IRC log started Sat Oct 23 00:00:01 1999 [msg(TUNES)] permlog 1999.1023 12:10am -:- Tril [dem@bespin.dhs.org] has joined #tunes -:- ServerMode/#tunes [+o Tril] by clarke.openprojects.net hey tril 12:40am well, i'll be back tomorrow -:- water [water@tnt-10-106.tscnet.net] has left #tunes [] 12:50am -:- _ruiner_ [DIY@ppp024.wi.centurytel.net] has joined #tunes -:- _ruiner_ [DIY@ppp024.wi.centurytel.net] has left #tunes [] -:- SignOff td: #TUNES (td has no reason) -:- SignOff Zhivago: #TUNES (Ping timeout for Zhivago[th.merddin.com.au]) -:- Zhivago [brian@th.merddin.com.au] has joined #tunes hmm 05:30am -:- fare [fare@quatramaran.ens.fr] has joined #Tunes -:- fare is now known as Fare -:- tcn [tcn@cci-209150250112.clarityconnect.net] has joined #tunes t c n ! 06:10am * tcn/#tunes is away. Automatically set away [SZon] hey fare i'm trying out a new irc client gotta turn off all the crap features.. :) 06:20am * tcn/#tunes is away. Automatically set away [SZon] douuugh 06:50am * tcn/#tunes is away. Automatically set away [SZon] hum 07:10am ... ping 07:20am * tcn/#tunes is away. Automatically set away [SZon] -:- SignOff tcn: #TUNES (That's it for today) -:- Netjoined: sterling.openprojects.net varley.openprojects.net -:- Tril [dem@bespin.dhs.org] has joined #tunes -:- ServerMode/#tunes [+o Tril] by sterling.openprojects.net -:- hcf [nef@me-portland-us205.javanet.com] has joined #tunes Tril 07:50am -:- SignOff Zhivago: #TUNES (Ping timeout for Zhivago[th.merddin.com.au]) -:- eihrul [eihrul@usr5-ppp216.lvdi.net] has joined #tunes -:- SignOff eihrul: #TUNES ([x]chat) -:- Zhivago [brian@th.merddin.com.au] has joined #tunes -:- NetSplit: clarke.openprojects.net split from sterling.openprojects.net [08:22am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [clarke.openprojects.net] -:- eihrul [eihrul@usr5-ppp216.lvdi.net] has joined #tunes re * eihrul/#tunes nods. 08:50am -:- AlonzoTG [Alonzo@216-164-136-171.s171.tnt4.lnhva.md.dialup.rcn.com] has joined #tunes -:- AlonzoTG [Alonzo@216-164-136-171.s171.tnt4.lnhva.md.dialup.rcn.com] has left #tunes [] -:- SignOff Fare: #TUNES (Ping timeout for Fare[quatramaran.ens.fr]) -:- Fare [fare@quatramaran.ens.fr] has joined #Tunes -:- NetSplit: adams.openprojects.net split from devlin.openprojects.net [10:27am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [adams.openprojects.net] -:- water [water@tnt-10-20.tscnet.net] has joined #tunes hcf: i noticed that jason marshall just now started speaking up i also noticed but then, Jim Little made his little announcement, too brb wow. i'm sooo thrilled that prism is working yeah, it's just what tunes needs... the ultimate super-kludge 10:40am water: u seen http://www.openave.com, http://www.sourcexchange.com, http://www.cosource.com? not specifically is there a reason why i should look there? may assist w/ getting developers but these sites are for the request/provide model hmm ok i'm not sure if i could find someone who could make the code i want, though -:- eihrul [eihrul@216.24.141.216] has joined #tunes 10:50am oh great hey eihrul, what do you think of prism? brb, switching logons -:- _water [water@tnt-10-69.tscnet.net] has joined #tunes 11:00am -:- SignOff water: #TUNES (Ping timeout for water[tnt-10-20.tscnet.net]) -:- _water is now known as water -:- td [x@1Cust248.tnt3.wilmington.nc.da.uu.net] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us734.javanet.com] has joined #tunes er... wb er... thx water: haven't looked at prism yet * water/#tunes attempts to provide the laymen's intro for ontologies, etc to tunes mlist * eihrul/#tunes claps. heh hopefully, this should clear some things up 11:30am well, it's on its way enjoy! 11:50am hcf: jason marshall also wrote me personally well, i'll be afk for a while oh? -:- fire [no@209-68-229-224.dialup.cust.tfb.com] has joined #tunes * fire/#tunes is back from his race :) 12:00pm how'd you do? hcf: yeah he asked about arrow, but i don't think he groks the tunes ideas yet ok 21:07 on 3.1 miles but of course im a freshman you can't expect much someday i hope to grok arrow :) heh what is DMOZ? www.dmoz.org k the mozilla open-directory where'd you hear about it? aka odp, aka noohoo -:- Kaufmann [Kaufmann@dial753.infolink.com.br] has joined #tunes hmm, i ges they did away w/ using noohoo ... 12:10pm "Coding has begun." !@#$ on what? TUNES on the DMOZ comment heh yeah it's embarrassing !@#$ -> (not (length $)) :) no one wants to write a bootable scheme water, what? what do you mean by bootable? schemeOS heh scheme has problems it's the best LISP tough that's what tunes is supposed to start on though rather well, it could be a lispos instead we should write a very simple HLL interpreter then bootstrap well, there is laurent martelli's idea for an aop lisp and he does have a basic lisp written in c well, alisp is not Scheme, but it's a helluva lot like it, and (although first it will run as a regular process on top of UNIX) I intend to make it bootable, eventually. i.e., capable of running by itself Of course, I really have nothing so far. :P hell we could write the HLL interpreter in COBOL for all i care -:- water has changed the topic on channel #tunes to: Let's actually make a Free Reflective Computing System at || think for a moment, that's right use your brain, shouldn't this discussion be taking place in #OsDev? -:- AlonzoTG [Alonzo@209-122-203-234.s488.tnt6.lnhva.md.dialup.rcn.com] has joined #tunes why don't we just not use the term "bootable " let's just call it a computing system om I love minix. =P wrong channel, atg ? my computer is dying. =( nobody really likes computers. atg: use #osdev okay... I'm there too but its empty. oh... Is music talk allowed here? i like tom's post got my response yet? yeah 12:20pm im going down the list gimme a second -:- SignOff td: #TUNES (Ping timeout for td[1Cust248.tnt3.wilmington.nc.da.uu.net]) damn i haven't heard from jim little in a long time yeah, it's a shame that he had to pollute the list again ;) heh prism doesn't look that exciting abi, prism? rumour has it prism is a way to program with multiple specialized, interoperable languages. or at http://www.teleport.com/~sphere or YABL BL? bad language? heh abi: yabl? yabl is Yet Another Burden Layer aaaaah lol water: what is an arrow? i think an arrow is a homo-iconic information manipulation system intended to support programming and human languages in the way that Tunes should. at http://www.tunes.org/papers/Arrow/, http://www.tunes.org/~water/ an arrow not arrow oh think of it as a cons cell a directed link er... not just a link i mean, both a cons cell and a link have certain connotations. arrows are general-purpose constructs just like a list is a general purpose construct for lisp? sure but lisp is just a programming language 12:30pm arrow ain't so you just apply arrow to a programming language? apply it? hmm make it arrow equivelent yeah, you can model a language abstractly with arrows i mean you're gonna have to have programming languages sure, but now they'd just be algorithm notations fully-fledged abstract representations water: hmm, i'm still having trouble picturing what part of the system is "active" :) well, just think of an arro-information evaluator do ontologies actively translate things or are they tools for some other active entity? they're tools for agents don't read too much into the agents idea so an ontology can be applied to a given arrow structure by some active thing (whatever that is at this point)? yeah, just an extension of the evaluator a "task" ultimately yielding a whole other arrow structure? yep encoded lazily ok but the arrow structure can be reflected on you can store arrow structure in other data-formats as long as the original arrow content can be obtained later * eihrul/#tunes was thinking something along those lines before; if you can reflect on any part of the system you might be able to ultimately change the representation of arrows themselves. cool 12:40pm the ontology system allows interface reflection and ultimately data-structure reflection and code-generation since code is just a data-format well, sort of :) * Kaufmann/#tunes has to admit that he's understanding very little of all this * fire/#tunes ditto damn did you read my latest email? water, to the Tunes list? yeah -:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes yes I did hey bin hey water hello my swedish rose :D LOL 'nyway... :) well, did it explain anything? did you check out those url's? water: I picked up your paper again bin: unfortunately (it really sucks) i read a story in english class about a swedish king that travels the country side looking for support to raise a new army and finds a silver mine way out in the country wait, what email are you talking about? The _where tunes is going_ thread-starter? water: I've found that it's pretty related to this thing I'm doing, AKOS the one i posted just this morning -:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[209-122-203-234.s488.tnt6.lnhva.md.dialup.rcn.com]) fire: A real king? beats me brb water, yes, I read that one too he goes into a town and the minister tells him about the mine but there are no links there no URLs, I mean huh? umm, no water: Have time for a few q's? bin: sure water: How can agents see and interact with the system? 12:50pm kauf: the mail i sent this morning has three url's, two of which have many links let's look again bin: well, they get information by linking domain-specific info with ontologies bin: it's like routing packets semantically bin: the agent is the router hm it picks the route based on its purpose for the info how are agents "associated" with an ontology? (because that's how ontologies are used, not?) what is an ontology? i guess an ontology is a statement of a logical theory in some domain or what a person believes to be provable. i.e. what _can_ possibly be done in the world well, i think that they actually just work with domains, and the ontologies link the domains. the agent has access to the ontology graph and activates its parts to translate info to the problem domain fire: check out my e-mail from this morning aaaaaaaah, /new/ email Sounds like domains are something different.. * binEng/#tunes looks in the paper it's just an information's original source like a data-structure format The 'real' structure of the arrow graph? bin: referring to what? oh domain hmm yes -:- Miellaby [sgarden@ppp-214-28.kd.club-internet.fr] has joined #TUNES hi miel maybe you mean that a domain is an isolated graph? But that sounds like it's against the spirit of it I'm leaving well, the ontology system has to built on top of arrow -:- SignOff Kaufmann: #TUNES ("Jesus would make a good main character for a Seinfeld spinoff" - Amber << Well, it got my attention... coming soon, the pilot episode of BEN YOSSEF! Sundays at Atheist News Network!) hi water you can 'subgraph' your graph to restrict it to a domain water, can we have a speak about your system? miel: sure, it seems it's the hot tunes topic bin: unfortunately, the way i just explained it isn't good enough for the docs water: I thought ontologies were a result of the structure of the graph? water:I'm totally seduced by the Arrow proposal. I wish I could understand more. yes, ontologies come from graphs ok miel: darn it, i was too convincing well, keep in mind everyone that i'm still working lots of things out but those arrows that makes of ontologies are 'on top' of the "real" information web, right? 01:00pm it _is_ weird that some of the nodes are links, too, isn't it? :) it begins to make some sense, methinks water: btw, did you work any further on the paper? What's the latest version? well, an ontology is a smenatic specification which has to exist somewhere and must be specified in some language bin: yes, but the new version will have to run over 100 pages and i'm nowhere near done. Specified in some language? That suggests that ontologies exist in a higher level than arrows and i've been having serious computer problems lately darn it, i chose the wrong words it all boils down to "how do you specify a graph's structure?" imo you need a semantic structure like logic to specify the graph you need an ontology to provide the semantic structure Well that's a snap when the epistemic building blocks are arrows (is that the right use of the word?) and you need graphs to construct the ontology yeah, but what if you want to build a grpah that has more than an infinite number of arrows? what if you want to have a graph of "all binary trees"? I don't see why you need semantic structures for the graph. Unless you mean a binary representation, read by some sort of interpreter? But that's very lowlevel i mean, how do you declare a graph Outside the system, it must be, . outside the whole system? Outside the arrow system. hell no why would i want the arrow system to be unable to manipulate itself ? I'd love to see you do the trick of a recursive definition, but I'm afraid that is a bit hard well, it involves refelction The system must start *somewhere*. It can't come out of thin air and define itself. at some point, you substitute "actual arrow system" with code that does the same thing yes i.e. a vm that supports reflection vm = lang spec Do you want to define graphs in and only in Arrows? sure eventually, i want to make the vm completely transparent 01:10pm ..and Arrows is then defined as a graph, which is defined in Arrows? How'd you do that? water : I got the feeling there is a lake of one or two fundamental operators in the system design. Those primitives should be seen as "special" arrows. right, and i've worked out a lot of those to who? bin: well, that's still an open question miel: right, and i've worked out a lot of those sorry for instance, a "car" and "cdr" graph what, fundamental building blocks that I've been kept uninformed of? ;) every arrow in the system (including those in those two graphs) would have one arrow in each of those graphs corresponding exactly to the arrow's references bin: yep That does not exactly ease the process of understanding.. those two graphs are essential constructs they must exist in arrow, and they must apply to the whole system \esc I imagined a dropped system following a part of 'Arrow' goals. I miel: huh? Do you have any additional available info more than the paper? Anything written? bin: not yet I failed in one step : migrate from an absolute naming method to a full relative names matching. bin: like i said, major computer problems :/ miel: ah How you do that ? bin: well, if you could direct a guru to me who could handle all of my computer problems... bin: preferrable one in seattle heh I'd like to know more about the way an agent can identify a given arrow inside the system. my computers have all of my tools and docs on them miel: well, the arrow is specified semantically miel: by some typ of declaration water: Back to my earlier question; each agent is associated with one ontology, right? the declaration actually refers to the arrow in question, so that referring to it later on is a simple matter of accessing it through the def bin: no, an agent is just a coherent set of activities I was disapointed by the design of the textual I/F in your smalltalk model. It doesn't seem to be reflexiv. I hope this is a temporary mechanism for debugging ;) miel: it is miel: i'm going to release some more substantial code very soon :) good new! 01:20pm water: explain yourself, just a set of activities? Then the system defines the agent! the problem is that it is quite unfinished bin: i haven't worked it out entirely... I'd like to know more about 'car' and 'cdr' arrows. what do they mean ? bin: it's supposed to be an emergent property of a running system Are you referring to agents? Agents emerge from running systems? miel: in lisp, you access the first and second references of a list by applying the 'car' and 'cdr' primitive functions, respectively bin: yes Then in such case, I don't see how a human or even a function could "run" - access the system. -:- Zhivago [brian@th.merddin.com.au] has joined #tunes it wouldn't be hard... interfaces _are_ ontologies and ontologies can provide functional interfaces yes that's what I thought. But you said agents where *not* associated with an ontology, i.e. interface well, they aren't you can use multiple interfaces for the same task, for instance so how do they communicate? zhivago well, they place information in various places yo :D Do you planify an arrow which will act as a swich from an analysing step to a declarativ step, during the OS graph processing, ie an arrow which make the OS clone the graph the arrow is linked to? water: They need an ontology for that, I tell you! if the ui ontology picks up information automatically from a certain source and re-directs it to the screen... -:- td [x@1Cust132.tnt1.wilmington.nc.da.uu.net] has joined #tunes miel: i'm not sure exactly what you mean neither I miel: probably yes heh You should explain the process which handle a new arrow declaration. There 's a need. yes, i know i wish i had one :) i'm working on it though. it's probably the major reason for waiting to update the paper 01:30pm bin: any new groking problems? i have a few Agents need (at least) an ontology to place information anywhere. i would really like the jist of arrow bin: right, but i don't want to restrict them to one ontology I can imagine this process partly. My problem is to understand when you intend to make the system to switch from a 'pattern matching' step to an action step. There will be a special arrow, isn't it? fire: i don't have a jist on me right now. Wouldn't a set of ontologies just be another ontology? miel: probably. but i'd need to formalize it bin: you can join them, yes. someone give me the jist of arrow -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us734.javanet.com]) i can't get through that whole paper Then I don't see any need for more than one ontology per agent. But on you it seems like ontologies aren't really connected to agents? bin: most likely there will be union and intersection operators for them as well as the arrow operators fire: arrow =hll, ontology = meta-translator That seems to me to be going beyond the pure arrow concept s/hll/tunes hll/ bin: well, arrow logic does support it Is arrow logic something else you didn't tell me about? bin: eek abi accial? i heard accial was "A Crash Course in Arrow Logic" at ftp://ftp.phil.uu.nl/pub/logic/PREPRINTS/preprint107.ps.Z it's also in the paper, but it has a very small section or are you speaking about relations? arrow logic relates to relational algebra relates to? not the same thing? it's a modal logic, so it has operators no, it's slightly less expressive (to varying degrees) than undecidable relational algebra arrow logic is actually a family of related logics, though hmm... my ghostview isn't installed on this computer... darn How much papers on miscelleous subjects do you think ppl will need to read to understand arrows, really? Seems like you often point to some obscure file on the net (i have to use a friend's computer for net connections) hehe most of the subjects that arrow builds on are obscure except to certain groups of ai and logic researchers * Miellaby/#TUNES whishes he had been a mathematician. it's been very difficult to make it easy to grok for programmers 01:40pm (mmm... higher mathematics :) Does it need all strict and narrow theory to be understood? well, _I_ have to understand it to communicate it to you Arrows? Arrows are loading, atleast loading? that's a start and i have to understand it and reason using it (the theory) to give coders specs for programs for it Can't the idea be captured without all fancy terms? about programmer : I disagree. I'm a programmer. And yet, I think that -if only there was an exaustiv list of Arrow OS fundamental primitives- I could make my imagination works. well, i don't have that list at all and i don't have anything to build that list from no theory actually applies to my idea, just parts of it You have your computer at least :> * water/#tunes smacks binEng with his 30 books on higher math that apply to Arrow in one way or another. water: gimme that url for a crash course on arrow logic again please i've corresponded with various professors, and they've all recognized that they're papers help my idea, but that they can't help with whole arrow picture No-mathematicians are not fan of mathematical proving. You'll be able to write the theory latter. :). Just give us foods :) fire: scroll up, scroll up :) and you want everyone that tries to understand Arrows to read up on those subjects? oh funny that bin: hell no .Z?!?! fire: "uncompress" it :) uh how dare you version me :( muahaha I know your OS, haha haha plus i have nothing to view postscript with right now bin: i just need help with the aspects that don't directly relate to theory, so that i can make up that list of arrow primitives * binEng/#tunes dances around fire hehe * water/#tunes dances a jig. * binEng/#tunes stops, because someone said something about playing with fire if i could hand people the arrow primitives and say "go code 'em", i'd be set for life water: do think I should reread all of the paper? but i don't have them formally worked out yet bin: sure, it'll give you a nice headache :) :P no, really, you should it's not that bad