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.
| 01:51:09 | <drewp> | man i'd love a URI for the twitter user with account id 16053131 (not the user with the name 'drewpca', since that's not stable) |
| 01:51:38 | <drewp> | getting kind of tired of http://bigasterisk.com/twitter/account/16053131 all over my data |
| 05:10:44 | <mhausenblas> | good morning Web of Data! |
| 05:17:04 | <mhausenblas> | hm. what's up with http://chatlogs.planetrdf.com? seems down ... |
| 05:36:35 | <logger> | ACTION is logging |
| 05:45:10 | <mhausenblas> | ta |
| 08:13:56 | <mhausenblas> | http://bnode.org/blog/2009/05/26/simple-rdfication-of-sparql-select-results-with-rdfa |
| 08:14:04 | <mhausenblas> | A:| Simple RDFication of SPARQL SELECT results with RDFa |
| 08:14:15 | <mhausenblas> | A: by bengee - cool stuff! |
| 09:03:20 | <tobyink> | BLURB:ttldent (pronounced "turtle dent") |
| 09:04:22 | <tobyink> | B: The idea is to post turtle to identi.ca (or another laconi.ca installation, or twitter even). |
| 09:05:06 | <tobyink> | B: There needs to be a script or a website or something to extract turtle from the microblog feeds. |
| 09:06:13 | <presbrey> | hey tobyink, there's a new tabulator release with RDFa parsing, but I can't get it to work with your site |
| 09:06:30 | <tobyink> | B: Certain conventions/shortcuts are applied - URIs are always consider relative to the status update page. Certain hard-coded URIs have special meanings, e.g. <#me> refers to the account holder. |
| 09:07:11 | <tobyink> | presbrey: Might be a MIME type thing. I use text/html I think, not application/xhtml+xml. I'll download and have a look soon. |
| 09:07:48 | <tobyink> | B: Commonly used prefixes are considered predeclared, including FOAF, Dublin Core, SKOS, SIOC, RDF, RDFS, OWL. |
| 09:09:27 | <melvster_> | presbrey: you can try http://www.ivan-herman.net/foaf.html or http://sw-app.org/mic.xhtml#i |
| 09:10:21 | <tobyink> | Comments in ttldent start with a hash sign like they do in turtle, but unlike turtle end with *any* white space (not just a new line character). This coincides with laconica/twitter's hashtag syntax. |
| 09:10:31 | <tobyink> | B: Comments in ttldent start with a hash sign like they do in turtle, but unlike turtle end with *any* white space (not just a new line character). This coincides with laconica/twitter's hashtag syntax. |
| 09:11:41 | <Anchakor> | B: sounds like ugly hack ;) |
| 09:12:17 | <tobyink> | B: All ttldent posts must include the hashtag #ttl to identify them as ttldent posts. Any additional hashtags are treated as tags for the status update using the holygoat ontology. |
| 09:13:02 | <tobyink> | B: Yes, it's a hack, but it could be fun. I'm going to attempt to write a ttldent extractor script on my train journey home this evening. |
| 09:16:15 | <tobyink> | B3: Certain conventions/shortcuts are applied - URIs are always consider relative to the status update page. Certain hard-coded URIs have special meanings, e.g. <#me> refers to the account holder. <@foo> refers to the holder of account "foo" |
| 09:25:10 | <melvster_> | quite an interesting thread about getting XMPP (openfire server + spark client) working with X.509, hopefully this could lead to a FOAF+SSL instant messaging system soon: http://www.igniterealtime.org/community/thread/37967 |
| 09:26:23 | <tobyink> | B7: All ttldent posts begin with the hashtag #ttl to identify them as ttldent posts. Text before #ttl is treated as comments. Any additional hashtags are treated as tags for the status update using the holygoat ontology. |
| 09:42:06 | <Phae> | has anyone written anything interesting about google's rdfa implementation and why it's not great? |
| 09:49:24 | <tobyink> | Phae: http://iandavis.com/blog/2009/05/googles-rdfa-a-damp-squib |
| 09:49:33 | <Phae> | <3 tah |
| 09:51:26 | <david`bgk> | Phae, http://www.jenitennison.com/blog/node/104 is interesting too |
| 09:53:30 | <Phae> | thanks |
| 09:53:35 | <CaptSolo> | mhausenblas: nice try |
| 09:53:49 | <mhausenblas> | CaptSolo: wrong channel ;) |
| 09:54:36 | <CaptSolo> | mhausenblas: right chan. your nick changes are global to the server :P |
| 09:54:56 | <CaptSolo> | you mean, #sioc is the right chan then? |
| 09:56:34 | <mhausenblas> | yes |
| 10:28:09 | <timbl> | has 'damp squib' been chumped? if not, verychumpable imho |
| 10:28:10 | <phenny> | timbl: 16 May 19:07Z <mhausenblas> tell timbl re Tabulator / Turtle fine now, oshani helped me out |
| 10:28:11 | <phenny> | timbl: 18 May 08:32Z <mhausenblas> tell timbl see http://esw.w3.org/topic/WriteWebOfData |
| 10:28:15 | <phenny> | timbl: 20 May 14:29Z <kidehen> tell timbl do you have an SVG variant of: http://www.w3.org/DesignIssues/diagrams/social/acl-arch.png? I want to tweak it a little i.e, expand [data file] box to a broader box that would look like this: [ data space [data file] [dbms] [others] ] . |
| 10:28:17 | <phenny> | timbl: 24 May 15:59Z <mhausenblas> tell timbl that I've updated http://esw.w3.org/topic/WriteWebOfData now - please review/comment |
| 10:38:20 | <presbrey> | mhausenblas: I implemented autoconf support for the WebID Apache modules to increase portability and hopefully support your win32 apache |
| 10:38:30 | <mhausenblas> | presbrey: very cool! |
| 10:39:01 | <mhausenblas> | will check out ASAP (currently preparing ESWC09 dogfood data and that takes my full attention ;) |
| 10:39:20 | <presbrey> | great! thanks for testing |
| 10:39:51 | <mhausenblas> | got an issue tracker somewhere? |
| 10:40:08 | <mhausenblas> | (just in case ;) |
| 10:42:13 | <tobyink> | http://iandavis.com/blog/2009/05/googles-rdfa-a-damp-squib |
| 10:42:31 | <tobyink> | C: Google?s RDFa a Damp Squib |
| 10:42:37 | <tobyink> | C:|Google?s RDFa a Damp Squib |
| 10:42:48 | <tobyink> | C1: No, it hadn't been chumped yet. |
| 10:46:40 | <presbrey> | mhausenblas: not yet; working on it, I'll be around too |
| 10:50:51 | <mhausenblas> | ok, presbrey |
| 12:09:29 | <melvster> | anyone know if there any linked data services for XMPP? |
| 12:09:29 | <phenny> | melvster: 17 May 16:52Z <mhausenblas> tell melvster see http://chatlogs.planetrdf.com/swig/2009-05-17#T16-47-58 |
| 12:10:44 | <melvster> | mhausenblas: wacup is cool! |
| 12:41:17 | <mhausenblas> | ta melvster - just started to toy around with it . do expect more ;) |
| 12:46:47 | <tobyink> | Is there a well-established predicate for "likes"? domain foaf:Person, range owl:Thing. |
| 12:48:47 | <tobyink> | Probably "likes" would be a rdfs:subPropertyOf foaf:topic_interest, but it's certainly not the same as foaf:topic_interest. |
| 12:49:27 | <tobyink> | e.g. you could be interested in the topic of facism without liking it. |
| 12:49:42 | <gromgull> | tobyink: aha - the new facebook-ish simplistic modelling of the world "I like this" :) |
| 12:50:03 | <gromgull> | In nepomuk we used ratings... rather than likes - but this is much more complicated, it's not a binary predicate, you have the issue of what ranks to use |
| 12:50:03 | <gromgull> | etc. |
| 12:50:16 | <gromgull> | Nepomuk in KDE4 has stars from 1 to 5 I think |
| 12:50:44 | <tobyink> | Ratings are certainly a good system, but I'm looking for something simpler. |
| 12:52:04 | <tobyink> | e.g. { a :Rating ; :rater ?person ; :rated ?thing , :score ?score . } and { ?score > 3 } => { ?person :likes ?thing . } |
| 12:52:21 | <gromgull> | yes |
| 12:52:30 | <tobyink> | So ratings would effectively be a reified version of my simpler "likes" predicate. |
| 12:52:32 | <gromgull> | I don't know of any such property though - I made up my own before |
| 12:52:36 | <gromgull> | but that doesn't help you |
| 12:52:46 | <tobyink> | open.vocab.org, here I come... |
| 12:53:05 | <tobyink> | Actually, I'll host it on ontologi.es . |
| 12:54:06 | <Phae> | oo.. |
| 12:54:07 | <Phae> | ratings |
| 12:54:13 | <Phae> | how very relevant to my interests |
| 12:57:16 | <danbri> | can you rate the relevance on a 1-5 scale, Phae? :) |
| 12:57:42 | <Phae> | hah |
| 12:58:15 | <besbes_> | my proposal is, whenever you do ratings, do not use any numeric scale, but semantically well-defined classes or instances |
| 12:58:27 | <besbes_> | because a rating "1" can mean anything |
| 12:58:35 | <Phae> | sure |
| 12:59:03 | <Phae> | we have a ratings system here |
| 12:59:16 | <Phae> | but people soemone restrict it to out of 2, so they can do a kind of binary for/against vote |
| 12:59:22 | <Phae> | so the number is irrelevant |
| 12:59:25 | <Phae> | it's just like or not like |
| 12:59:35 | <Phae> | but i have a meeting. bbl. |
| 13:08:35 | <gromgull> | i believe - although I could be mistaken, that nepomuk in kde4 has 5 stars shown the UI, but actually the rating are stored as floats between 0 and 1 |
| 13:09:15 | <tobyink> | http://ontologi.es/like |
| 13:09:26 | <tobyink> | D:"like" ontology. |
| 13:09:47 | <tobyink> | D: early draft, includes cwm-compatible inference rules. |
| 13:09:49 | <besbes_> | gromgull, this is equally problematic, since you never know if 0 means "good" or "bad" |
| 13:10:27 | <besbes_> | i mean, we have the semantic web, why do not express ratings in a semantically unambiguous way? |
| 13:10:30 | <melvster> | 1 == 100% so normally means good |
| 13:10:45 | <besbes_> | melvster, the problem lies in "normally" :-) |
| 13:10:51 | <besbes_> | who defines what this is? |
| 13:12:13 | <besbes_> | a set of rating classes or instances can be defined in an ontology, and can be described so everybody can look up their meaning |
| 13:13:56 | <melvster> | foaf privacy concerns: http://my.opera.com/community/forums/topic.dml?id=277003 this could be a good use case for web access control, i wonder if every foaf should have an ACL file |
| 13:14:10 | <danbri> | i've pinged chaals |
| 13:14:28 | <danbri> | the basic guideline should be "don't put more into rdf/xml than into xhtml or html..." |
| 13:14:33 | <danbri> | same acls for each format |
| 13:15:25 | <melvster> | im wondering how to link the WebACL to a FOAF, can we put it in the FOAF, link it in, or should we store it seperately? |
| 13:18:50 | <gromgull> | besbes_: that is documented somewhere of course :) which is sort of good enough - since real automatic processing will never happen anyway :) |
| 13:19:48 | <besbes_> | gromgull, the advantage of using concepts/instances is that you have the immediate link to the documentation (via the concept/instance URI, which is *of course* dereferencable) |
| 13:20:09 | <besbes_> | and, as you see by this discussion, this is needed not only in cases where there is automatic processing :-) |
| 13:46:08 | <tobyink> | cwm is awesome. It really is. |
| 14:04:07 | <timbl> | :-)) A wile since anyone has said that |
| 14:04:16 | <timbl> | maybe we should maintain it a bit |
| 14:04:21 | <timbl> | s/wile/while/ |
| 14:13:34 | <cygri> | tobyink, i'm curious, what's awesome about it? |
| 14:14:13 | <tobyink> | cygri: what's *not* awesome about it. |
| 14:14:47 | <tobyink> | Right now, I'm most impressed with the maths+logic inferencing. |
| 14:15:36 | <tobyink> | e.g. the second to last rule here <http://ontologi.es/like> actually works. |
| 14:16:58 | <tobyink> | Yes, I know it's nothing that prolog can't handle easily, but still, as RDF tools go, it's very clever.\ |
| 14:18:00 | <cygri> | ok. cheers tobyink |
| 14:18:07 | <cygri> | cwm webservice anyone? ;-) |
| 14:18:49 | <kwijibo> | how performant is it? |
| 14:20:36 | <tobyink> | kwijibo: I've only really used it on very small data sets, so wouldn't like to say how fast it could reason on larger sets. |
| 14:25:07 | <shellac> | iirc it was a fairly naive implementation |
| 14:26:26 | <shellac> | I think euler could handle n3 rules, and was faster (used rete?) |
| 14:28:22 | <shellac> | http://eulersharp.sourceforge.net/ |
| 14:28:54 | <shellac> | E:|Euler: compatible with cwm |
| 14:38:53 | <tobyink> | shellac: thanks for the link. I will be sure to try that out later. |
| 14:39:53 | <tobyink> | Possible tagline for Euler - "Delivers as much awesomeness as cwm, but faster". |
| 14:43:16 | <danbri> | i tried to make an euler in ruby once via some programmatic python-to-ruby convertor, but it didn't go... |
| 15:18:58 | <gromgull> | the euler code is rather on the unreadable side... |
| 15:20:03 | <shellac> | I remember it was one huuuuge java method |
| 15:20:20 | <shellac> | but that may be an exaggeration :-) |
| 15:56:57 | <PovAddict> | shellac: you haven't seen huge until you see C code written by FORTRAN programmers |
| 15:57:02 | <PovAddict> | "you can write FORTRAN in any language" |
| 15:57:04 | <PovAddict> | http://pastebin.com/f6d215fd8 |
| 15:57:14 | <PovAddict> | <.< |
| 15:57:34 | <PovAddict> | F:"It's possible to write FORTRAN code in any language. That is definitely FORTRAN." |
| 15:58:07 | <PovAddict> | F:main() starts in line 123 and ends in line 4197 |
| 16:19:03 | <shellac> | PovAddict: ouch. ouch ouch ouch. |
| 16:19:24 | <PovAddict> | :D |
| 16:20:06 | <PovAddict> | they declare two hundred variables at the beginning of the function |
| 16:20:33 | <scor> | mhausenblas: ping |
| 16:27:41 | <mhausenblas> | pong |
| 16:27:52 | <mhausenblas> | (here or over there?) |
| 16:28:01 | <mhausenblas> | over there, obviously ;) |
| 18:37:22 | <mhausenblas> | http://data.semanticweb.org/conference/eswc/2009/ |
| 18:37:49 | <mhausenblas> | G:| ESWC 2009 publication data in RDF and HTML |
| 18:38:13 | <mhausenblas> | G: fresh dogfood, served with spicy semantics ;) |
| 18:44:10 | <PovAddict> | mhausenblas: 12. Decidability of $\mathcal{SHI}$ with transitive closure of roles |
| 18:44:24 | <PovAddict> | isn't that something that failed to be replaced? |
| 18:55:12 | <mhausenblas> | yep, PovAddict ;) |
| 18:55:16 | <mhausenblas> | thanks |
| 18:55:30 | <mhausenblas> | for spotting it |
| 19:54:08 | <Anchakor> | anyone knows of article about rdfs/owl ontology design malpractices? |
| 20:16:37 | <timbl> | leef, http://www.w3.org/2009/recovery/all.n3 |
| 20:39:31 | <mhausenblas> | good nite Web of Data |
| 20:45:21 | <Anchakor> | any ideas why rdfs spec says "rdfs:Datatype rdfs:subClassOf rdfs:Class"? |
| 20:49:26 | <Shepard> | I'd much rather someone explained to me the definitions of owl:Thing and owl:Nothing - they make no sense to me whatsoever :) |
| 21:02:39 | <Anchakor> | I guess Ill read that like "instances of rdfs:Datatype are rdfs:Classes too" |
| 21:03:56 | <lbjay> | @dict artisan |
| 21:04:05 | <lbjay> | dang... |
| 21:04:08 | <lbjay> | lurk fail |
| 21:08:13 | <Anchakor> | Shepard: see this and scroll 3 lines up, that is only thing I know :) http://www.w3.org/TR/2004/REC-owl-ref-20040210/#EnumeratedClass |
| 21:10:49 | <Anchakor> | "class extenstion" == "instance" |
| 21:11:55 | <Shepard> | it's the definition under http://www.w3.org/TR/2004/REC-owl-ref-20040210/#appB that confuses me |
| 21:16:39 | <Anchakor> | yeah it's a use of this kind of class description http://www.w3.org/TR/2004/REC-owl-ref-20040210/#Boolean |
| 21:18:06 | <Anchakor> | I guess the editors thought it would be useful for something, though I can see what now :) |
| 23:10:08 | <CaptSolo> | yvesr: hi |
| 23:10:55 | <CaptSolo> | ACTION could you ping me when online? |
| 23:10:59 | <CaptSolo> | ups |
| 23:43:02 | <rockyrock> | do you use Prolog in semantic web? |
Back to channel and daily index: content-negotiated html turtle
