00:04:52 dajobe has quit 00:49:35 dmiles has quit 00:52:43 AaronSw has changed the topic to: Life, the Semantic Web and Everyone - http://www.w3.org/RDF/Interest/ 01:00:43 .email me@aaronsw.com to read // http://www.ietf.org/internet-drafts/draft-baker-http-resource-state-model-01.txt 01:00:44 email successfully sent. 01:02:36 sbp has joined #rdfig 01:10:28 * AaronSw disconnects 01:46:53 sbp has quit 02:09:40 sbp has joined #rdfig 02:10:24 hi sean' 02:10:35 Hi there 02:12:57 ... so logical operations like (and, or, not, => etc) things with truth tabels have their domain and range both on logical values .. agree? 02:13:30 * sbp looks around... me? 02:13:38 logical values usually boolian {true, fales} but might be other things too 02:14:17 yeah you ... dont think anybodty else is listening 02:14:24 :-) 02:14:46 Can you give me an example of use in N3? 02:15:07 ... i dont know th n3 vocabulary well enough 02:15:14 I'm not really sure what the range and domain of log:implies are 02:15:31 n3:StatementSet I guess 02:15:50 or, as TimBL would have us say: n3:Formulae 02:15:52 :-) 02:15:58 let's try model.n3 02:16:28 well i dont think log:implies is in the set of logical operations .. that is my point actually ... because log:implies does **not** have its range and domain as a logical value .. it's range and domain are contexts (or statements) 02:16:50 StatementSet according to model.n3 02:17:08 Formula according to log.n3 02:17:13 [[[ 02:17:13 log:implies a rdf:Property; 02:17:13 rdfs:comment "Logical implication. Understood natively by cwm."; 02:17:13 rdfs:range log:Formula; 02:17:13 rdfs:isDefinedBy <>; 02:17:13 rdfs:range log:Formula. 02:17:17 ]]] - http://www.w3.org/2000/10/swap/log.n3 02:17:35 so there must be some property that connects a logical value to a statement !!! <--- my point 02:17:42 s/range/domain/ 02:17:47 for one of the above... 02:18:12 well, there's a property that connects two statement sets right there, yeah 02:20:04 but the logical operation 'or' must join only other logical expressions or values ... a statement is not a logical value .. therefore there must be some other property arc to connect the logical operations to the statements for us to be able to use them on statements 02:21:02 { :Bob a :Male } :or { :Bob a :Female } . 02:21:11 they look like statements to me 02:21:30 (or rather, they are statements, hidden in context brackets) 02:22:00 [ ifTrue {:Bob a :male}] :or [ifTure {:Bob a :Female}] is what i say we must do 02:23:34 or more precisely .. dropping the colons ... 02:24:51 sbp has quit 02:25:11 [or [ifture {Bob a male}], [iftrue {Bob a female}], [iftrue {Bob a Asexual} ] ] a LogicalValue 02:28:22 sbp has joined #rdfig 02:28:39 did you catch my last? 02:28:42 [or [ifture {Bob a male}], [iftrue {Bob a female}], [iftrue {Bob a Asexual} ] ] a LogicalValue 02:29:58 I'd do it like... well, I was thinking of using a log:Truth type things, but that's quite inefficient 02:30:48 i dont care about efficienty here .. how do you do it with log:Truth ? 02:31:21 ({ { :Bob a :Male } a log:Truth } { { :Bob a :Female } a log:Truth }}) a :OR . # Is my best guess for now 02:33:13 of course, when you state something in RDF, then you're saying it's a log:Truth anyway, so that reduces to:- 02:33:15 thing is that { {bob a male} a truth} is still and all just a statement .. is is *not* a member of the set {true, fales} 02:33:31 ({ :Bob a :Male } { :Bob a :Female }) a :OR . 02:33:58 nope makes :or a class 02:34:02 well, how do you otherwise say something is false? You can't in RDF. We established that 02:34:12 yeah, OR can be a class 02:34:17 also you want something that nests on the subject 02:34:33 it's the class of lists for whom the members are a list of logical alternatives 02:35:43 like [ifTrue {bob a male} ] is an anomous node that can only be a logical value .. therefore you can nest it with other such nodes and make a logical and\or\not expression 02:36:42 the domain of 'iftrue' is a logical value and the range is a statement .. unless i got them switched 02:36:57 nope, that seems right 02:37:22 Hmm... that actually makes sense to me 02:37:48 no hang on, I'm thinking of it like an implication 02:38:26 [ :ifTrue { :Bob a :Male } ] => _:x :ifTrue { :Bob a :Male } . => _:x log:implies { :Bob a :Male } . 02:38:30 i got this recently form Jon Awbrey .. i could site the posts if you are interested .. seems as if it is grounded in the logical work of Quine 02:38:44 cite: yes, please :-) 02:39:34 a moment please 02:40:06 hang on, this isn't right:- 02:40:07 thing is that { {bob a male} a truth} is still and all just a statement .. is is *not* a member of the set {true, fales} 02:40:18 {bob a male} is clearly a member of the set true 02:40:43 and if you assert the whole thing, then you're saying that's true too 02:41:07 we're saying that """bob a male" is a truth" is a truth" 02:41:26 because the last bit can only ever be "is a truth" in RDF 02:41:52 so the only way to say that bob a female is false is to do """Bob a Female" a falsehood" a truth" 02:44:34 ... http://suo.ieee.org/email/msg07343.html the citations are at the bottom of the email 02:45:30 {bob a male} memberOf Statements 02:45:57 True memberOf LogicalValues 02:46:06 False memberOf LogicalValues 02:48:22 * sbp reads Awbrey's stuff 02:48:34 O.K., what is your question/comment again? 02:49:21 i have to noodel about how to meld this idea with the n3 idea that there is a class of true statements called log:Truth 02:49:42 yes :-) 02:50:43 the question is how to make nestable logical expressions using the classical logical operators of (and, or, not, =>, <=>) which can only operate on logical values 02:51:42 i think n3's proplem is that you couldnt make a complex logical expression in the log: vocuabulary 02:51:54 well, we have AND. We can make NOT using :F a log:Falsehood . We can do OR using the :OR class, or (heh) with NANDs, and we have predicates for => and <=> 02:52:26 yeah, you're probably right about that. But I don't think it's a problem; at least, it hasn't proven itself to be one yet 02:52:38 but my point is that they do not nest 02:54:02 give me something that I cannot convert into N3 02:54:23 s/something/something realistic/ 02:56:00 (=> A (or C D (and E G (not G)))) where the uppercase letters are statements 02:56:24 ooh, a challenge! :-) 02:56:51 (=> A (or C D (and E G (not G))) H) 02:57:09 => must have two operands 02:57:35 er... it already has two operands: A, and the rest of the bracketed stuff 02:58:04 oh .. yeah your right 02:58:37 :A log:implies [ a :OR; daml:first :C; daml:rest (:D {:E :G [ :not G ]}) ] . 02:59:07 [a :or] ????? 02:59:24 yeah, because I couldn't do () a :OR . 02:59:47 probalby cause n3 doesnt nest 03:00:31 but I met your challenge, did I not? 03:00:41 yeah, it is a bit odd... 03:02:04 [+ [* x, [/ y, 2] ] 5, g] .. is that legit n3 ? 03:02:29 nope... 03:02:51 note [] to eliminate the subjects and make ann nodes all of which represent numbers 03:03:05 how would you do it? 03:04:13 [ :add [ :multiply :x, [ :divide :y, "2" ] ], "5", :g ] . 03:04:19 pretty much the same :-) 03:04:51 fine, so just do the same with logical operators 03:05:19 yeah, and it comes out very much like yours. In fact, you just left a comma out 03:05:37 perhaps the operators could be added to N3. You 03:05:47 but you still have to have a property that maps a statement into a logical value .. course you could assume it is there and say the operations have a range that is either a logical operator or the truth of a statement 03:05:50 You'll have to beg Tim and convince him that it's worth using CWM as a calculator 03:06:57 hang on, my way wouldn't work 03:07:05 y/2 != 2/y 03:07:13 y-2 != 2-y 03:07:23 y*2 == 2*y 03:07:30 y+2 == 2+y 03:07:33 so... 03:07:41 things like entails, and log:implies that work on quoted statements are different than the logical operators that work on logical values and need to make nested expressions 03:07:42 [ :add [ :multiply :x, [ :divide (:y "2") ] ], "5", :g ] . 03:08:34 well we can leave the non commutative operations till we can have a way to order things 03:09:02 heh, "x" is short for putwhateveryouwanthere:x 03:09:12 (:y "2") being a sequenced bag? 03:09:16 yes 03:11:04 so my point .. and Awbrey's is that there are two different sets of relationships here (and, or, not, =>, <=>) and then the set that takes statements (log:implies, |-, |=) so if we want to be able to deal with both kinds of properties 03:11:09 these datatype discussion are weird, I think. I reckon that they may even put some implementors off... they certainly put me off 03:11:52 "if we want to be able to deal with both kinds of properties [stop]" I think you got cut off 03:12:21 log:implies == => 03:12:35 then we have to know the difference, cause they do operate on different things ... statements are not logical values 03:13:03 what *are* logical values, then? 03:13:28 the set (true, fales) if you want to be boolian about things 03:13:42 (1, 0) 03:14:18 well, you can do boolean logic if you want in RDF, but it's a bit boring 03:14:19 those are the only things that (and, or, ....) are defined on 03:15:06 we need boolian operators to be able to express complex rules 03:15:37 we have them, as classes, surely? THe class of things that are true, the class of things that are false 03:17:54 but the only thing you can do with a class is state that a individual is a member of it ... how do you get a boolian value (0,1) out of that ?? 03:17:57 and to represent all things that are true: ?x a log:Truth . 03:18:22 well, make a boolean value then - :True 03:18:40 :BooleanTrue 03:18:46 :BooleanFalse 03:19:25 hense: [ifTrue {bob a male}] a :BooleanTrue 03:20:18 I still don't get the "ifTrue" predicate. But I'll take your word for it that it does what it's supposed to do :-) Can you provide a synonym? 03:20:38 isTrue 03:20:54 is asks the yes or no question, is the statement true or not 03:21:18 so it sounds like a reverse implication then, to me 03:21:37 if bob is indeed a male, then the subject of the triple (the anonymous subject) is true... yes? 03:21:49 so the subject of the statement [isTrue {bob a male]] is a boolina value if bob has a Y choromsome 03:22:18 O.K., I think I get it now 03:22:22 yes the anonous subject represents a boolian value 03:22:45 So, you can say :BooleanTrue :isTrue { :Bob a :Human } . 03:22:51 (if Bob is indeed a human) 03:23:01 what if I say:- 03:23:07 :BooleanFalse :isTrue { :Bob a :Human } . 03:24:02 it might need to go the other way ... ie its not a symetric property 03:25:05 hmmm .. no i think you had it right 03:25:12 what does this mean: :BooleanTrue :isTrue :BooleanFalse . 03:25:41 its a category error 03:26:00 :isTtru domain TruthValue 03:26:05 { :BooleanTrue :isTrue :BooleanFalse } a :CategoryError . 03:26:09 :isTrue range Statement 03:26:51 ooh, a schema inconsistency. O.K., I grok that 03:27:03 actually :CatagoryError is a perfectly good :LogicalValue .. so 03:27:03 how about... 03:27:50 nope, I think I'm getting there 03:27:53 :CategoryError :isTrue {:BooleanTrue :isTrue :BooleanFalse} 03:28:02 heh, cool 03:28:19 which is what i used to call trancedental logic 03:28:52 ah, perhaps: :isTrue daml:inverse :truthType . :truthType rdfs:subPropertyOf rdf:type . 03:29:28 i doubt it 03:29:42 yeah, it'd make :BooleanTrue a class again 03:30:01 rdf:type an :isTure are totally different properties 03:30:09 can you give another example of this being used? 03:30:31 the class woud be :TruthValues .. but it is more of a set than a class 03:30:58 or do you call a set a class ? 03:31:00 :TruthValues daml:oneOf (:BooleanTrue :BooleanFalse) . 03:31:14 that looks right 03:31:47 thing is what is the relationship between :BooleanTrue and log:Truth 03:32:09 I've got no idea :-) 03:33:12 { ?X a log:Truth} log:implies {[:isTrue ?X]} 03:33:34 where ?X is any statement 03:33:40 ooh 03:34:12 that does make it seem that ":isTrue daml:inverse :truthType . :truthType rdfs:subPropertyOf rdf:type ." 03:34:57 is there a real :truthType in some real namespace? 03:35:10 nope, I just made it up to fit the occasion 03:35:48 use :truthType in a statement excpet for the one above 03:35:54 :truthType a rdf:Property; rdfs:domain log:Formula; rdfs:range :TruthType . :TruthType daml:oneOf (log:Truth log:Falsehood) . 03:36:12 { :Bob a :Male } :truthType log:Truth . 03:36:35 Hmm... odd grammar 03:36:38 { :Bob a :Male } :truthType log:True . 03:37:31 i think your saying that :isTrue subProperty :truthType ... are you? 03:37:42 I'm saying that it's the inverse 03:37:48 :isTrue daml:inverse :truthType . 03:38:10 so: { ?x :truthType ?y } log:implies { ?y :truthType ?x } . 03:38:29 oic 03:40:59 i was thinking of menotgraphing the complete log: namespace 03:41:16 wow, that would be a task and a half. There are quite a lot of terms in there 03:41:24 but it would still probably be a useful thing to do 03:41:41 well i was just thinking of it 03:42:55 its the relationship between log: and KIF and CycL that interestes me ... these all should form a single coherant vocabulary .. but everyone is fighting that :( 03:43:05 yeah... 03:43:33 thing is that everybodies logics want to differ .. and they should be able to differ 03:44:20 nephrael has joined #rdfig 03:44:29 * sbp waves to dmiles 03:44:35 hey there 03:44:48 so to allow people to specify the kinds of logics that apply to theire epistles, we need to get into the smenatic kind of operators .. like (|- , |=, log:implies) .. and be able to specify what kind of logic applies in some particular context 03:45:12 yeah, I agree with you there 03:46:10 but it's all "semantics of semantics" discussions... it's really tricky to follow, and makes it a nightmare to implement stuff 03:46:30 we just need standardization, and quickly. The SW evolves at a snail's pace at the moment 03:46:35 also each context should have it's own universal quanification.. but i am sure you have already been there 03:47:02 yeah 03:47:23 the problem with ?x is that you can't get it to mean: { this log:forAll :x . :x :y :z } . 03:47:55 what does that mean? 03:48:18 it means that the context is true for all x 03:48:24 ive been drawing the quantification between the variable and a variable context 03:48:34 (forall ?X ( holds ?x ?y ?z)) 03:48:45 yeah 03:48:56 ( but case sensitive ) 03:49:13 err insensitive 03:49:26 is that the sumo:holds ??? is that the same property ? 03:49:36 yes 03:49:52 [[[ 03:49:53 consider: (forall (?p) (=> (forall (?x) (likes ?x ?p)) (type ?p GoodGuy)). 03:49:53 i.e. if everybody likes P, then P is a good guy. 03:49:53 that's *very* different from (forall (?p ?x) (=> (likes ?x ?p) (type ?p GoodGuy))) 03:49:54 ]]] 03:50:05 the predicate is best always understood as a 1st order problem of holds 03:50:43 the latter is "if anybody likes P..." 03:51:48 the quanifier arround ?p in eigfther case means that thatis axiom is all subsumming 03:52:11 its more to close the well formmed formula then anything 03:52:45 by being all subsumming it means that the contropositive is definately deducable 03:53:13 what does ( holds ?x ?y ?z) mean ? 03:53:29 (?y ?x ?z), I guess 03:53:42 well could mean all things known.. for all giving facts... 03:53:59 oh heh, sorry... 03:54:08 what does (forall ?X ( holds ?x ?y ?z)) mean? 03:54:13 yes its just so we den't in CycL/KIF get it confused with a reifiable function 03:54:35 what things do the ?x ?y ?z bind with? 03:55:25 (assignmentFn ?x ?y ?z) if opposi8te of (holds ?x ?Y ?Z) 03:55:51 inh the first case ?x is a bound to a fountion.. in the sencond case bound to a predicate 03:55:58 err function 03:57:05 i can't parse: "(holds P N1 ... NK) is true just in case the tuple of objects denoted by N1,..., NK is an element of the &%Relation P.") 03:57:06 in sumo/suo it depends on the domain of what (?x ?Y ?Z) is used in 03:57:46 i grok the tupel, but what does 'being an element of the relation' mean 03:57:50 it is legal in suo/sumo to say (forall (?x ?Y ?Z)(insance (?x ?y ?Z) Entity)) 03:58:35 sean, were talking about 03:58:38 http://ontology.teknowledge.com:8080/rsigma/SKB.jsp?req=SC&name=holds&caseSensitive=on&skb=Merge 03:58:39 A: http://ontology.teknowledge.com:8080/rsigma/SKB.jsp?req=SC&name=holds&caseSensitive=on&skb=Merge from SethR 03:58:56 (and (forall (holds ?x ?Y ?Z)(insance (?x ?y ?Z) Entity)) (forall (holds ?x ?Y ?Z)(insance (assingmentFn ?x ?y ?Z) Entity)) 03:59:05 oops 03:59:09 A:| What does (forall ?X ( holds ?x ?y ?z)) mean ? 03:59:09 titled item A 03:59:31 (and (forall (?x ?Y ?Z)(insance (holds ?x ?y ?Z) Entity)) (forall (holds ?x ?Y ?Z)(insance (assingmentFn ?x ?y ?Z) Entity)) 03:59:50 still oops but i think you know what i mean 04:00:20 holds just disambiguates the syntax of (?x ?Y ?Z) 04:01:00 i am going to statr prefixing sumo:holds 04:01:27 but what does 'being an element of the relation' mean ?? 04:01:41 * sbp finds some really cool new properties 04:01:46 element of the tuple? 04:02:32 elelmet of the relation 04:02:37 oh Ians text 04:03:01 he shouold not have said element.. then he was implying sets 04:03:20 he should have said "respective arguments" 04:03:23 what is an element of a relation ? 04:03:34 in my opinion nothing is 04:03:54 it doesnt apply in any theory i know of 04:03:54 well what should he have meant? 04:04:04 { :y a :AntisymmetricRelation } log:implies { this log:forAll :x , :z . { :x :y :z . :z :y :z } log:implies { :x daml:equivalentTo :y } } . 04:05:09 rdf:type a :AntisymmetricRelation . 04:05:21 is it really? 04:05:32 he should have said each element corisponds to an agument in a tuple 04:05:52 it is .. but not really nessarily to assert it 04:06:05 wait 04:06:20 does this have anything to do with the natural language word 'holds' ? 04:06:38 :Class .a :Class 04:06:52 { :Class a :Class } 04:07:05 you i guess so 04:07:10 err yeah i guess so 04:07:19 ah, so it is 04:07:45 heh, that's pretty cool... 04:07:52 im lost 04:07:54 by being declared antisymemtric yuou can safely skip some inference backchains 04:09:16 nolthing toi do with natural language holds unless by coincidence 04:10:06 now I'm lost :-) 04:10:16 but there is some ambiuguity in holds.. means holds true"? holds unrefuted? holds always? 04:10:43 nessarily holds? unnessarlyu holds sometimes? 04:11:08 let me ask it differently 04:11:22 holds = "it is true that there is a statement with this subject, predicate, and object" 04:11:36 hehe sbp... exactly 04:11:42 if i want to say that a set of rules can be applied to a specific context .. how do i do it?? 04:12:26 there is no public releaization of context in SUmo unfortuantly 04:12:43 internally i use (ist ?Context (holds ?Rel ?arg1 ?arg2)) 04:12:44 i thojugh Adam just added it 04:13:09 i'll loook treasl quick 04:14:06 if so i can stop ussing cyc structures 04:15:32 http://suo.ieee.org/email/msg07262.html 04:15:32 B: http://suo.ieee.org/email/msg07262.html from SethR 04:15:53 B:| is this the start of contexts in SUO or not ? 04:15:53 titled item B 04:15:56 * sbp works on ReflexiveRelation 04:17:04 B: yes but it is just comment based 04:17:04 commented item B 04:17:13 B: ;; INCLUDES 'STRUCTURAL ONTOLOGY' 04:17:14 commented item B 04:17:39 Oh .. YAY!!! 04:18:01 good god.. finally 04:18:16 B: i am wrong .. Yay! 04:18:17 commented item B 04:18:21 so does (ist MyChiggyContext (holds instance ?X Entity)) say that everything in MyChiccgContext is an Entity ? 04:19:12 well looks loike asserted has replaced ist 04:19:19 yay! 04:19:29 nephrael is now known as dmiles 04:19:45 hrrm didn't notice i was ussing my efnet nick 04:20:27 B: 'asserted' is like cyc's 'ist' 04:20:27 commented item B 04:20:42 Well, I'd better be going now. If you have a suggestion for SWN, you know where to mail me :-) 04:20:45 Gotta run 04:20:51 i have to run as well 04:21:15 so how do we say that all the rules (asserted statements) in some context are asserted in another context ? 04:21:36 is that 'includes' ? 04:21:49 i guess but you could just say... 04:22:22 (=> (asserted ParentOnto ?X) (asserted ChildOnto ?X)) 04:22:35 sbp has quit 04:24:39 well thanks dmiles .. you cleard up a lot .... i got to go pay attentions to the better half now :) 04:25:20 SethR has quit 04:42:14 * AaronSw reconnects 04:57:47 mnot has joined #rdfig 05:38:34 tav` has quit 05:56:26 * AaronSw disconnects 07:42:09 mnot has quit 10:28:39 oierw` has quit 10:29:42 oierw` has joined #rdfig 10:41:04 oierw` has quit 10:41:18 oierw` has joined #rdfig 13:30:40 dajobe has joined #rdfig 14:06:48 dmiles has quit 15:07:23 sbp has joined #rdfig 15:10:37 sbp has quit 15:47:48 sbp has joined #rdfig 15:48:24 http://lists.w3.org/Archives/Public/www-rdf-interest/2001Nov/0175 15:48:25 C: http://lists.w3.org/Archives/Public/www-rdf-interest/2001Nov/0175 from sbp 15:48:27 chump? aw, man... 15:48:39 take your time :-) 15:48:54 C:|Re: Namespaces wihtout "#" Was: Few CWM Bugs 15:48:54 titled item C 15:49:12 C:Proposal to end the "what is identified" discussions; a new HTTP header 15:49:13 commented item C 15:49:52 C:"Entity-Type: Resource" - the data returned by the HTTP transaction *directly* corresponds with the resource requested 15:49:53 commented item C 15:50:26 C:"Entity-Type: Representation" - the data returned by the HTTP transaction is an indirect representation corresponding to the resource requested 15:50:26 commented item C 15:53:46 C:This has the cool advantage that you can add it to an (X)HTML file: 15:53:46 commented item C 15:54:30 C:Hmm... and perhaps N3: <> http:Entity-Type "Resource" . 15:54:30 commented item C 15:56:22 sbp has quit 16:47:44 * AaronSw reconnects 16:48:32 aaron, you there? 16:49:12 or is it just your client :) 16:49:26 yep 16:49:38 this time it's me :) 16:50:45 what kind of URI would you use to refer to a person? 16:50:52 (in RDF, of course) 16:51:20 Hmm. Well, if sandro would finish up and get tag: approved, I'd use it. 16:51:27 http://www.taguri.org/ 16:51:46 I wouldn't. I'd relate them to other URIs for less politically sensitive types of thing. Mailboxes, homepages, etc etc. 16:52:08 politically sensitive? 16:52:09 ok, so tag is one alternative 16:52:25 danbri, so you would use the person's home page to identify them? 16:52:33 or their mailbox? 16:52:43 danbri would probably also use the hash of the person's code... oops, I've said too much. 16:53:02 larsbot, only indirectly: i.e. [ a :Person ; foaf:homepage ] 16:53:17 C:Interesting idea, but I can't really imagine that anyone would want to say Entity-Type: Resource -- then they'd never be able to change the page! 16:53:17 commented item C 16:53:29 ah, "would relate them to"; I missed that 16:53:49 but what are the alternatives to tag URIs? 16:54:04 what would danbri use as the URI of the person resource? 16:54:42 He'd use an "anonymous" node, which on most days doesn't have a URI... 16:54:59 Well, there are other proposals simialr to tag -- PTS, NSL, etc. There's the centralized urn:pin system... 16:55:04 but some days they do? :-) 16:55:22 Yeah, my system only works on those days. :-) 16:55:33 right, but in general the RDF opinion is: don't use URLs? 16:55:35 ambient has joined #rdfig 16:56:35 Well.... that's one of the things we're still arguing about. Theoretically using a URL would be fine as long as the dereferenced page makes it clear that the resource involved is the person, not the document. 16:56:55 C:Nor would they be able to use content negotiation, I think... 16:56:55 commented item C 16:57:09 C:Perhaps this would be better if TimBL clearly defined what a document is. 16:57:10 commented item C 16:57:16 oh, that's interesting, because it's exactly the same as the idea of subject indicators in topic maps, which people think is one of the main XTM/RDF differences 16:57:47 sbp has joined #rdfig 16:58:16 but this is only in theory? what does that mean? 16:58:51 Well, it means that TimBL only likes this idea if the URL has a '#' in it. 16:58:59 See sbp's proposal in C for a little more. 16:59:19 And that Roy Fielding, myself, and some others only like this idea if the URL *doesn't* have a '#' in it. 16:59:43