Available formats: content-negotiated html turtle (see SIOC for the vocabulary)
Back to channel and daily index: content-negotiated html turtle
These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.
| 05:25:26 | <mhausenblas> | good morning Web of Data |
| 07:54:25 | <Wikier> | http://realtech.burningbird.net/semantic-web/rdf-and-rdfa/wolfram-alpha-what-rdf |
| 07:54:34 | <Wikier> | A:|Wolfram Alpha: What is RDF? |
| 08:04:53 | <KjetilK> | ACTION isn't too impressed with Wolfram Alpha either |
| 08:05:24 | <KjetilK> | I did some searches on more obscure mountains, and it returned very little of interest |
| 08:05:26 | <mhausenblas> | +1 |
| 08:06:13 | <KjetilK> | I tried the difficult question of "the highest mountain on earth outside of Himalaya-Karakoram", then just outside of Himalaya, then just outside of Asia, then just on earth |
| 08:06:38 | <KjetilK> | got no response to any of them |
| 08:09:38 | <KjetilK> | they have probably spidered Wikipedia quite a lot though, but we allready have that... |
| 08:40:35 | <melvster_> | facebook becomes largest openid relying party: http://www.techcrunch.com/2009/05/18/facebook-becomes-largest-openid-relying-party/ |
| 08:44:14 | <drewp> | i look forward to working on the oss version of wolfram alpha, though |
| 08:44:41 | <drewp> | where we can write our own very special purpose types of result boxes, and more importantly, incorporate our personal data in the result boxes |
| 08:45:08 | <drewp> | e.g. "ed" -> "here's a graph of how many emails you exchanged with ed over the last few years" |
| 08:46:05 | <melvster_> | drewp: you'd probably want that client side |
| 08:46:38 | <drewp> | i'm not sure what 'client side' is, but yes i would run an instance at my house |
| 08:47:26 | <IanDMclean> | How do I say that rdf is a dialect? |
| 08:50:21 | <idmclean> | I assume there is a class of thing defined in RDF which represents RDF |
| 08:51:43 | <Shepard> | why a class? what would be its instances? |
| 08:52:48 | <idmclean> | N3 notation, turtle, rdfxml? |
| 08:53:24 | <Shepard> | oh you mean the class of formats, representations |
| 08:53:30 | <Shepard> | dunno |
| 08:53:58 | <idmclean> | I'm writing the definition for a Language of Languages(aLoL) |
| 08:54:01 | <drewp> | i don't know which format 'rdf' is, if 'rdfxml' is something separate |
| 08:54:09 | <Shepard> | you can get identifiers for the individual formats quite easily, not sure if anyone defined that class |
| 08:55:09 | <idmclean> | In aLoL, there are Languages and Dialects. Dialects are subclasses of Languages. RDFXML is a subclass of Dialect. |
| 08:55:41 | <idmclean> | More generally, RDFXML, Turtle, N3, are all subclasses/instantations of RDF. |
| 08:56:36 | <idmclean> | Shepard, I need an explination of Class and Instances in this context. |
| 08:56:38 | <drewp> | it seems very confusing to call RDF a language, then. What are some other examples of languages out there? |
| 08:56:57 | <Shepard> | I don't agree with that, a dialect of RDF would be a variation of the data model, not a concrete syntax |
| 08:57:29 | <idmclean> | RDF is a Dialect of aLoL |
| 08:57:31 | <drewp> | if i have some zsh, tcsh, bash, and ksh, do you say those are all dialects of "shell" or something? |
| 08:57:35 | <Shepard> | apart from that: http://en.wikipedia.org/wiki/A_language_is_a_dialect_with_an_army_and_navy |
| 08:58:19 | <Shepard> | idmclean: did you talk to KickAssClown? they wanted to define programming languages in RDF as well |
| 08:58:40 | <drewp> | Shepard: i think they've met :) |
| 08:58:58 | <idmclean> | I am KickAssClown. ^_^ |
| 08:59:23 | <Shepard> | oh |
| 08:59:48 | <idmclean> | An example is Scheme is a Dialect of LISP |
| 09:00:06 | <idmclean> | N triples is a dialect of N3 |
| 09:00:33 | <drewp> | ok that lisp one i can understand. They're both languages, and they have a strong relationship with their syntax and other features |
| 09:00:38 | <idmclean> | N triples is also a dialect of Turtle. |
| 09:01:32 | <Shepard> | then dialect seems like a property to me :) |
| 09:01:56 | <Shepard> | :NTriples alol:dialect :Turtle . |
| 09:02:10 | <idmclean> | A language in my mind is effectively a collection of closely related dialects using the same or similar symbols (semiotics?). |
| 09:03:13 | <Shepard> | I strongly discourage you from making a distinction between dialect and language, it doesn't work for human languages :) |
| 09:03:25 | <melvster_> | http://lists.foaf-project.org/pipermail/foaf-protocols/2009-May/000599.html |
| 09:03:31 | <idmclean> | I'm mainly concerned with Formal languages. |
| 09:03:38 | <melvster_> | B: Web Access Control demo and source code |
| 09:03:52 | <Shepard> | idmclean: still, you need to clearly define the difference |
| 09:04:23 | <idmclean> | The difference is simple, a dialect is a subset of a larger set: a langauge. |
| 09:05:00 | <Shepard> | are Java and C++ dialects of one language? |
| 09:06:28 | <Shepard> | i.e. what is in the set of languages that is not in the set of dialects and how do you define that difference? |
| 09:06:50 | <idmclean> | Java and C++ might be dialects of C or Assembly, but at this point the model I'm building isn't intended to be used generally. |
| 09:07:01 | <drewp> | Suppose you don't know about a common parent so you call X and Y two languages. Later, you learn about the common parent, so do you change X and Y from language to dialect? |
| 09:07:05 | <Shepard> | and if you say that N-Triples is a dialect of N3 and of Turtle then are N3 and Turtle languages? |
| 09:08:03 | <idmclean> | N3 and Turtle are Languages in the context of examining N-Triples. |
| 09:08:43 | <tobyink> | But Turtle is a dialect of Trig. |
| 09:08:58 | <idmclean> | as in trigonometry? |
| 09:09:01 | <Shepard> | ah, contexts come into play, one of RDF's hardest problems ;) |
| 09:10:28 | <idmclean> | The trick I'm trying to play here is that a language isn't strictly defined. A language is an abstraction of syntax, morphology, semantics, grammar etc. |
| 09:11:06 | <idmclean> | Dialects are implementations defining the specifics of the language. |
| 09:11:53 | <melvster_> | i think normally you'll say one language is equivalent to another if there is a regular expression that converts back and forth |
| 09:12:26 | <idmclean> | melvster_, yes. I would agree with that. |
| 09:13:05 | <idmclean> | I would say that Java and C++ are part of a larger language and share common dialects. |
| 09:18:39 | <idmclean> | Anyway, is there a thing in RDF/Turtle/N3 which represents the set of things which is RDF? |
| 09:22:11 | <idmclean> | Can I say "The thing at this URI represents RDF which is Dialect"? |
| 09:22:32 | <idmclean> | *which is a Dialect |
| 09:22:49 | <idmclean> | if so, how? |
| 09:25:27 | <idmclean> | drewp, I wouldn't call X or Y languages, I would call them a Dialect. If they have a common parent, it is a language which they can be said to be a subset of. I don't know what the overall language/languages are, but I do know that the programming "languages" I am going to be dealing with are subsets of larger "languages". |
| 09:26:40 | <kwijibo> | <http://dbpedia.org/resource/Resource_Description_Framework> a :Dialect . |
| 09:31:54 | <tobyink> | I have stuff to say on this topic but am on phone. |
| 09:39:17 | <idmclean> | for simplicity, let say I have a Language, A, and two dialects, B and C. the Language is defined as everything that is in B and C, but nothing which is in both B and C. |
| 09:41:06 | <idmclean> | I don't know what A is. |
| 09:44:29 | <idmclean> | I suppose A is the Universe that B and C occupy. |
| 09:52:57 | <Shepard> | I'm still not comfortable with using the terms dialect and language. how about Java and C++ are languages of a language family? |
| 09:53:17 | <Shepard> | WP: A language family is a group of languages related by descent from a common ancestor, called the proto-language of that family. |
| 09:55:30 | <idmclean> | yuck |
| 09:56:30 | <idmclean> | Shepard, Java and C++ are in the same conceptual space for the model because of the commonalities of their grammar, semantics, and structure. |
| 09:57:43 | <idmclean> | If you were to remove all the conflicting portions of Java and C++, you could merge them. |
| 09:58:02 | <idmclean> | They effectively, would be a new language if you did so. |
| 09:58:32 | <tobyink> | Something like this? http://esw.pastebin.com/f75d76f16 |
| 10:01:37 | <idmclean> | tobyink, yes. |
| 10:04:06 | <tobyink> | That way you can have :someFunction :written_in :Python . :written_in rdfs:subPropertyOf rdf:type. :Python rdfs:subClassOf :ProgrammingLanguage . |
| 10:05:01 | <idmclean> | tobyink, definitely what I was trying to express. |
| 10:06:54 | <tobyink> | rdfs:Class/owl:Class is useful beyond just being fodder for rdf:type to point at. |
| 10:14:39 | <idmclean> | I fail to comprehend what rdfs:subPropertyOf [ owl:inverseOf rdfs:subClassOf ] . is expressing. |
| 10:15:36 | <swh> | I think it means ":superClass", but why you need that I have no idea |
| 10:18:21 | <Kerwork> | heh, what an unfortunatly named channel |
| 10:19:56 | <Shepard> | Kerwork: it's actually the channel of Alcoholics Anonymous, in disguise |
| 10:20:49 | <Kerwork> | in disguise? I thought it was quite obvious, why do you think I am here? :P |
| 10:21:10 | <Shepard> | ;) |
| 10:22:34 | <Kerwork> | semantic web is also interesting tho ;) |
| 10:22:36 | <tobyink> | idmclean: rdfs:subPropertyOf [owl:inverseOf rdfs:subClassOf] is an unfortunately necessarily long-winded way of saying rdfs:subPropertyOf rdfs:superClassOf . |
| 10:23:07 | <tobyink> | Because rdfs:superClassOf is not defined. |
| 10:24:30 | <idmclean> | tobyink, dialect is then defined as a superClassof? |
| 10:24:37 | <tobyink> | It effectively means that :dialect and :serialisation point from a language to a superclass of that language. |
| 10:24:44 | <tobyink> | Which is probably wrong. |
| 10:24:51 | <tobyink> | Let me have another go. |
| 10:25:24 | <tobyink> | No, it's right. |
| 10:25:32 | <idmclean> | I'm pretty sure N3 is the superclass of Turtle rather than the otherway around . ^_^ |
| 10:26:08 | <tobyink> | { :N3 :dialect :Turtle . } => {:Turtle :subClassOf :N3 . } |
| 10:26:56 | <tobyink> | And { :N3 :dialect :Turtle } can be written { :Turtle is :dialect of :N3 } using the "is ... of" notation. |
| 10:27:46 | <idmclean> | tobyink, may I use your code as the basis for alol? |
| 10:27:58 | <tobyink> | Yes - sure. |
| 10:29:19 | <tobyink> | Still not sure about definition of :dialect, :serialisation. Difficult to get it straight in my head. |
| 10:30:12 | <tobyink> | Because rdfs:subClassOf is kinda backwards to begin with. |
| 10:30:42 | <tobyink> | It's right on pastebin. I'm sure it must me. |
| 10:30:49 | <tobyink> | s/me/be/. |
| 10:32:52 | <tobyink> | Really though :RDFXML is :serialisation of [ rdfs:subClassOf :RDF ] . |
| 10:33:08 | <tobyink> | Because not all RDF graphs may be represented in RDF/XML. |
| 10:37:56 | <idmclean> | Why owl:Class instead of rdfs:Class? |
| 10:40:23 | <tobyink> | Using owl:Class instead of rdfs:Class means that OWL Full reasoning might not be needed. OWL Full is computationally more expensive than OWL Lite or OWL DL. |
| 10:40:51 | <tobyink> | (I'm pretending I know the difference between owl:Class and rdfs:Class.) |
| 10:41:18 | <tobyink> | slightly longer version - http://esw.pastebin.com/f3bd77ad6 |
| 10:42:32 | <tobyink> | Oh, and let's include a definition for written_in - http://esw.pastebin.com/d2975d9e4 |
| 10:47:34 | <tobyink> | http://esw.pastebin.com/d4c41bb0d |
| 10:48:25 | <tobyink> | C:|My proposed representation of programming languages, their relationships with each other, and their relationships with programs in OWL. |
| 10:49:05 | <idmclean> | logger, pointer. |
| 10:49:05 | <logger> | See http://chatlogs.planetrdf.com/swig/2009-05-19#T10-49-15 |
| 10:50:15 | <idmclean> | http://web.cs.mun.ca/~ulf/pld/index.html |
| 10:50:55 | <tobyink> | C: The :Language, :ProgrammingLanguage and :DataLanguage classes are perhaps badly named. They represent the class of all things written in a language. So :RDF is not a :DataLanguage, but rather :RDF is a subclass of :DataLanguage. |
| 10:51:27 | <idmclean> | D:| A Programming Language Design Reference by Ulf Schünemann |
| 10:52:22 | <Count-Duckula> | predicate to say 1 subject is related to another? |
| 10:52:43 | <Count-Duckula> | or 'similar to' |
| 10:53:05 | <tobyink> | There's http://open.vocab.org/terms/similarTo . |
| 10:53:33 | <tobyink> | By "subject" do you mean as in "subject predicate object" or as in "topic". |
| 10:53:55 | <tobyink> | If the latter, then SKOS probably provides you with better tools that ov:similarTo. |
| 10:54:47 | <tobyink> | For example, there's skos:broader to point from a concept to a broader topic e.g. Algebra to Mathematics . |
| 10:55:00 | <tobyink> | skos:narrower, skos:close, etc. |
| 10:56:39 | <tobyink> | idmclean: A while back I wrote a basic scripting language with its tokeniser, parser, interpreter in PHP. |
| 10:57:23 | <danbri> | also lots of data expressed in skos |
| 11:00:34 | <mhausenblas> | ping LeeF |
| 11:01:21 | <mhausenblas> | how comes http://iswc2009.semanticweb.org/wiki/index.php/ISWC_2009_Semantic_Web_In_Use_Track is sort of orphaned (incl. PC) one month before submission deadline ;) ? |
| 11:01:24 | <idmclean> | tobyink, the parser, interperter, compiler(?), and profiler for this is going to be very weird to develop. |
| 11:01:58 | <mhausenblas> | (to be fair, the same is true for the research track ...) |
| 11:02:50 | <tobyink> | idmclean: Well, you can skip a lot of the parsing, as you can just use an off-the-shelf RDF parser. The rest will be challenging though. |
| 11:03:33 | <tobyink> | With most languages you build an "abstract syntax tree", but because of RDF's nature you'll end up with an "abstract syntax web". |
| 11:03:50 | <idmclean> | tobyink, depends on what this looks like once I introduce another dialect. |
| 11:04:08 | <tobyink> | Most of the standard text books on compiler design will abandon you at that point. |
| 11:04:21 | <idmclean> | tobyink, exactly. I'm building a language which is a graph rather than a tree or a line. |
| 11:05:27 | <idmclean> | tobyink, I'm hoping the compiler, interperter, and profiler gurus will hear me out once I reach that point. |
| 11:08:29 | <Count-Duckula> | tobyink: It's more specific <www.example.com/123123> relatedTo? <www.example.com/123123> ... something like that |
| 11:08:45 | <Count-Duckula> | different URI's obv |
| 11:09:53 | <tobyink> | As I said, there is <http://open.vocab.org/terms/similarTo> but that's very general, with a very "fluffy" definition. |
| 11:10:40 | <Count-Duckula> | tobyink: Thanks, looks about right :-) |
| 11:11:10 | <Count-Duckula> | lol there is some very random terms in ov |
| 11:11:27 | <tobyink> | I use it occasionally as an alternative to owl:sameAs when the match has been made by a script and not manually verified. |
| 11:12:26 | <tobyink> | But the definition of "similar to" is quite vague. For example, I am similar to you in that we're both (and I'm making some assumptions about you here...) humans, with heads, arms and legs. |
| 11:12:51 | <tobyink> | So are we ovterms:similarTo each other? |
| 11:13:26 | <tobyink> | Am I similar to my desk because we're both solid objects occupying roughly the same geographic co-ordinates? |
| 11:14:36 | <tobyink> | Very vague. I'd generally recommend only using ovterms:similarTo when you can't find anything more specific. Of course, it's easy enough to define your own more specific term and make it an rdfs:subPropertyOf ovterms:similarTo. |
| 11:16:21 | <Count-Duckula> | ov:Friend of a Fro -- got to find a way to use that :-) |
| 11:24:18 | <Count-Duckula> | tobyink: Thanks makes sense |
| 11:30:26 | <mhausenblas> | phenny, tell LeeF see http://chatlogs.planetrdf.com/swig/2009-05-19.html#T11-01-31 |
| 11:30:26 | <phenny> | mhausenblas: I'll pass that on when LeeF is around. |
| 11:30:34 | <mhausenblas> | ACTION off for lunch |
| 11:31:56 | <tobyink> | iand: 404 error at http://purl.org/vocab/bio/0.1/ |
| 11:32:26 | <tobyink> | Though individual properties and classes deref OK. |
| 11:35:27 | <tobyink> | The example pages and changelog 404 too. |
| 11:35:56 | <tobyink> | phenny, tell iand <http://chatlogs.planetrdf.com/swig/2009-05-19.html#T11-32-06> |
| 11:35:56 | <phenny> | tobyink: I'll pass that on when iand is around. |
| 11:38:49 | <thomase> | tobyink: looks like "0.1.html" is the correct file?! |
| 11:39:23 | <tobyink> | Yes - I was able to find it eventually, but the link still needs fixing. |
| 11:39:46 | <thomase> | yes... |
| 11:42:32 | <Anchakor> | idmclean: hi |
| 11:43:02 | <Anchakor> | ACTION is back from school from physics exam he just screwed up |
| 11:43:48 | <Anchakor> | idmclean: If I were you I'd think dialect = subLanguage |
| 11:44:08 | <tobyink> | Anchakor: oh well, there's always the glamorous world of environmental health to fall back on. |
| 11:44:08 | <Anchakor> | similar as subClass |
| 11:44:39 | <idmclean> | Anchakor, Hi. I am using the term Dialect in this context to refer to a sub-language of some unknown larger language. My assumption is that all languages are a subset of a universal language. |
| 11:46:15 | <Anchakor> | idmclean: maybe, all rdfs:Classes are subClasses of rdf:Thing |
| 11:46:26 | <tobyink> | idmclean: in my model, all the classes are subclasses of :Language which is meant to represent all the languages that have been, will be, or could be written, spoken, programmed in or thought about. |
| 11:46:57 | <tobyink> | And in fact, even the languages which have not been written, spoken, programmed in or thought about. |
| 11:47:00 | <idmclean> | Based on what I found out about BNF and vW-Grammars, I've come to the conclusion that there is a more powerful language than the one I use and know of. |
| 11:48:05 | <Anchakor> | idmclean: but certainly I wouldn't think of rdf as a language - it isn't, it's just a model |
| 11:48:07 | <idmclean> | tobyink, I had noticed which is why I agreed that your implementation would seem to be correct in spirit if not correct in detail. |
| 11:49:10 | <idmclean> | Anchakor, I would agree with you about the reality of it, but in aLoL, any expression/symbol/statement/object is a linguistic construct. |
| 11:49:49 | <idmclean> | rdf is sufficiently complex as to be a set of rules defining an abstract family of expressions. |
| 11:49:53 | <tobyink> | Anchakor: you just need to be flexible about what you define as a language. e.g. English is a language with various dialects (American English, Australian English, *proper* English, ...) and various serialisations (written, spoken, etc). |
| 11:50:11 | <idmclean> | What tobyink said. |
| 11:50:56 | <Anchakor> | idmclean: how about drawing a venn diagram explaining that? ;) |
| 11:51:00 | <idmclean> | In the context that we're trying to define, constructs are parts of a Language of Languages. |
| 11:51:15 | <tobyink> | Sign language (and there are various types of sign language) is mostly a distinct language from English. But if you restrict it to just the signs which represent letters, you can come up with something which is a serialisation of English plus a dialect of Sign Language. |
| 11:51:50 | <tobyink> | Where is danbri in this conversation. Seems like something that would be right up his street. |
| 11:52:31 | <idmclean> | U is a Language of Languages. B and C are distinct Dialects in a Language of Languages. U is defined as being all the elements in B and C, but none of the Elements at the intersection of B and C. |
| 11:53:04 | <danbri> | ACTION can't stop now |
| 11:53:05 | <danbri> | but |
| 11:53:10 | <tobyink> | In my model I treat :dialect and :serialisation as different sorts of subsetting relationships. |
| 11:53:10 | <danbri> | .g tom baker pidgin dublin core |
| 11:53:11 | <phenny> | danbri: http://www.dlib.org/dlib/october00/baker/10baker.html |
| 11:53:27 | <danbri> | see also |
| 11:53:27 | <danbri> | http://dublincore.org/news/2009/#dcmi-news-20090518-03 |
| 11:54:45 | <tobyink> | Awesome - http://www.dlib.org/dlib/october00/baker/10baker.html is written by Doctor Who! |
| 11:57:22 | <idmclean> | I think my reasoning about the intersection of B and C maybe too general. |
| 11:58:15 | <idmclean> | I'm thinking where two dialects intersect, the machine would be unable to parse because of ambiguity of conflicting rules/syntax/semantics/definitions, etc |
| 11:58:47 | <Shepard> | tobyink: sign language is a type of language, not a language in itself :) |
| 11:59:19 | <tobyink> | :BritishSignLanguage rdfs:subClassOf :SignLanguage . |
| 12:00:56 | <tobyink> | Under my model, the distinction between serialisations, languages and types of language is intentionally vague. Instead, the preciseness lies in the relationships between them all. |
| 12:01:42 | <Anchakor> | how about ":dialect rdfs:subPropertyOf rdfs:subClassOf . :Language rdfs:subClassOf rdfs:Class ."? |
| 12:02:02 | <Anchakor> | s/dialect/dialectOf/ |
| 12:02:23 | <tobyink> | This is what enables it to cope with the fact that there exists a language :X such that :X is :dialect of :BritishSignLanguage , is :serialisation of :English . |
| 12:03:12 | <tobyink> | :dialectOf rdfs:subPropertyOf rdfs:subClassOf works. |
| 12:03:56 | <tobyink> | I chose to define it as :dialect rdfs:subPropertyOf [ owl:inverseOf rdfs:subClassOf ] instead though because I prefer terms not to have "of" at the end. |
| 12:04:53 | <tobyink> | This is because N3 already has the "is ... of" pattern for representing the inverse meaning of a predicate. |
| 12:05:05 | <tobyink> | And using "is ...Of of" gets confusing. |
| 12:05:13 | <Anchakor> | would ":Language rdfs:subClassOf rdfs:Class ." work? |
| 12:05:52 | <Anchakor> | then I could say ":Python a :Language" |
| 12:06:04 | <tobyink> | Yes. I use owl:Class instead though. No particular reason, but OWL reasoning tends to behave better with owl:Class than rdfs:Class . |
| 12:06:58 | <tobyink> | Anchakor: you can't do {:Python a :Language} under my model. |
| 12:07:20 | <tobyink> | :Language is poorly named in my model. Instead of representing the class of all languages, it represents the class of all things written in any language. |
| 12:07:51 | <Anchakor> | if ":Language rdfs:subClassOf rdfs:Class ." is defined, it should. |
| 12:08:04 | <tobyink> | Thus :Python is a subclass of :Language because the set of all things written in Python is a subset of the set of all things written in any language. |
| 12:08:46 | <idmclean> | Anchakor, I wouldn't allow a direct statement like ":Python a :Language" because while it seems intuitively right, I have to ask you first what is a language? What are it' |
| 12:08:57 | <idmclean> | it's necessary and sufficient qualities? |
| 12:09:35 | <idmclean> | Once again, what tobyink said. |
| 12:10:18 | <tobyink> | Basically, ":Python" in my model doesn't represent a programming language (as it might appear to) but rather represents the set of all things written in Python. |
| 12:10:49 | <idmclean> | We sidestep having to define a language, by simply pointing at an example/implemenation of language and going that is a language and implying that it is a subset of a language. |
| 12:10:57 | <tobyink> | So :Python isn't "a language". |
| 12:11:42 | <idmclean> | Instead of defining strictly and percisely what a language is, we build up the definition through inference from examples. |
| 12:11:52 | <tobyink> | exactly. |
| 12:12:51 | <idmclean> | What Python would most likely get pointed to first is the library implementation of the "language" definition. |
| 12:13:54 | <idmclean> | A computer can extract things like the semantics, syntax, grammar, and such from that and compare it to other sets of semantics, syntax, grammar, and such. |
| 12:15:13 | <idmclean> | If we break up the sets so that each set is compatible and nonconflicting with any other set, we can merge the sets into one set. That set is a Language of Languages. |
| 12:16:03 | <tobyink> | This conversation is starting to get distinctly "Snow Crash". |
| 12:20:35 | <Anchakor> | idmclean: yeah, noone knows what is languague anyways so that might be best :) |
| 12:21:04 | <tobyink> | idmclean: If you want, I can provide a stable URI at http://ontologi.es for a language ontology, and a subversion repo for developing the ontology. |
| 12:22:41 | <idmclean> | tobyink, at this point you've done the most work implementation wise to be honest. I'm trying to teach myself the notation and concepts. I was tripping over thinking of the things-written-in-alol as documents which is basically incompatible with RDF and graphs. |
| 12:23:30 | <idmclean> | tobyink, if you would like to host the ontology, I would be grateful for the assistance. |
| 12:33:42 | <idmclean> | Anchakor, vW-Grammars were very useful to learn about in this context. A two-level or three-level vW-Grammar can generate an infinite number of "languages", so it's too powerful of a language to be used in computing unrestricted because parsing it is inheriently undecidable. |
| 12:41:26 | <idmclean> | Anchakor, if there exists a language which is capable generating an infinite number of languages, then any universal language is going to be on the same order of infinity or greater. All languages that can be expressed then should be subsets in that infinite set of langauges. |
| 12:44:36 | <Anchakor> | yeah, thats better than try to define what language is... anyway I am more interested in semantics for which the languages were created to capture and their interprets/compilers... |
| 12:44:48 | <edsu> | published some linked data in an app at loc.gov last night ; e.g. http://linkeddata.uriburner.com/about/html/http://chroniclingamerica.loc.gov/lccn/sn85066387;thing |
| 12:49:25 | <yvesr> | just stumbled across http://sameas.freebase.com/ - it looks like a base of sameAs relationships from Freebase to other LOD datasets |
| 12:49:42 | <yvesr> | anyone knows how that work? |
| 12:51:40 | <tobyink> | idmclean, Anchakor: ontology has a URI now. <http://ontologi.es/lang/core#> for core terms ; <http://ontologi.es/lang/programming-languages#> for one particular view of how programming languages relate together ; <http://ontologi.es/lang/data-languages#> for one particular view of how data languaes fit together. |
| 12:53:42 | <idmclean> | tobyink, thank you. |
| 12:58:02 | <idmclean> | tobyink, minor quibbles: AnyLanguage is equivalent to All Languages. AnyLanguage can only represent the set of all things written in any known/documented language. |
| 12:59:34 | <idmclean> | MachineLanguage is the set of all things written in language/format that *is* machine-readable. |
| 13:01:16 | <idmclean> | tobyink, Anchakor, I'm going to look at putting together a specification draft. On that note, anyone know of resources for writing a w3c specification? |
| 13:03:25 | <Anchakor> | idmclean: dunno if that needs specification... what is it usefull for anyway? |
| 13:04:02 | <Anchakor> | tobyink: getting this from cwm: "HTTP Error 406: Not Acceptable" |
| 13:04:26 | <Anchakor> | did "./cwm http://ontologi.es/lang/programming-languages#" |
| 13:04:47 | <idmclean> | Anchakor, it's useful to document things so that people like tobyink and yourself can reference it in expanding, examining, critiquing, or revisioning the implementation. |
| 13:05:32 | <idmclean> | So it can be shared with people asynchronistically who might be able to act on implementing the specification in the mean time. |
| 13:05:39 | <Anchakor> | idmclean: ok, maybe use some wiki |
| 13:18:32 | <idmclean> | tobyink, I posted the links at http://hypography.com/forums/computer-science/9168-self-referential-language-3.html |
| 13:24:53 | <iand> | tobyink: thanks |
| 13:24:54 | <phenny> | iand: 11:35Z <tobyink> tell iand <http://chatlogs.planetrdf.com/swig/2009-05-19.html#T11-32-06> |
| 13:25:05 | <iand> | tobyink: http://vocab.org/bio/0.1.html works though :( |
| 13:30:26 | <iand> | tobyink: http://vocab.org/bio/0.1/.html is fixed now |
| 13:32:31 | <tobyink> | Anchakor: could you let me know the output of "./cwm http://richard.cyganiak.de/2008/03/rdfbugs/accept.php" |
| 13:34:08 | <Anchakor> | tobyink: http://codepad.org/SysDMyt2 without the warnings for deprecated python features |
| 13:35:31 | <tobyink> | Looks like CWM sends the obsolete text/rdf+n3 content-type in its Accept header. |
| 13:35:53 | <tobyink> | I should be able to do a little Apache magic to fix it at my end. |
| 13:37:12 | <Anchakor> | we need some rdf data about the serialization and the content-types they use :) |
| 13:37:29 | <Anchakor> | s/the serialization/rdf serializations/ |
| 13:38:37 | <mhausenblas> | any SPARQL Update guys around? |
| 13:39:03 | <mhausenblas> | what is the proposed HTTP method for sending SPARQL diffs? |
| 13:39:41 | <shellac> | POST, I assume |
| 13:39:47 | <shellac> | what are the options? |
| 13:40:21 | <thomase> | shellac: PUT maybe the other one :) |
| 13:40:56 | <mhausenblas> | any concrete evidences? |
| 13:41:06 | <mhausenblas> | my hunch would be that as well, but .... |
| 13:41:13 | <shellac> | joseki |
| 13:41:38 | <mhausenblas> | I seem not even be able to find the section in the current SPARQL spec |
| 13:41:52 | <mhausenblas> | shellac: a standard, not an implemention ... |
| 13:42:02 | <mhausenblas> | even though Joseki is somewhat special, yes ;) |
| 13:42:09 | <shellac> | you said concrete |
| 13:43:08 | <mhausenblas> | concrete in the sense of paragraph in spec, yes, sorry ;) |
| 13:43:12 | <mhausenblas> | ok, http://www.w3.org/TR/rdf-sparql-protocol/ seems to tell |
| 13:43:23 | <mhausenblas> | now, still, what are the SPARUL WG chaps after |
| 13:43:40 | <mhausenblas> | please don't tell me that you want to continue GET for updating stuff |
| 13:43:57 | <Shepard> | I don't think PUT works for diffs. with PUT you're supposed to send a complete representation as a replacement |
| 13:44:03 | <Shepard> | PATCH maybe? |
| 13:44:09 | <tobyink> | Hmmm... Apache doesn't seem to have a mechanism for indicating that a single file has mutliple content-types. |
| 13:44:11 | <shellac> | I assumed PUT was a joke :-) |
| 13:44:19 | <shellac> | PATCH doesn't exist |
| 13:44:33 | <thomase> | hehe I just mentioned that it _might_ be a possible method |
| 13:44:33 | <swh> | we use PUT to write to sparql stores |
| 13:44:37 | <shellac> | POST is fine, as long as you only use it for one thing |
| 13:44:49 | <swh> | POST has slightly vague semantics |
| 13:44:54 | <swh> | whereas PUT is very clear |
| 13:45:11 | <Shepard> | hmm, I thought PATCH existed in WebDAV |
| 13:45:25 | <tobyink> | My thoughts on PUT/POST/etc - http://buzzword.org.uk/2009/posted-data/spec |
| 13:45:31 | <shellac> | oh, maybe in webdav. |
| 13:46:28 | <shellac> | swh: PUT is clear, but PUT insert { } seems clearly wrong |
| 13:47:16 | <swh> | oh, sure |
| 13:47:17 | <oshani> | mhausenblas, tabulator uses POST |
| 13:47:27 | <shellac> | aiui POST is fine. the issue with POST is overloading. stick to one thing, and it's ok |
| 13:47:30 | <mhausenblas> | both PUT and POST have clear semantics (FWIW), that was not the question ;) |
| 13:47:45 | <mhausenblas> | thanks oshani |
| 13:47:51 | <mhausenblas> | for SPARQL Update, correct? |
| 13:47:58 | <oshani> | Yes |
| 13:48:01 | <mhausenblas> | thanks |
| 13:48:05 | <mhausenblas> | btw, all well now? |
| 13:49:52 | <shellac> | ACTION chuckles at the clear semantics of POST |
| 13:50:32 | <shellac> | I'd suggest it contains clearly delimited regions of vagueness, at best |
| 13:50:38 | <mhausenblas> | what is the problem re PSOT in your opinion? |
| 13:51:26 | <shellac> | I have no problem with it. |
| 13:51:46 | <mhausenblas> | oshani: now, as you have endless time, again ;) did your review http://esw.w3.org/topic/WriteWebOfData already? |
| 13:52:09 | <mhausenblas> | shellac: no offence meant. I'm eager to learn what other people might have experienced ... |
| 13:52:33 | <shellac> | I was just reading the specs and it does stand out from the others |
| 13:52:53 | <shellac> | it enumerates what you might want to do with it |
| 13:54:24 | <mhausenblas> | ACTION sorry, gotta run quickly, BRB ... |
| 13:57:45 | <oshani> | mhausenblas, nope not yet, but I will have at it look later :-) (right now I'm in a meeting) |
| 14:02:07 | <mhausenblas> | ACTION back ... sorry shellac |
| 14:02:24 | <mhausenblas> | good, oshani - happy met-up :) |
| 14:07:19 | <Shepard> | mhausenblas: basically there is POST(a) where you append something to a resource or create a new resource - this is quite RESTful. then there's POST(p) where you send some data and it processes it in some way - richardson&ruby also call this "overloaded POST" |
| 14:09:09 | <shellac> | are they saying either is bad? or simply 'don't do both'? |
| 14:12:35 | <shellac> | ah, nevermind, I hadn't fully understood the nature of "overloaded POST" |
| 14:14:25 | <Shepard> | they say overloaded POST is bad because it can turn stuff into RPC-style but sometimes it's unavoidable. but they also say that you should introduce new resources to stay inside the uniform interface |
| 14:14:46 | <Shepard> | so I guess what they propose is not to use both things on one resource |
| 14:15:45 | <Shepard> | so it's ok if a resource accepts POST for processing stuff but then the same resource shouldn't be used for accepting POST(a) |
| 14:17:34 | <Shepard> | I think it's written a bit confused and that by "overloaded POST" they don't just mean a resource that does some processing instead of appending but a resource that does several things |
| 14:17:56 | <shellac> | I suspect it was mnot who made POST issues a bit clearer for me. bad sign: looking at posted content then deciding what to do |
| 14:19:20 | <shellac> | ACTION has tried and failed to find mark baker blog post around restful sql |
| 14:20:03 | <shellac> | ah, http://dowhatimean.net/2006/11/restful-sql |
| 14:29:08 | <mhausenblas> | ACTION wondering if cygri has changed his mind since :) |
| 14:30:36 | <mhausenblas> | http://people.csail.mit.edu/eob/2009/05/19/a-simple-extension-for-microformat-rdfa-table-support/ |
| 14:30:50 | <mhausenblas> | F:| A Simple Extension for Microformat & RDFa Table Support |
| 14:31:23 | <mhausenblas> | F: Hu? YAEMF (yet another embedded metadata format proposal) |
| 14:31:58 | <mhausenblas> | F: I think we have already some choices; please, let's stick with that, don't come up with new proposals ;) |
| 14:32:56 | <mhausenblas> | F: /me wondering who wrote that. /me likes to understand who states what. unsigned posts don't help there |
| 14:34:55 | <PovAddict> | agree with F2 completely |
| 14:40:07 | <mhausenblas> | Shepard: I don't understand your distinction. Seems artificial and not HTTP spec-conforming to me |
| 14:40:18 | <mhausenblas> | Shepard: did you read http://roy.gbiv.com/untangled/2009/it-is-okay-to-use-post ? |
| 14:40:29 | <idmclean> | mhausenblas, did you see the prototype ontology that tobyink threw together for describing a Language of Languages? |
| 14:40:45 | <mhausenblas> | (and I guess Roy can be seen as a sort of expert in the REST area, do we agree? :) |
| 14:41:00 | <mhausenblas> | idmclean: yes. but I fail to see its use :) |
| 14:41:23 | <mhausenblas> | is this just for fun (which is fine, don't get me wrong) |
| 14:41:31 | <mhausenblas> | or does this solve a problem? |
| 14:41:35 | <Shepard> | mhausenblas: I didn't make that distinction, it's one the REST community came up with |
| 14:41:46 | <mhausenblas> | Shepard: see my last comment ;) |
| 14:41:59 | <idmclean> | mhausenblas, in a more filled at form, it will be useful for automatically analyzing, altering, comparing, etc different langauges. |
| 14:42:09 | <PovAddict> | hmm |
| 14:42:22 | <mhausenblas> | idmclean: cool. please keep me posted when there is an implementation that consumes/processes this |
| 14:42:25 | <tobyink> | Well, for one, you can say { <foo> :written_in :Turtle } and use OWL reasoning to figure out { <foo> :written_in :RDF } . |
| 14:42:31 | <PovAddict> | what about sending a delta http://tools.ietf.org/html/rfc3229 in PUT? |
| 14:42:43 | <mhausenblas> | tobyink: what's the use case? |
| 14:42:57 | <kidehen> | iand: u there ? |
| 14:43:12 | <tobyink> | Classifying what languages things are written in. (Including human and machine languages). |
| 14:43:12 | <PovAddict> | I think if you follow that spec, sending an IM header (Instance Manipulation), and you know the server can take it, PUTting a diff would have clear semantics |
| 14:43:59 | <mhausenblas> | again, PovAddict, and FWIW, Shepard - I don;t think there's an issue with the HTTP method's semantics |
| 14:44:03 | <kidehen> | iand: need response re. "influenced_by" and relationship ontology, I assume you got my phenny message? |
| 14:44:14 | <mhausenblas> | just with the way people and systems use them ;) |
| 14:44:18 | <PovAddict> | I wasn't talking about POST (which most of the discussion was about) |
| 14:44:19 | <iand> | kidehen: hi |
| 14:44:23 | <kidehen> | iand: actually relationship vocab :-) |
| 14:44:37 | <mhausenblas> | PovAddict: me neither but about *all* HTTP methods |
| 14:44:39 | <iand> | kidehen: I don't have an answer on that, i.e. I don't have a view as to why it isn't there |
| 14:44:51 | <kidehen> | iand: could it be added quickly? |
| 14:45:02 | <idmclean> | mhausenblas, for me, the use case is making a programming "language" using the aLoL schema for the declarative portion of the "language". The purpose of the language is to develop and analyze languages. The ontology will help with making inferances about the existence of undeclared languages. |
| 14:45:16 | <PovAddict> | PUT is supposed to send a complete replacement, not a patch (that's what PATCH is for) but I think PUT + RFC3229 would be ok |
| 14:45:52 | <iand> | kidehen: is there a definition I can use? is it reflecting something else in xfn or other? |
| 14:46:50 | <mhausenblas> | idmclean: that sounds rather like a solution, not a use case to me ;) |
| 14:47:03 | <idmclean> | mhausenblas, generally, you can identify a piece of source code written in python and map it to source code for java or anyother language. |
| 14:47:52 | <idmclean> | mhausenblas, I'm unfamiliar with the formal specification of a use-case. Would you enlighten me? |
| 14:48:26 | <mhausenblas> | idmclean: can you come up with a story like: Jim has a yadayaa and wants to blabla, hence he needs to dadada ... |
| 14:48:34 | <mhausenblas> | lemme grab an example for you |
| 14:48:36 | <kidehen> | iand: I'll lookup now and provide a definition in the style of relationship ontology |
| 14:48:45 | <iand> | kidehen: I have just added one |
| 14:48:55 | <iand> | I can publish... what's the urgency? |
| 14:48:59 | <mhausenblas> | idmclean: for example see http://www.w3.org/2008/WebVideo/Fragments/WD-media-fragments-reqs/#use-cases |
| 14:49:30 | <mhausenblas> | that is, idmclean, you don't tell about the solution but about the problem at hand, describe the task |
| 14:49:33 | <kidehen> | iand: I am working on some simple Linked Data demos that start with profiles. The goal is to show that describing yourself with granularity is one of the keys to Web karma :-) |
| 14:50:05 | <iand> | see http://vocab.org/relationship now |
| 14:50:20 | <mhausenblas> | from that you can theoretically derive requirements and once your stuff is in place show how you solve the UC using your technology |
| 14:50:50 | <idmclean> | Use case: Abe wants to implement reflection in C++. Abe knows that Java has an implementation. |
| 14:51:27 | <idmclean> | mhausenblas, like that? |
| 14:52:03 | <kidehen> | iand: definition sent via mai. |
| 14:52:10 | <kidehen> | iand: s/mai/mail |
| 14:55:10 | <idmclean> | mhausenblas, Abe can utilize an n3 representation of the abstract semantic graph for Java and C++ to map java.lang.reflect to C++ thus generating most if not all of the implementation in a direct translation. |
| 14:55:14 | <PovAddict> | if I wanted to implement FOAF+SSL, I need a proper ($$) SSL certificate on the server? |
| 14:57:57 | <Anchakor> | PovAddict: no, you can use self-signed |
| 14:58:20 | <mhausenblas> | idmclean: I think you start to get the idea ;) |
| 14:58:31 | <PovAddict> | well, but to use FOAF+SSL, my site needs to be in HTTPS, right? |
| 14:58:35 | <iand> | kidehen: I don't have your email yet, but does http://vocab.org/relationship/influencedBy.html meet your needs? |
| 14:58:51 | <mhausenblas> | the more you describe the problem and certain actions and the less potential solutions the better it is (for you and your readers ;) |
| 14:59:03 | <Anchakor> | PovAddict: yes, it has to communicate with https |
| 14:59:10 | <PovAddict> | and a self-signed cert in HTTPS means browser nagbox saying it's not a valid cert are you absolutely completely sure you want to connect anyway? |
| 14:59:39 | <Anchakor> | PovAddict: yes, some dialog window |
| 14:59:55 | <PovAddict> | in Firefox 3 it's quite a task to get past that (intentionally) |
| 15:05:13 | <idmclean> | tonyink, you seem to understand the utility of aLoL. What're some use-cases we can examine? |
| 15:06:56 | <tobyink> | idmclean: Distributed programming languages? If every node in the "abstract syntax web" has a URI, they can all live on different servers. Then that abstract syntax web can be compiled down into a local representation written in the language of your choice. |
| 15:07:14 | <kidehen> | iand: yep, and you will see it practically identical to what I sent :-) |
| 15:09:11 | <idmclean> | tobyink, that is a really nifty use for it. |
| 15:10:20 | <tobyink> | Indeed. I don't know enough about parsing / compiling to write such an app myself, but it would be hella cool. |
| 15:12:11 | <tobyink> | Anchakor: try the following... cwm http://ontologi.es/lang/data-languages http://ontologi.es/lang/core --think='http://goddamn.co.uk/viewvc/misc/rdfs-rules.n3?revision=39&view=co' |
| 15:13:48 | <idmclean> | Feature/bug mapping between languages, distributed programming languages, automated source code translation between languages, automated programming language analysis... |
| 15:14:33 | <idmclean> | Generating documentation. |
| 15:21:50 | <tobyink> | Imagine you could write a program in RDF - e.g. a sequence of steps would be an rdf:List ; an "if" clause would be an rdf:Resource with an ex:condition predicate and an ex:result predicate, etc. (This would not necessarily be a pleasant language to program in, but imagine you could write your program in another language and then run a translation into RDF.) |
| 15:22:18 | <tobyink> | Each function, class, module gets given a URI. |
| 15:22:28 | <idmclean> | tobyink, you describe exactly what I mean to do with this. |
| 15:22:41 | <sandro> | i think it's called LISP ? |
| 15:22:57 | <idmclean> | LISP wishes it was this cool |
| 15:23:04 | <tobyink> | It certainly has similarities with LISP. |
| 15:23:33 | <idmclean> | The major difference being the standard production of this will be programming languages instead of programs. |
| 15:25:14 | <tobyink> | Then you can run a local program which fetches all these module, function and class definitions off the web and compiles them down to a file on your disk, translating [ a ex:IfClause ] into whatever syntax an if clause has in the target programming language. |
| 15:27:20 | <idmclean> | tobyink, thanks for all the help today. When I show up next, I think we should talk about one of the other dialects in the core aLoL: the metalanguage, the function dialect, the state dialect, or the automata dialect. |
| 15:27:35 | <tobyink> | No problem. |
| 15:29:16 | <idmclean> | May y'all be well. |
| 16:04:33 | <tobyink> | My contribution to the RDFa-in-HTML debate for the day ... http://www.w3.org/Bugs/Public/show_bug.cgi?id=6854#c1 |
| 16:05:35 | <tobyink> | yvesr: dbtune.org/musicbrainz down? |
| 16:08:55 | <yvesr> | tobyink: d2r unexpectedly crashed, should be fine now |
| 16:49:28 | <mhausenblas> | http://dbooth.org/2009/lifecycle/ |
| 16:49:40 | <mhausenblas> | G:| The URI Lifecycle in Semantic Web Architecture |
| 16:50:04 | <mhausenblas> | G: by David Booth - worth investing the time. very nice read. |
| 18:43:57 | <mhausenblas> | good nite Web of Data |
| 19:12:42 | <Anchakor> | Tomothy: this is what I am doing (and idmclean too, but from different approach), rdf programming inspired by lisp :) |
| 19:13:11 | <Anchakor> | phenny: tell tobyink this is what I am doing (and idmclean too, but from different approach), rdf programming inspired by lisp :) |
| 19:13:12 | <phenny> | Anchakor: I'll pass that on when tobyink is around. |
| 19:13:29 | <Anchakor> | Tomothy: sorry this was meant for tobyink ;) |
| 19:14:46 | <Anchakor> | phenny: tell idmclean do you consider bytecode a programming language? ;) |
| 19:14:46 | <phenny> | Anchakor: I'll pass that on when idmclean is around. |
| 19:45:10 | <Shepard> | phenny, tell tobyink: you can write programs in N3. not in an imperative way of course. another one is http://en.wikipedia.org/wiki/Adenine_(programming_language) |
| 19:45:11 | <phenny> | Shepard: I'll pass that on when tobyink is around. |
| 20:35:05 | <LeeF> | http://www.w3.org/blog/SW/2009/05/19/sparql_working_group_holds_1st_face_to_f |
| 20:35:05 | <phenny> | LeeF: 11:30Z <mhausenblas> tell LeeF see http://chatlogs.planetrdf.com/swig/2009-05-19.html#T11-01-31 |
| 20:35:15 | <LeeF> | H:| SPARQL working group holds 1st face-to-face, decides upcoming features |
| 20:35:26 | <LeeF> | H: Summary of SPARQL WG first face-to-face meeting |
| 20:36:00 | <LeeF> | H: Decided features include aggregates, subqueries, update, projected expressions, negation, and service description |
| 20:36:23 | <LeeF> | H: Time-permitting features include SPARQL/OWL semantics, basic federated query, property paths, and core functions & surface syntax |
| 20:58:59 | <Shepard> | LeeF: that is ace. I think it's quite a good selection. :) |
| 21:01:20 | <LeeF> | Shepard, me too :-) it's also an ambitious selection, so we'll see how it plays out :) |
| 21:01:50 | <LeeF> | Lots of good stuff missed the cut also, but you can't have everything |
| 21:02:08 | <Anchakor> | LeeF: are there any plans to provide rdf vocabulary for expressing rdf queries? |
| 21:02:24 | <Anchakor> | s/rdf queries/sparql queries/ |
| 21:03:42 | <Shepard> | Anchakor: there is http://spinrdf.org/sp.html is you don't want to wait for an official one :) |
| 21:21:39 | <danbri> | H:Am a bit suprised to see federated query on this list, but hey, go for it! :) |
| 21:24:14 | <kasei> | why's that danbri? |
| 21:28:47 | <Anchakor> | ambitious I would guess :) |
| 21:30:24 | <kasei> | the federated feature under consideration has a fairly narrow scope. |
| 21:31:01 | <kasei> | there are some security concerns, but I don't think it's all that ambitious |
| 22:24:47 | <dajobe> | H:I hope "time-permitting" means you only get to start those features *IF* the decided features are completed |
| 23:42:27 | <kasei> | dajobe: for some of the time-permitting features, work can go on in parallel to the primary work on the required features. |
| 23:53:36 | <LeeF> | Anchakor, it was a proposed feature, but it did not have a critical mass of support within the WG, so no, we won't be pursuing that in this round of specification. |
| 23:54:08 | <kasei> | heya LeeF |
| 23:55:35 | <LeeF> | hi kasei |
| 23:59:09 | <kasei> | ACTION feels a bit bad for dropping off the call this morning as you were about to throw work my way :) |
| 23:59:31 | <kasei> | it was all skype's fault |
Back to channel and daily index: content-negotiated html turtle