me to the next item under root_diktuon, 'todo'
09:22:53 <corey> hrm
09:23:52 <corey> but since you always see the children, is that necessary?
09:24:26 <hcf> i guess not
09:25:57 <corey> its only 14k
09:28:50 <hcf> going 'up' from the front page is odd
09:29:02 <corey> well its root_intro
09:29:26 <hcf> put it above root_dikt
09:29:37 <corey> oh
09:29:44 <corey> that list is alphabetized
09:29:51 <corey> i was thinking about that
09:30:02 <corey> maybe i should have an order attribute
09:30:22 <corey> that way
09:30:36 <corey> when someone is adding things they want sequential
09:30:44 <corey> it will be.
09:31:04 <hcf> yep
09:31:21 <corey> adding to the todo.
09:33:07 <corey> anything else?
09:33:30 <hcf> reuse of previous poop i'v said
09:33:43 <corey> i think the default being intro is a bit awkward
09:33:44 <hcf> like the who.cgi thing
09:33:55 <hcf> instead of plain text names
09:34:05 <hcf> buts thats later
09:34:10 <hcf> s/ts/t/
09:34:23 <hcf> a help system
09:34:48 <hcf> non numeric id node access
09:34:54 <corey> hmm
09:34:59 <hcf> w/o removing numeric id access
09:35:09 <corey> thats difficult
09:35:18 <corey> limitations of urls
09:35:22 <hcf> numeric is fine for now
09:35:28 <corey> well
09:35:31 <corey> its all in the db
09:35:49 <corey> so i dont think itd be difficult to write other front ends for the data
09:36:02 [NICK] ult[class] changed nick to: ult
09:36:49 <corey> ult: you want a user?
09:37:18 <ult> Sure
09:38:10 <hcf> corey: u should (at some point) do a dikt aware 404 page, instead of falling thru to bespin's
09:38:32 <corey> hmm
09:38:42 <corey> thats apache.conf
09:38:46 <corey> ill look at it
09:39:19 <hcf> search and login should be node unto root_dikt ;)
09:39:24 <hcf> s/node/nodes/
09:39:29 <hcf> s/unto/under/
09:39:31 <hcf> geez
09:40:17 <hcf> what r u using for auth nowadays?
09:41:32 <corey> http post to db lookup that sets cookie
09:42:28 <corey> ok well i have a date
09:42:30 <corey> bbl
09:42:35 [NICK] corey changed nick to: cor[lunch]
09:48:28 water joined #tunes
09:48:41 <water> hey all
09:48:46 <ult> 'lo water
09:51:52 <water> hm my post to the tunes list yesterday still hasn't made it
09:53:15 <ult> weird
09:53:21 <ult> tunes list seems kind of sluggish/broken
09:53:43 <water> well, it's more related to the fact that i posted from the ship
09:53:57 <water> in fact, i know exactly what's wrong
09:54:10 <water> because i made sure to cc it to this accout
09:54:15 <ult> doh!
09:54:16 <water> and it hasn't shown up
09:54:45 <water> so it seems those geniuses (ship's server admins) have screwed up again
09:55:17 <water> those fellows are so unbelievably stupid
09:55:55 <water> they wouldn't know tcp/ip from one of their bodily orifices
09:57:37 <water> "He may find there aren't enough atoms in the universe to store the yes-no questions that might be required even by a fourth-grader to do a homework problem." 
09:57:37 <water> — Doug Lenat
09:59:02 <water> hm i missed corey again
10:03:06 [QUIT] MysticOne quit: irc.linux.com carter.openprojects.net
10:03:06 [QUIT] cor[lunch] quit: irc.linux.com carter.openprojects.net
10:10:49 cor[lunch] joined #tunes
11:03:30 [NICK] JALH changed nick to: WindowsGold
11:09:01 <corey> water: hey
11:09:39 <water> yo
11:09:52 <water> i was about to go shower, but...
11:10:04 <water> how're things going? i looked at dikt2
11:11:34 <corey> good
11:11:44 <water> damn, it looks like my email from the ship will take forever to reach the tunes mlist
11:12:09 <corey> yeah with that one thread, i got the replies days before i got your original
11:13:21 <corey> how about them slate docs
11:13:31 <corey> what do you need from me?
11:13:33 <water> still working on theory and mop design
11:13:35 <corey> ok
11:14:03 <water> i'm working hard on it, but the language paradigm needs careful attention before i try to explain it
11:14:12 <corey> agreed. :)
11:14:17 <water> heh
11:15:22 <corey> well ill keep trudging on with dik2
11:15:27 <water> ok thx
11:16:06 <corey> did you try to add a node?
11:16:11 <water> not yet
11:41:49 [QUIT] Downix quit: Read error to Downix[frednet-mcm-e-015.students.frednet.uwrf.edu]: EOF from client
11:42:21 <corey> hmm
11:42:32 <corey> question, should nodes be case sensitive or not
11:43:10 <corey> its not a problem either way, cept which is easier for people to remember.
11:43:41 <corey> i vote for case sensitive
11:44:08 <water> oh?
11:44:18 <water> hm
11:44:53 <water> doesn't bother me either way
11:44:56 <corey> ok
11:48:23 <water> hm "an equational definition of partial evaluation"
11:48:33 <water> (is within pebook)
11:51:00 <water> heh
11:51:29 <water> the authors suggest using p/e to specialize a ray tracer to particular attributes of a scene
11:52:05 <water> yes, indeed, someday i will create that real-time ray-tracing game engine >:)
11:54:07 <hcf> corey: what r the pros for the user is they r case sensitive?
11:55:24 <corey> hcf: some names need to be cased
11:55:50 <hcf> such as?
11:56:50 <corey> i dunno
11:56:52 <corey> heh
11:56:57 <water> geez
11:57:03 <water> how about "HLL"?
11:57:04 <corey> TUNES
11:57:15 <water> "StdLib"?
11:57:22 <hcf> will there be both 'HLL' and 'hll'?
11:57:36 <corey> hcf: you gave me an os name before
11:57:42 <corey> i cant remember what it was
11:57:54 <hcf> v2?
11:58:16 <hcf> was on the subject of chars in the url
11:58:29 <hcf> cuz it can be v2/os v2-os v2_os etc
11:58:36 <hcf> wasnt a case issue
11:58:40 <corey> hmm
11:59:15 <corey> i like case
12:00:01 <corey> and thats the way it is now
12:00:13 <corey> so if there is a great case for case, ill change it :P
12:00:56 <corey> s/case/case insensitive
12:04:09 <hcf> its fine, as long as an error page offers names close to the given
12:05:56 <corey> yeah
12:06:04 <corey> exactly
12:06:38 <corey> does anyone have any idea on how to implement that? btw... :)
12:06:58 <corey> i dont know which angle would be best to attack it from
12:07:04 <water> take the lowercase of the search string and compare with lowercase of nodenames
12:07:27 <water> return the results of that query
12:07:31 <corey> ok
12:07:40 <hcf> corey: perl has string::approx mod
12:07:48 <hcf> i assume php has smth like it
12:08:09 <corey> how does a word processor spell checker guess words?
12:08:22 <corey> hcf: does string::approx act like that?
12:08:29 <corey> not just the initial similiar
12:08:30 <hcf> corey: that and more
12:08:33 <corey> hmm
12:08:34 <water> probably a totally different process than you need
12:08:49 <water> (from word processors)
12:09:34 <corey> levenshtein hmm
12:09:46 <water> huh?
12:10:16 <corey> The Levenshtein distance is defined as the minimal number of characters you have to replace, insert or delete to transform str1
12:10:17 <corey> into str2.
12:10:29 <water> oh yeah i've heard of that
12:10:50 <corey> hmm should be easy now
12:12:30 <corey> oh im cluttering logs
12:12:35 [NICK] corey changed nick to: cor[dikt2]
12:13:26 [QUIT] JALH quit: BitchX has bite! (Just ask Mike Tyson!)
12:20:50 <hcf> wow, php.net's 404 page really sucks (ex: http://www.php.net/search.ph)
12:21:15 <water> heh maybe that says something? ;)
12:22:07 <hcf> http://www.php.net/search works sorta
12:22:40 <hcf> uses ht:dig
12:44:20 <water> ok... i'm heading out for a few hours
12:44:34 <water> i'll probably be up all night, and i have tomorrow off as well
12:44:48 <cor[dikt2]> later
12:45:00 <water> i'll write up some side threads i intended to start when i get back
12:45:16 <water> it looks like the original thread will take a while to get delivered
12:45:27 <water> cya corey, hcf, lurkers
12:45:31 [QUIT] water quit: The Tao went that-a-way!
13:42:15 [QUIT] Geeky quit: Leaving
14:15:53 [QUIT] JALH quit: Ping timeout for JALH[host5-99-57-175.btinternet.com]
14:49:10 Fare joined #tunes
15:02:57 [QUIT] JALH quit: ....zzzzzZZZZZZZzzzzzzz.......zzzzzzZZZZZZZzzzzzzz......zzzzzZZZZZZzzzzzz....yyyyyyYYYYYYYyyyyyy....xxxxxxXXXXXXxxxxx... g'night :)
15:07:50 <ult> i dunno
16:23:14 |U_Killed_Kenny| joined #tunes
16:23:22 [QUIT] |U_Killed_Kenny| quit: Read error to |U_Killed_Kenny|[s029.arbv.res-hall.gac.edu]: Connection reset by peer
16:35:41 lar1 joined #tunes
16:36:13 [NICK] ink|sleep changed nick to: ink
16:37:33 <Fare> Ok, so where were we?
16:37:54 <lar1> 17:00:08 <lar1> Does Retro not have a FORTH interpeter?
16:37:55 <lar1> 17:00:29 <Fare> interpret = compile execute
16:37:55 <lar1> 17:00:46 <lar1> What more do you speak of?
16:37:55 <lar1> 17:11:32 <Fare> anything
16:37:55 <lar1> 17:12:06 <lar1> Could you give an example?
16:37:57 <lar1> 17:18:12 <Fare> well, ok, let's take a pattern matcher
16:38:12 <Fare> Ok.
16:38:32 <Fare> We want to extend a base language with some pattern-matching capability
16:39:25 <Fare> FORTH is a bad example, for you'd have to add structures first, before you can even match them.
16:39:45 <Fare> So by the time you add the feature, it's no more forth, already.
16:39:53 <lar1> Pattern matching for all datatypes?
16:40:23 <Fare> Yes; or at least in a way that can be extended for yet unsupported datatypes.
16:40:37 <Fare> more importantly, pattern-matching for source statements
16:40:57 <Fare> without having to invent and make explicit a new syntax for internalized statements
16:41:06 <Fare> as you do in ML.
16:41:09 <lar1> Wouldn't that just be matching strings?
16:41:13 <Fare> No
16:41:18 <Fare> You'd match structures
16:41:57 <lar1> Source statemnts in structures?  I don't understand.
16:42:14 <Fare> just like, in ML, you match term with (App (Lambda var body) t) => replace var t body | x => x
16:42:47 <lar1> Ah, ok.
16:43:00 <Fare> well, you'd write match term with `((lambda ,var ,body) ,t) => replace var t body | x => x
16:43:55 <Fare> i.e. unless you specifically want to, you don't end up with lots of ugly stuff that depends on a lower-level representation of source than you use
16:44:56 <Fare> ok. So how do you add such a feature?
16:45:50 <lar1> I don't know. :)
16:45:55 <Fare> In LISP, you might end up with something that interacts badly with macros, for instance
16:46:40 <Fare> just like Screamer has a code-walker that transforms whole programs so as to add backtracking ability to CL
16:46:59 <lar1> What does that have to do with pattern matching?
16:47:46 <Fare> well, so as to understand arbitrary structures, the matcher must expand macros, etc, so as to go down to whatever primitives there are that it understands
16:48:16 <Fare> basically, it has to re-do in "userland" a large part of what the compiler does.
16:48:39 <Fare> Instead, it ought to be able to add the processing at the right moment directly into the compiler
16:49:42 <Fare> (another possibility is to not understand macros or any kind of complex patterns)
16:50:05 <lar1> So what qualifies as a pattern?  Everything?
16:50:39 <Fare> and of course, if you try to use at the same time two packages that reimplement or dirtily try to work around the compiler, they might interact bizarrely
16:51:02 <Fare> lar1: anything is a pattern; only a pattern can be used constructively or destructively
16:52:07 <lar1> So pattern matching is part of the foundation of a reflective system? 
16:53:05 <Fare> that is, the same `(lambda ,var ,body) can be used constructively to make a new lambda abstraction, or destructively to analyze one.
16:53:31 <Fare> yes, pattern matching is essential to achieve a _high-level_ reflective system
16:53:50 lar1 does not know ML
16:53:55 <Fare> or else, you must enforce explicit low-level representations
16:54:09 <Fare> (like LISP SEXP, for instance)
16:54:19 <Fare> (or those funky types in ML)
16:55:39 <Fare> oops, IRQ
16:55:53 <lar1> Aren't there already functions that pattern match?
17:29:37 eihrul joined #tunes
17:37:53 <Fare> RTI
17:38:22 <Fare> lar1: yes, they can match precisely one kind of very restrictive patterns
17:38:33 <Fare> (I presume you mean in LISP)
17:38:58 <Fare> (actually, in CL, the patterns are more complex than in Scheme, since you have optional arguments, etc)
17:39:22 <lar1> There isn't a pattern matcher that works on the source level of yet?
17:39:23 <Fare> similarly, ocaml now has keyword arguments
17:40:16 <Fare> lar1: not that I know of; or more precisely, none that has good symmetry between patterns as constructors and patterns as destructors
17:40:36 <Fare> the Coq grammar extension mechanism somehow is a source-level pattern matcher.
17:41:05 <lar1> Hmm, ok.
17:41:14 <Fare> but it's not embedded in the normal language; it's an ad-hoc extension stuff.
17:41:56 <lar1> So the LLL Sub-Project requrirements would be met by extending LISP's pattern matching facilities?
17:42:09 <Fare> not exactly;
17:42:32 <Fare> rather, adding a pattern matcher to a lisp skeleton would give a HLL-
17:42:51 <Fare> and then you can write an open HLL- compiler in HLL-
17:43:01 <Fare> and the lower layers of this open compiler is the LLL
17:43:18 <lar1> Ah
17:43:54 <lar1> Ok, thank you for you time.  I must be off now.
17:44:06 <lar1> I hope I wasn't too annoying.
17:47:42 <Kyle_L> Fare: what do you mean by patterns as constructors and patterns as destructors?
18:02:32 <eihrul> well, if you construct something... you obviously want to destruct it to get useful information out of it eventually
18:03:02 <Fare> depends on whether the pattern is on the lhs or rhs of a matching construct
18:04:07 <Kyle_L> thanks, but now I have to go.
18:04:10 [QUIT] Kyle_L quit: Leaving
18:05:04 <eihrul> Fare: what if the monitor is turned upside down?
18:22:01 <Fare> AND your font is such that...
18:28:44 hcf joined #tunes
18:29:56 water joined #tunes
18:30:07 <water> hey
18:30:33 <water> ah. my post finally made it
18:31:17 [QUIT] MysticOne quit: [BX] I'm out like a light...
18:31:30 <water> hm Fare has been active today
18:32:32 <water> hey eih!
18:34:42 <eihrul> 'lo
18:34:50 <water> how're things going?
18:35:08 <eihrul> #{}
18:35:18 <water> i have a lot of things to talk about, perhaps best done via email (for tunesers' benefit)
18:43:11 [NICK] corey[away] changed nick to: corey
18:50:53 ult joined #tunes
19:09:24 [QUIT] morton quit: Ping timeout for morton[freenet.nether.net]
19:15:10 [QUIT] Fare quit: Ping timeout for Fare[APastourelles-101-1-2-94.abo.wanadoo.fr]
19:23:48 morton joined #tunes
19:24:03 <water> hey morton
19:24:08 <morton> yes?
19:24:19 <water> slate discussion is on #{} right now and in #{}ogs
19:24:29 <water> s/ogs/ logs/
19:24:41 <morton> ah
19:24:42 [QUIT] _QZ quit: Leaving
19:24:58 <morton> when did i become a slate fan?
19:25:01 <morton> -)
19:25:24 <water> you're arriving just reminded me to say that
19:25:46 [QUIT] air_ quit: BitchX-74a14 -- just do it.
19:28:24 [QUIT] brand quit: BitchX-74a14 -- just do it.
19:35:46 [NICK] nate37_pastey changed nick to: nate37_desk
19:41:56 [NICK] ult changed nick to: ult[sleep]
21:07:00 [QUIT] MysticOne quit: Ping timeout for MysticOne[porpoise13.panama.gulf.net]
21:19:33 [QUIT] eihrul quit: Leaving
21:34:20 [QUIT] hcf quit: Leaving
21:38:10 [NICK] lar[dinner] changed nick to: lar1
21:38:34 [NICK] ink|food changed nick to: ink
22:55:33 [QUIT] lar1 quit: Moon Baby.
23:36:24 water joined #tunes
00:03:12 witten joined #tunes