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.
| 00:03:12 | <drewp> | dbpedia down? :( |
| 00:07:26 | <snail> | I'm getting (104) Connection reset by peer errors |
| 04:38:39 | <mhausenblas> | good morning Web of Data |
| 04:46:34 | <LeeF> | g'morning, mhausenblas |
| 04:52:08 | <mhausenblas> | hey LeeF |
| 05:34:42 | <drewp> | it's been several years since my last bad experience with nagios, and i need something like that, so i'm trying it again |
| 05:35:31 | <drewp> | why oh why can't the monitoring plugins just assert some RDF statements, we collect them in some trivial way, and then the monitoring interface can just be sparql queries over those statements? |
| 05:36:52 | <drewp> | each report from a monitoring plugin gets its own subgraph/context, so when you ask "why" about any given statement, we can tell you when and where it was determined |
| 05:40:31 | <Shepard> | build it :] |
| 05:41:10 | <mhausenblas> | ACTION was waiting for someone to say that ;) |
| 05:41:29 | <drewp> | sounds like a fun project to me, but i have too many other things broken right now (which is why i'm installing nagios) |
| 05:41:41 | <mhausenblas> | however, I know that drewp has the knowledge to do so ... |
| 05:42:02 | <drewp> | i'm trying to get someone else excited about it, here |
| 05:42:33 | <mhausenblas> | yeah, shame, isn't it. there are always so many things around that would be nice or even great to have ... but unless someone invents the 48h day ... |
| 05:42:46 | <mhausenblas> | drewp: how about Shepard ? :D |
| 05:44:16 | <mhausenblas> | ACTION gotta grab some coffee ... BRB |
| 05:46:01 | <Shepard> | hey, I got tons of projects already ;) |
| 06:00:20 | <mhausenblas> | Shepard: well ... ;) |
| 06:01:24 | <mhausenblas> | timbl, got time for some quick Qs re WAC? |
| 06:22:38 | <kwijibo> | mhausenblas: bug report for lidaman ? |
| 06:23:25 | <kwijibo> | the "is of type" link seems to get mixed up with other links |
| 06:24:45 | <kwijibo> | for instance it says is of type kwijibo, creator: owl:Thing |
| 06:24:59 | <kwijibo> | instead of the other way round |
| 06:26:00 | <mhausenblas> | right, thanks kwijibo. |
| 06:26:31 | <mhausenblas> | would it be too much to ask to put it on http://bitbucket.org/mhausenblas/lidaman/issues/ ? |
| 06:26:32 | <mhausenblas> | :) |
| 06:26:51 | <kwijibo> | consider it done |
| 06:26:53 | <mhausenblas> | (maybe even with a link to an example where it fails) |
| 06:26:56 | <mhausenblas> | cheers! |
| 06:32:53 | <kwijibo> | done |
| 06:33:20 | <mhausenblas> | thanks, kwijibo |
| 06:33:31 | <mhausenblas> | seen Manu's proposal on RDFa list? |
| 06:33:41 | <kwijibo> | not yet |
| 06:34:00 | <mhausenblas> | see http://lists.w3.org/Archives/Public/public-rdfa/2009May/0030.html |
| 06:34:07 | <kwijibo> | ah yes, i did |
| 06:36:38 | <mhausenblas> | and? your thoughts? |
| 06:37:11 | <kwijibo> | hmm, it's not super easy :) |
| 06:37:17 | <kwijibo> | does neologism not do RDFa ? |
| 06:37:38 | <kwijibo> | or is that a "could do ... later" issue ? |
| 06:38:57 | <kwijibo> | I think this is interesting: # There are currently no online tools to correct common vocabulary authoring mistakes. |
| 06:39:13 | <mhausenblas> | did you also read my reply? :P |
| 06:39:18 | <kwijibo> | yes |
| 06:39:30 | <mhausenblas> | well, I thought I've listed at least some :D |
| 06:39:35 | <kwijibo> | i read yyour reply first :) |
| 06:39:38 | <mhausenblas> | hehe |
| 06:40:01 | <mhausenblas> | anyways, I think whatever the outcome is it should not be RDFa-specific |
| 06:40:20 | <mhausenblas> | this is really a general RDF issue and people should think big here |
| 06:40:33 | <mhausenblas> | maybe open.vocab.org is a good starting point |
| 06:40:48 | <mhausenblas> | maybe knoodle is, I dunno, but nothing RDFa-only, IMHO |
| 06:41:20 | <kwijibo> | the validation/sanity checking thing is interesting to me: what /are/ the common mistakes? |
| 06:41:54 | <kwijibo> | one is not giving terms rdfs:isDefinedBy properties |
| 06:42:04 | <kwijibo> | well |
| 06:42:17 | <kwijibo> | there are few "mistakes" in RDF land I suppose |
| 06:42:40 | <kwijibo> | but lets say, deviations from most helpful practice |
| 06:46:40 | <mhausenblas> | there are at least two classes of issues, in my understanding |
| 06:47:01 | <mhausenblas> | one is around *using* the schema/modelling language correctly |
| 06:48:29 | <mhausenblas> | see for example http://chatlogs.planetrdf.com/swig/2009-05-17.html#T10-01-25 |
| 06:49:11 | <kwijibo> | haha "using Protege is even greater masochism ;)" |
| 06:49:27 | <mhausenblas> | this class can more or less easily be healed by good practice and active community mechanisms such as #swig channel or mailing list |
| 06:49:28 | <mhausenblas> | ;) |
| 06:49:38 | <mhausenblas> | the second class is maybe more serious |
| 06:49:46 | <mhausenblas> | it's about the actual semantics |
| 06:50:37 | <mhausenblas> | looking at people that do, for example RDBMS schemas for 10 years already, you'll find that they have a certain experience |
| 06:50:53 | <mhausenblas> | they can tell what works or what is not a good idea for a certain use case |
| 06:51:15 | <mhausenblas> | e.g. de-normalised scheme for certain Web apps |
| 06:51:18 | <kwijibo> | like what? |
| 06:51:33 | <kwijibo> | ah |
| 06:51:47 | <mhausenblas> | or when to better do integrity checks and when not, etc. |
| 06:52:06 | <mhausenblas> | however, we as a community have not yet reached this stage, I believe |
| 06:52:45 | <mhausenblas> | additionally it is one thing to create a vocabulary for esp. one app/use case |
| 06:53:20 | <mhausenblas> | and another to create it so flexible that future (re)use and serendipity is sort of built-in |
| 06:54:47 | <mhausenblas> | this is something what we can't really put into software to enforce it or validate and tell people 'hey, if you model this prop like that, you'll run into troubles' |
| 06:55:19 | <mhausenblas> | just recall the many discussions we had about each and every property and class in the voiD voc |
| 06:55:45 | <mhausenblas> | we would dig into it, find out what is available, find indications how and where people use it, etc. |
| 06:55:47 | <kwijibo> | mhausenblas: mainly about naming things ;) |
| 06:56:01 | <mhausenblas> | not really! I mean, yes that was one thing |
| 06:56:15 | <mhausenblas> | but think about the LS modelling |
| 06:56:25 | <mhausenblas> | about the entire stats module |
| 06:56:50 | <mhausenblas> | about how and where to attach provenance information |
| 06:57:15 | <kwijibo> | yeah |
| 06:57:15 | <mhausenblas> | even the seemingly simple thing re exampleResource would take us some time |
| 06:57:21 | <kwijibo> | :D |
| 06:57:26 | <mhausenblas> | because we did it properly |
| 06:57:39 | <mhausenblas> | based on researching what and how things are used out there |
| 06:58:03 | <mhausenblas> | now, the point here is not to praise our work and tell how well we did |
| 06:58:24 | <kwijibo> | mhausenblas: was there anything that was easy to decide on in voiD ? |
| 06:58:37 | <kwijibo> | maybe void:vocabulary |
| 06:58:39 | <mhausenblas> | the point is to list all the things that are necessary to create terms. mainly a social thing |
| 06:58:44 | <mhausenblas> | yep, I guess so |
| 06:59:18 | <mhausenblas> | there were two or three 'obvious' things, incl. vocabulary, yes |
| 06:59:45 | <mhausenblas> | however, this reminds me that we really should sit together soon to discuss what goes into voiD 2.0 |
| 06:59:59 | <kwijibo> | yes |
| 07:00:12 | <mhausenblas> | esp. re SPARQL endpoints and distributed queries there is a lot going on |
| 07:00:32 | <mhausenblas> | and I think next to fixing the stats module this should be our main focus |
| 07:00:39 | <mhausenblas> | will you be around ESWC09? |
| 07:00:46 | <kwijibo> | i don't know |
| 07:00:49 | <kwijibo> | when is it ? |
| 07:01:04 | <mhausenblas> | 31 May to 4 June I believe |
| 07:01:18 | <kwijibo> | no, i don't think so then :/ |
| 07:01:35 | <mhausenblas> | however, IIRC cygri will also not be there and dunno about Jun |
| 07:02:04 | <mhausenblas> | anyways, let's schedule a skype telecon ASAP. how about this week? |
| 07:02:13 | <mhausenblas> | ACTION preparing doodle |
| 07:03:15 | <kwijibo> | this week would be good for me |
| 07:04:04 | <mhausenblas> | ACTION wondering how I get a list of all release 2.0 issues ... |
| 07:12:58 | <mhausenblas> | sent out voiD 2.0 kick-off scheduling ... |
| 07:18:53 | <mhausenblas> | kwijibo: ta. and: what's up? why is Saturday bad? :D |
| 07:20:45 | <kwijibo> | mhausenblas: I dunno, vainly optimistic I might be able to go play outside on saturday |
| 07:21:01 | <mhausenblas> | ah, I see ;) |
| 07:26:42 | <kwijibo> | it is really, really, weird that the HTML5 WG are arguing about "link rot" :s |
| 07:27:44 | <mhausenblas> | ACTION doesn't comment on this |
| 07:27:46 | <mhausenblas> | :) |
| 07:52:54 | <kwijibo> | why be interested in writing a markup language for an environment you don't accept the core working principle of ? |
| 08:00:17 | <mhausenblas> | hu? kwijibo, whatdayamean? |
| 08:02:30 | <kwijibo> | mhausenblas: well, html(5) is markup language for the WWW right ? |
| 08:05:14 | <kwijibo> | web works by being able to follow links / dereference URLs |
| 08:06:53 | <kwijibo> | so if someone doesn't like that, why be interested in a markup language for that system ? |
| 08:08:12 | <Shepard> | I bet they will be happy when RDFa gets included in HTML5 together with a parsing algorithm that contains 25 pages of description of the error handling for dereferencing URIs in the parsed result :) |
| 08:11:34 | <kwijibo> | :D |
| 08:11:53 | <kwijibo> | doesn't seem like RDFa will get included though http://www.webstandards.org/2009/05/13/interview-with-ian-hickson-editor-of-the-html-5-specification/ |
| 08:13:21 | <tobyink> | "ivan" here is Ivan Herman? |
| 08:13:51 | <kwijibo> | "We considered RDFa long and hard (in fact this is an issue that’s a hot topic right now), but at the end of the day, while some people really like it, I don’t think it strikes the right balance between power and ease of authoring. For example, it uses namespaces and prefixes, which by and large confuse authors to no end. Just recently though I proposed something of a compromise which takes some of RDFa’s better ideas and puts them into HTML 5, so ho |
| 08:15:43 | <tobyink> | ... so hop... (truncated) |
| 08:15:58 | <mhausenblas> | kwijibo: yep |
| 08:19:01 | <kwijibo> | mhausenblas: yep it does take care of the main needs that caused people to invent RDFa ? :p |
| 08:20:08 | <mhausenblas> | ? |
| 08:20:12 | <mhausenblas> | ahm no. |
| 08:20:16 | <kwijibo> | :D |
| 08:20:32 | <mhausenblas> | yep, seems RDFa is ignored by HTML5 |
| 08:20:37 | <mhausenblas> | sad but true |
| 08:20:52 | <kwijibo> | ACTION mentally replaces "RDFa" with "Javascript" in that passage to see how reasonable it is |
| 08:21:06 | <mhausenblas> | NIH syndrome + some political issues I guess |
| 08:22:13 | <mhausenblas> | ACTION still find it strange that no one seems to have told Hixie in Google beforehand re RDFa adoption ;) |
| 08:22:59 | <kwijibo> | mhausenblas: maybe no one wanted to be the one to tell him ... |
| 08:23:23 | <Shepard> | lol |
| 08:23:35 | <gromgull> | morning |
| 08:23:40 | <Shepard> | hi |
| 08:24:05 | <mhausenblas> | kwijibo: quite likely. or he didn't listen. |
| 08:24:15 | <mhausenblas> | no. sorry, *that* can't be ;) |
| 08:24:23 | <mhausenblas> | morning gromgull |
| 08:24:35 | <kwijibo> | morning gromgull |
| 08:24:49 | <kwijibo> | nah, I dunno, on the face of it, maybe google's adoption of RDFa is like what gsnedders said about rss 1.0 |
| 08:25:32 | <gsnedders> | Hixie said Google parses RDFa using regex; I presume that refers to what was recently announced. |
| 08:25:44 | <mhausenblas> | yeah |
| 08:25:45 | <kwijibo> | it's open to the charge that they aren't using it to its full potential, just as an alternate serialisation of microformats |
| 08:26:02 | <kwijibo> | oh that's awful :p |
| 08:26:10 | <mhausenblas> | I think the step it self was important, not the concrete implementation |
| 08:26:22 | <mhausenblas> | however, back to 'work' ;) |
| 08:26:24 | <kwijibo> | mhausenblas: yeah that's true too |
| 08:26:27 | <mhausenblas> | http://esw.w3.org/topic/WriteWebOfData |
| 08:26:36 | <mhausenblas> | A:| Realizing a write-enabled Web of Data |
| 08:26:51 | <gsnedders> | mhausenblas: FWIW my opinion on the issues with RDFa in HTML 5 is that RDFa is heavily reliant on namespace prefixes (due to CURIEs), which HTML 5 doesn't have, which makes it hard to start with |
| 08:26:52 | <mhausenblas> | A: trying to gather an overview of what is going on in this area |
| 08:27:27 | <mhausenblas> | gsnedders: maybe you can tell me why XMLNS are such a bad thing, I still don't get it |
| 08:28:15 | <kwijibo> | I thought html 5 would have an xml serialisation as well ? so couldn't one just use xhtml5 with RDFa ? |
| 08:28:15 | <gsnedders> | mhausenblas: I don't have time, but there are compat issues too in HTML (a fair amount of text/html has @xmlns in it and would break if you start parsing it per XML-Names) |
| 08:28:21 | <gsnedders> | kwijibo: Indeed |
| 08:29:05 | <mhausenblas> | ok, gsnedders. |
| 08:29:10 | <gsnedders> | (I might have misunderstood what Hixie said about Google and RDFa, FWIW) |
| 08:29:15 | <kwijibo> | ACTION doesn't understand that |
| 08:29:24 | <mhausenblas> | maybe I should say name spaces and not XMLNS in future ;) |
| 08:29:57 | <gsnedders> | mhausenblas: Then you get divergent syntax for the same thing, which isn't so nice. |
| 08:30:57 | <gsnedders> | ACTION goes back to doing schoolwork |
| 08:30:59 | <mhausenblas> | gsnedders: yep, I think you have a point, there |
| 08:31:06 | <mhausenblas> | NJoy! gsnedders |
| 08:31:09 | <mhausenblas> | last year? |
| 08:31:28 | <gsnedders> | mhausenblas: Indeed |
| 08:31:44 | <gsnedders> | mhausenblas: Had my first exam on Friday, and have been on study leave since last Wednesday |
| 08:32:20 | <mhausenblas> | phenny, tell timbl see http://esw.w3.org/topic/WriteWebOfData |
| 08:32:21 | <phenny> | mhausenblas: I'll pass that on when timbl is around. |
| 08:32:28 | <gsnedders> | It's anyone's guess whether the current microdata proposal in HTML 5 will stick, though |
| 08:33:00 | <kwijibo> | gsnedders: re your question "who uses RSS as RDF" - short list: all/most applications on talis platform, many applications that use ARC I think (eg knowee, paggr), triplr ... |
| 08:33:24 | <mhausenblas> | I see gsnedders - anyway, good luck with the exam |
| 08:33:41 | <gsnedders> | I'd personally like something closer to RDFa, and I said I'd write up some proposal after my exams. But for the two weeks I have free, I've said I'll do a lot :) |
| 08:34:08 | <gsnedders> | kwijibo: That has any significant usage? |
| 08:34:40 | <mhausenblas> | ACTION off for a coffee. again. OMG, I'm drinking too much of it I guess ;) |
| 08:34:46 | <kwijibo> | gsnedders: i dunno what you consider significant, but the fact that it is both rdf vocab and syndication format is quite handy |
| 08:35:22 | <gsnedders> | Also, I don't see what advantage you get with processing RSS 1.0 as RDF/XML from a feed reader POV. |
| 08:35:44 | <gsnedders> | (Which is probably a rather provocative thing to say in here :D) |
| 08:35:55 | <gsnedders> | Now, really, I am off, before I am shot. :) |
| 08:36:24 | <kwijibo> | gsnedders: only if your feed reader has other RDF data i suppose |
| 08:36:58 | <kwijibo> | just saying, some applications do use it as RDF, even if the more commons ones don't |
| 08:37:10 | <libby> | ...which is exactly what diso are lookign at for activity streams / atom right now... |
| 08:37:29 | <libby> | well not rdf |
| 08:37:35 | <libby> | but rdf-esque (re diso) |
| 08:37:35 | <kwijibo> | diso ? |
| 08:39:31 | <kwijibo> | i think in the ml, rss was cited for its long standingness as a deployed and still dereferencable namespace, which is fair ... |
| 08:39:44 | <kwijibo> | libby: what is diso? |
| 08:40:00 | <kwijibo> | ah |
| 08:40:03 | <kwijibo> | ACTION sees |
| 08:40:11 | <kwijibo> | http://diso-project.org/ |
| 08:40:38 | <libby> | jus facing similar problems - be intersting to see what they come up with |
| 08:40:49 | <kwijibo> | B: DiSo Project Distributed Social Networking |
| 08:42:40 | <libby> | B:[this is the message what I was thinking of|http://groups.google.com/group/activity-streams/msg/26f03f1a39ca43ef] |
| 08:42:49 | <kwijibo> | on the RDF formats being parsed with XML / regex etc to get at specific bits of information, people have written screen scrapers for dbpedia |
| 08:42:57 | <kwijibo> | *on the meme of |
| 08:43:30 | <kwijibo> | which is sort of funny ... |
| 08:43:41 | <libby> | oh, really? |
| 08:43:44 | <libby> | yeah |
| 08:44:03 | <kwijibo> | yeah, it said "Easier to parse version of wikipedia" on the description |
| 08:44:17 | <kwijibo> | i dunno, something of a FAIL on both parts maybe |
| 08:46:45 | <kwijibo> | a failure on their part to understand dbpedia and RDF, and failure on RDF and dbpedia's part that developers don't understand them sufficiently to reach for an RDF parser before writing a screenscraper |
| 08:49:27 | <kwijibo> | ah |
| 08:49:28 | <kwijibo> | http://parselets.com/parselets/dbpedia/3 |
| 08:50:14 | <kwijibo> | C: a screen scraper for dbpedia! - thought of as "An easier to parse wikipedia mirror" |
| 09:20:48 | <danbri> | gsnedders, good luck with exams! |
| 09:22:02 | <danbri> | one thought re "parses RSS1 as RDF or not", one thing RDF did give RSS was rules about composition of extension namespaces, so that people could invent addon elements without having to sit around the same committee table and argue whose elements are allowed inside whose |
| 09:22:16 | <danbri> | of course in practice it all splintered into rss1, atom, rss2 etc strands :/ |
| 09:22:59 | <danbri> | at the time people didn't believe us when we said there'd be demand for syndicating arbitrary data over rss... |
| 09:23:31 | <libby> | bah |
| 09:23:42 | <libby> | ACTION shakes fist at world |
| 09:25:19 | <gsnedders> | danbri: You forget there are several versions of RSS 2.0, all of which are incompatible with one-another |
| 09:27:30 | <gsnedders> | My advice for the sane: don't ever try and write a feed reader. |
| 09:29:32 | <mhausenblas> | update on my WAC experiment from the weekend: |
| 09:29:34 | <mhausenblas> | http://webofdata.wordpress.com/2009/05/18/embedded-web-access-control/ |
| 09:29:59 | <mhausenblas> | D:| Toying around with (embedded) WebAccessControl |
| 09:32:59 | <tobyink> | dajobe's Raptor does a surprisingly good job making sense out of tag soup feeds and representing them as RSS 1.0-compliant RDF graphs. |
| 09:33:46 | <gromgull> | Is Alexander Passant here? He is apassant, right? |
| 09:33:49 | <tobyink> | Though its Atom handling could benefit from a look at bblfish's Atom-OWL work. |
| 09:33:58 | <gromgull> | .phenny, seen apassant? |
| 09:34:05 | <thosch> | gromgull: alex' nick is terraces |
| 09:34:16 | <gromgull> | ah right |
| 09:34:24 | <gromgull> | I knew that :) |
| 09:34:30 | <gromgull> | Oh well - he's not here... |
| 09:43:51 | <danbri> | http://ilrt.org/discovery/2000/11/rss-query/ |
| 09:44:08 | <danbri> | E:|RSS query, 8+ years ago flavoured |
| 09:45:07 | <libby> | I was looking for that |
| 09:52:01 | <mhausenblas> | gromgull: if it is urgent I can relay ... he's just walking around here ... |
| 09:57:32 | <mhausenblas> | danbri: re embedded WAC, IIRC you were playing around with a related idea (I guess it was during ESWC08), right? |
| 09:58:46 | <melvster_> | mhausenblas: nice blog post s/foafm/foafme |
| 09:59:27 | <melvster_> | in the github link |
| 09:59:49 | <mhausenblas> | ta |
| 09:59:51 | <mhausenblas> | done ;) |
| 10:00:49 | <melvster_> | ive realised i need to compile bruno harbulots secure delegation change into /login/ but it's already in the main index.php |
| 10:01:03 | <melvster_> | re: interaction with openid |
| 10:01:56 | <melvster_> | i think one solution is follows: you can use openid when the openid in question 1. is a URI (99.99+% of cases) 2. when it uses FOAF autodiscovery |
| 10:03:33 | <melvster_> | re: oauth, attrbiute exchange, sreg etc. you should be able to write an RDF/FOAF bridge to the relevant voacabs, tho someone would need to take on that work |
| 10:05:51 | <melvster_> | transactions is an interesting idea |
| 10:06:38 | <melvster_> | though im guessing most SPARUL updates with be in the body of a single HTTP POST, therefore would be processed relatively atomically |
| 10:08:29 | <melvster_> | but of course the "sparql server" or the "webdav server" indicated in the diagram would handle that (wevdav has locking built in) |
| 10:09:37 | <Count-Duckula> | can I do a filter of the type ..... A && (B || C ||D)... which would match AB AC AD? at the moment it seems to ignore to brackets and I get AB C D |
| 10:09:43 | <Count-Duckula> | sparq |
| 10:26:31 | <indifferen7> | hello! can any of you recommend a property for pointing at a URL where an information resource resides? for example, dbpedia uses dbpprop:url on http://dbpedia.org/page/Wikipedia to point at the url of the Wikipedia site. Is it okay to use that property or do you have any other suggestions? thanks! |
| 10:29:27 | <mhausenblas> | indifferen7: you could have a look at http://esw.w3.org/topic/AwwswGenOntDiagrams |
| 10:29:37 | <mhausenblas> | but beware ... work in progress ;) |
| 10:31:48 | <indifferen7> | ah i see..i have been thinking hard about how to make explicit that two information resources are the same. this ontology looks really interesting. are the gen:sameWorkAs/gen:InformationResource stable resources? |
| 10:31:49 | <Hory> | how about foaf:page ? |
| 10:32:33 | <indifferen7> | foaf:page would have worked but i am not focusing only on web pages but other types of information resources as well |
| 10:33:38 | <Hory> | well foaf:page has foaf:Document as it's range |
| 10:33:56 | <Hory> | foaf:Documents can be non-web-pages |
| 10:35:44 | <indifferen7> | Hory: true, while it might not be semantically correct when you read it a jpg file could probably be the object of the statement |
| 10:39:42 | <danbri> | an image is a Document, sure |
| 10:40:35 | <danbri> | "embedded WAC" mhausenblas? remind me :) |
| 10:42:23 | <indifferen7> | using foaf:page to link to any type of information resource makes my life easier so I'll stick with that. thanks for your feedback! |
| 10:42:56 | <mhausenblas> | danbri: see http://chatlogs.planetrdf.com/swig/2009-05-18.html#T09-29-41 |
| 10:43:08 | <tommorris> | xmlns.com/foaf/ isn't redirecting to /foaf/spec/ |
| 10:43:09 | <mhausenblas> | I think you did something with XMLSig in RDFa? |
| 10:43:23 | <danbri> | tommorris, it never did |
| 10:43:29 | <danbri> | try 0.1/ subdir |
| 10:43:38 | <tommorris> | as for documents: "We do not (currently) distinguish precisely between physical and electronic documents, or between copies of a work and the abstraction those copies embody." |
| 10:43:48 | <tommorris> | seems easy enough. |
| 10:44:12 | <danbri> | note that there is a giant inconsistency with that and saying that foaf:Person and foaf:Document are disjoint: |
| 10:44:13 | <tommorris> | you might want to say "this group of people are collaborating on this document" but it doesn't have to be so that the document is on the Web. |
| 10:44:29 | <danbri> | i.e. we ignore the tattood community |
| 10:44:39 | <tommorris> | heheh |
| 10:44:44 | <danbri> | if a written manuscript can be a foaf:Document, .... |
| 10:46:27 | <tommorris> | I can see it now: SELECT * WHERE { ?person a foaf:Person; foaf:made ?x . ?x a foaf:Document, tattoo:BigFuckOffEagle; tattoo:location body:chest . } |
| 10:47:00 | <tommorris> | the BMEZine folks need to get on the Linked Data train. |
| 10:47:47 | <danbri> | i think being a Document is more like a role than an objective characteristic |
| 10:47:50 | <danbri> | ditto Agent |
| 10:48:24 | <danbri> | ditto OnlineAccount |
| 10:48:28 | <danbri> | if someone can be its author, or if it can have a number-of-words property, hell sure, call it a Document |
| 10:48:38 | <tommorris> | #include <putnam_twinearth.h> |
| 10:48:53 | <danbri> | Agent in particular is quite relational. you could go quite nutty trying to objectively say which things are "really" agents ... |
| 10:48:56 | <tommorris> | #include <socialconstructivistcritique.h> |
| 10:49:01 | <kwijibo> | can it stop being a document and be something else ? |
| 10:49:01 | <danbri> | :) |
| 10:49:30 | <tommorris> | I'm wishing the University of London's examiner worked more like gcc or javac. |
| 10:49:47 | <tommorris> | Why can't I just go into the exam, see a question and write a whole bunch of include statements? ;) |
| 10:50:55 | <tommorris> | kwijibo: I think so: a piece of canvas can go from being a piece of canvas into being an Image. |
| 10:51:47 | <tommorris> | ACTION wanders off to purchase caffeine. |
| 10:51:55 | <danbri> | re artifact vs information aspects, until we have something that smells a lot like FRBR, any notion of "Document" is going to be pretty woolly |
| 10:55:41 | <mhausenblas> | gromgull: relayed message to terraces - he'll ping you |
| 10:55:57 | <mhausenblas> | ACTION off for lunch now, catch ya laters ;) |
| 10:56:14 | <gromgull> | thanks! |
| 11:45:03 | <danbri> | http://failblog.org/2009/05/17/website-fail/ |
| 11:45:27 | <danbri> | F:|Why can't phone numbers get you to websites? |
| 11:46:26 | <Shepard> | call them and ask for the website address, social engineering ;) |
| 11:52:29 | <danbri> | http://sindice.com/search?q=%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23domain%3E+*+%3Chttp%3A%2F%2Fxmlns.com%2Ffoaf%2F0.1%2FPerson%3E+&qt=advanced |
| 11:52:45 | <danbri> | G:Sindice query: properties with domain foaf:Person |
| 11:52:47 | <danbri> | G:" Search results for advanced “<http://www.w3.org/2000/01/rdf-schema#domain> * <http://xmlns.com/foaf/0.1/Person> ”, found 187 " |
| 12:05:00 | <kjetilkWork> | In Turtle, if I have a " in a literal that's quoted with """, do I still need to escape the " ? |
| 12:06:06 | <Shepard> | no |
| 12:12:05 | <kjetilkWork> | iv_an_ru_, then you have a bug :-) |
| 12:12:42 | <iv_an_ru_> | ok, will check |
| 12:15:13 | <kjetilkWork> | thanks! :-) |
| 12:16:26 | <kjetilkWork> | iv_an_ru_, should I try to come up with a minimal example? |
| 12:34:56 | <tobyink> | http://ontologi.es/rail/void |
| 12:35:13 | <tobyink> | H:|Linked Railway Data voiD |
| 12:35:48 | <tobyink> | H:Published in RDFa with server-side conversion to other formats using mod_rewrite, PHP and ARC2. |
| 12:36:23 | <tobyink> | H: e.g. http://ontologi.es/rail/void.ttl , http://ontologi.es/rail/void.json?callback=someFunction |
| 12:42:57 | <karlcow> | http://www.google.com/search?hl=en&safe=off&tbo=1&tbs=ww%3A1&q=html5&btnG=Search&tbo=1#q=quirks+mode&hl=en&safe=off&tbs=ww:1&start=0&sa=N&tbo=1&fp=L_i8LCwZOrM |
| 12:43:16 | <karlcow> | I:|graph view of search results in Google from Google itself |
| 13:05:13 | <mhausenblas> | tobyink: great to see the your void file |
| 13:06:01 | <mhausenblas> | just one thing; I'd not use stuff from http://ld2sd.deri.org/lde/void-seeds.ttl (though I put that together) |
| 13:06:18 | <mhausenblas> | as this stuff has been manually assembled |
| 13:06:56 | <mhausenblas> | I know LDE is using it (again, because I did it ;) but I'm gonna switch to OpenLink's http://lod.openlinksw.com/void/ ASAP |
| 13:07:14 | <mhausenblas> | karlcow: hehe |
| 13:08:07 | <mhausenblas> | karlcow: yeah, has been reported here already (see http://chatlogs.planetrdf.com/swig/2009-05-13.html#T15-23-59) |
| 13:08:29 | <mhausenblas> | has been a time since you have been around, right? :D |
| 13:17:43 | <mhausenblas> | phenny, tell tobyink see http://chatlogs.planetrdf.com/swig/2009-05-18.html#T13-05-22 |
| 13:17:44 | <phenny> | mhausenblas: I'll pass that on when tobyink is around. |
| 13:18:05 | <tobyink> | yes, yes. I'm here. |
| 13:18:05 | <phenny> | tobyink: 13:17Z <mhausenblas> tell tobyink see http://chatlogs.planetrdf.com/swig/2009-05-18.html#T13-05-22 |
| 13:19:22 | <mhausenblas> | heya - yeah, just to mention it ;) |
| 13:19:57 | <tobyink> | so, this would be the URI to use for dbpedia's dataset, then? http://lod.openlinksw.com/void/dbpedia/Dataset |
| 13:20:28 | <mhausenblas> | great stuff, very nice usage of RDFa and pointer to more RDF data! |
| 13:20:35 | <mhausenblas> | lemme check |
| 13:21:03 | <mhausenblas> | yep, seems so :) |
| 13:21:30 | <tobyink> | But there doesn't seem to be a corresponding one for geonames. |
| 13:21:42 | <mhausenblas> | no? hu ... |
| 13:21:58 | <mhausenblas> | any OpenLink chaps around? |
| 13:23:09 | <kidehen> | mhausenblas: yes |
| 13:23:09 | <phenny> | kidehen: 15 May 20:27Z <timbl> tell kidehen sorry no .. no colons in URIs unesaped. |
| 13:23:57 | <mhausenblas> | ahm tobyink was wondering if there is no Geonames voiD URI in http://lod.openlinksw.com/void ? |
| 13:24:39 | <melvster_> | mhausenblas: did you get rows returned from http://pastebin.com/f6d163e05 ? |
| 13:25:00 | <kidehen> | mhausenblas: is there a specific dataset with genoames to dbpedia mappings? If so, and its missing from the VoiD, its due to the fact that live system is a little out of data re. Graph Groups and Graph within Groups partitioning |
| 13:25:12 | <mhausenblas> | melvster_: yes, why? |
| 13:25:25 | <mhausenblas> | ah, thanks, kidehen |
| 13:25:40 | <mhausenblas> | I see |
| 13:25:41 | <kidehen> | mhausenblas: in the VoiD graph. What should happen is this. Each dataset source filename is now a Graph IRI within the Group DBpedia. |
| 13:25:42 | <melvster_> | hmm strange im not gettnig anything back from ARC, but that's probalby just me ... |
| 13:26:05 | <mhausenblas> | from ARC, melvster_ ? |
| 13:26:40 | <melvster_> | im trying to run the sparql with arc2, but it's not giving me any row, but i suspect im doing something wrong ... ill debug |
| 13:27:02 | <mhausenblas> | kidehen: I guess it would still be good for tobyink to use http://lod.openlinksw.com/void/geonames/Dataset |
| 13:27:02 | <kidehen> | mhausenblas: I am going to have the guys kick off a DBpedia2 reload, and post reload you should see the partitioning I describe in the VoiD graph. DBpedia doesn't take long to load |
| 13:27:12 | <tobyink> | kidehen: You may be interested in http://ontologi.es/rail/links_dbpedia.ttl |
| 13:27:19 | <kidehen> | mhausenblas: yes, |
| 13:27:25 | <kidehen> | tobyink: yes, of course :-) |
| 13:27:27 | <mhausenblas> | ACTION notes that the resource exists, it doesn't 404 |
| 13:27:41 | <kidehen> | tobyink: that we can load too, very quickly |
| 13:27:41 | <mhausenblas> | it just happens to be empty |
| 13:28:22 | <mhausenblas> | melvster_: trying again in sparql.org ... |
| 13:28:23 | <melvster_> | i guess i can run that query from *any* sparql endpoint, yes? |
| 13:28:35 | <mhausenblas> | seems fine for me melvster_ |
| 13:28:38 | <mhausenblas> | see http://sparql.org/sparql?query=prefix+rdfs%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E+%0D%0Aprefix+rdf%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%23%3E+%0D%0Aprefix+foaf%3A+%3Chttp%3A%2F%2Fxmlns.com%2Ffoaf%2F0.1%2F%3E+%0D%0Aprefix+dcterms%3A+%3Chttp%3A%2F%2Fpurl.org%2Fdc%2Fterms%2F%3E+%0D%0Aprefix+acl%3A+%3Chttp%3A%2F%2Fwww.w3.org%2Fns%2Fauth%2Facl%23%3E+%0D%0A%0D%0Aselect+distinct+%3Fresource+%3Fperson |
| 13:28:38 | <kidehen> | mhausenblas: sorry, didn't read you URI completely assumed you meant: /void/Dataset/geonames, which will exist is there is a source file with geonames to DBpedia mappings. |
| 13:28:41 | <mhausenblas> | ups |
| 13:28:52 | <mhausenblas> | kidehen: np, seems sorted |
| 13:28:54 | <mhausenblas> | ;) |
| 13:29:15 | <kidehen> | mhausenblas: what's sorted, I am getting confused here :-) |
| 13:29:18 | <mhausenblas> | anyway, good to see you chaps more or less 24/7 around :) |
| 13:29:24 | <kidehen> | mhausenblas: URI of result, then I should be clear |
| 13:29:37 | <melvster_> | ah very nice |
| 13:29:37 | <mhausenblas> | kidehen: http://lod.openlinksw.com/void/geonames/Dataset |
| 13:29:48 | <mhausenblas> | exists, doesn't 404, just is empty ATM |
| 13:30:01 | <mhausenblas> | ACTION sorry for flooding the IRC :( |
| 13:30:46 | <mhausenblas> | melvster_: yes, any sparql-endpoint should do |
| 13:31:04 | <mhausenblas> | however, if you have your local ARC2 you first have to LOAD <> it :) |
| 13:31:16 | <mhausenblas> | it wont fetch the stuff for you, AFAIK |
| 13:31:28 | <melvster_> | i think i needed to add .ttl and it worked |
| 13:31:43 | <mhausenblas> | ah, right |
| 13:31:53 | <mhausenblas> | there are conneged .rdf and .ttl versions out there |
| 13:32:17 | <mhausenblas> | maybe not all endpoints handle that stuff the way sparql.org does |
| 13:32:55 | <mhausenblas> | while I'm at it: still, whoever is out there producing SPARQL engines and is listening out here: |
| 13:33:08 | <mhausenblas> | start supporting RDFa, please :) |
| 13:33:15 | <tobyink> | OK, now my voiD has... void:objectsTarget [ is void:subset of <http://lod.openlinksw.com/void/geonames/Dataset> ] . |
| 13:33:25 | <mhausenblas> | mhm |
| 13:33:47 | <mhausenblas> | ACTION gotta run ... BIAB |
| 13:54:10 | <melvster_> | libACL checked in, example at: http://foaf.me/git/testACL.php |
| 13:55:12 | <melvster_> | code example: http://gist.github.com/113476 |
| 14:03:25 | <kidehen> | tobyink: transpose Dataset and geonames in your URI |
| 14:05:40 | <kidehen> | tobyink: so since that the URIs for accessing data sponged or uploaded from geonames.org. For the DBpedia links it will be: ../void/Dataset/DBpedia/links_geonames_en |
| 14:05:44 | <tobyink> | kidehen: So <http://lod.openlinksw.com/void/Dataset/geonames>? And also <http://lod.openlinksw.com/void/Dataset/dbpedia>? |
| 14:06:29 | <kidehen> | tobyink: the last URI show the "en" geonames data sets graph IRI under the Grp: DBpedia |
| 14:07:47 | <kidehen> | tobyink: so if data comes from geonames.org its goes under geonames grp as in: ./void/Dataset/geonames/geonames_en |
| 14:08:27 | <kidehen> | tobyink: if a DBpedia linkset then its: ./void/Dataset/DBpedia/links_geonames_en |
| 14:09:12 | <kidehen> | tobyink: covering two different data set types. Note, the Graph Grp and Graph within Grp IRIs are accessible to SPARQL |
| 14:10:02 | <tobyink> | What I'm interested in is the URI representing geonames' entire dataset. |
| 14:10:04 | <kidehen> | tobyink: so you can scope sparql to a Group IRI or specific data set IRI |
| 14:11:10 | <kidehen> | tobyink: so you should use: http://lod.openlinksw.com/void//Dataset/geonames/geonames_[en] (depending on if its specifically en or other langs. |
| 14:11:39 | <kidehen> | tobyink: http://lod.openlinksw.com/void//Dataset/geonames would be a safe URI to make subset claim against |
| 14:12:07 | <kidehen> | tobyink: so back to where we started re. transposition of the scheme you currently use |
| 14:12:28 | <kidehen> | tobyink: one / not two |
| 14:12:43 | <kidehen> | tobyink: http://lod.openlinksw.com/void/Dataset/geonames |
| 14:14:05 | <tobyink> | OK, so now I'm using <http://lod.openlinksw.com/void/Dataset/dbpedia> and <http://lod.openlinksw.com/void/Dataset/geonames> . |
| 14:14:27 | <Anchakor> | mhausenblas: how is sparql engine related to rdfa? why not extract the rdf from rdfa and query that? |
| 14:14:40 | <kidehen> | tobyink: correct |
| 14:15:35 | <kidehen> | Anchakor: when you put RDFa in a Web Page the URL of the Web Page becomes a bona fide URI for SPARQL queries, assuming the SPARQL processor includes RDFa processing capability. Example: Virtuoso :-) |
| 14:15:38 | <tobyink> | Anchakor: one might as well ask how a SPARQL engine is related to RDF/XML or to Turtle. RDFa is just one more RDF serialisation that a good SPARQL engine should support as input. |
| 14:16:28 | <tobyink> | And indeed, with RDFa being a W3C Recommendation, should probably get higher priority than other serialisations (except RDF/XML). |
| 14:16:30 | <kidehen> | Anchakor: so the RDFa just makes it easier to RDFize the host HTML page |
| 14:17:09 | <kidehen> | tobyink: even higher than RDF/XML when talking to the broader Web community where ground zero is HTML :-) |
| 14:18:38 | <tobyink> | I think if I was writing a SPARQL engine (not that I have any plans to do so) and (for some reason) not using an existing parsing library, I'd probably aim for RDF/XML first because there's so much existing data in it. |
| 14:19:14 | <tobyink> | And a SPARQL engine with no RDF/XML support would be mocked mercilessly. |
| 14:19:53 | <kidehen> | tobyink: I mean, in SPARQL you can say: select * from <URL-of-HTML-with-RDFa-inside> where {?s ?p ?o} |
| 14:22:57 | <LeeF> | kidehen, do you have any sense for how broadly supported SPARQL'ing RDFa is yet? |
| 14:23:59 | <kidehen> | LeeF: no broader than what it was when you gave the Cambridge Semantic Web Gathering presentation :-) But this will change with Google's forays into RDFa, I suspect. Ditto some SPARQL tutorials aimed at RDFa/ |
| 14:24:17 | <shellac> | select * from http://torrez.us/services/rdfa/<url> { ... } |
| 14:24:49 | <kidehen> | shellac: that's part of LeeF's presentation re. my last comment :-) |
| 14:25:57 | <Anchakor> | tobyink: ok, if document with rdfa is a valid rdf serialization, does it have a mime type so crawlers can identify it fast? |
| 14:26:23 | <kidehen> | LeeF: There should be some RDFa recipe documents that include SPARQL examples. We will make some as part of demonstrating what the Googlevocab mapper ontology (from last week) was all about. Worthwhile since it covers RDFa, RDFs, OWL, and SPARQL issues |
| 14:26:42 | <tobyink> | Anchakor: application/xhtml+xml |
| 14:27:29 | <kidehen> | Anchakor: crawlers have to be RDFa processors i.e. agents to make triples from RDFa embedded in (X)HTML docs |
| 14:27:30 | <Anchakor> | tobyink: that means that 100% there will be some rdfa in it? I guess not... that's what makes me dislike rdfa a bit |
| 14:27:37 | <kidehen> | Anchakor: s/to/that |
| 14:28:20 | <shellac> | Anchakor: you have to grab the dtd, I guess, so the first 20 or so bytes need to be read |
| 14:28:46 | <Anchakor> | kidehen: I thought the exchange serialization is standartized as rdf/xml (nevermind how I hate it) |
| 14:28:47 | <kidehen> | phenny, tell iand: any reason why relationship vocab doesn't have a property for expressing "inspired_by" |
| 14:28:48 | <phenny> | kidehen: I'll pass that on when iand is around. |
| 14:29:06 | <tobyink> | Anchakor: plain old XHTML (i.e. 1.0, 1.1) can be parsed as RDFa and should yield sensible triples. |
| 14:29:41 | <kidehen> | Anchakor: when dealing with URIs the representation is negotiable, RDF/XML is but one representation |
| 14:30:01 | <tobyink> | kidehen: you could try <http://vocab.sindice.com/xfn#muse>. |
| 14:30:09 | <Anchakor> | I thought that one of advantages of semantic web, was that you don't have to parse the tons of text to get some data |
| 14:30:17 | <shellac> | 'There SHOULD be a @version attribute on the html element with the value "XHTML+RDFa 1.0"' |
| 14:30:27 | <Anchakor> | turns out RDFa is undoing that effort |
| 14:30:58 | <tobyink> | Anchakor: most RDF serialisations allow for some of the bytes in the file to be non-triples. |
| 14:31:18 | <kwijibo> | Anchakor: don't think that's ever been a stated goal that I've seen ... |
| 14:31:33 | <tobyink> | e.g. in N3/Turtle/N-Triples, you can have comments (prefixed with #). |
| 14:31:55 | <tobyink> | In RDF/XML, TriX and other XML-based RDF serialisations, there are <!-- XML comments -->. |
| 14:32:21 | <tobyink> | And in RDFa, there are <!--XML comments-->, *plus* elements which simply don't generate any triples. |
| 14:32:38 | <kwijibo> | maybe json is the only one with comments |
| 14:32:44 | <kidehen> | tobyink: see: http://data.openlinksw.com/about/html/http://vocab.sindice.com/xfn#muse (which is using SPARQL and RDFa Cartridge underneath) |
| 14:32:45 | <kwijibo> | *without |
| 14:33:08 | <LeeF> | kidehen, ok, i was hoping it had improved since then :-) :-/ |
| 14:33:37 | <Anchakor> | tobyink: yes, but sooner or later the people would want to query very large heterogenous datasets (WoD), which would put lot of demand on speed, and I would guess that RDFa is serialization which takes the longest to parse... (for now) |
| 14:33:55 | <kidehen> | tobyink: or: http://proxy.uriburner.com/about/html/http://vocab.sindice.com/xfn#muse |
| 14:34:32 | <kwijibo> | I dunno, looking at bengee's code the RDF XML parser has the most lines |
| 14:34:36 | <kidehen> | tobyink: proxy | linkeddata are the same machine re. uriburner.com and the sparql endpoint is at: http://proxy.uriburner.com/sparql |
| 14:34:51 | <kwijibo> | so i'd guess it's maybe the most intensive to parse |
| 14:35:23 | <tobyink> | Anchakor: some serlialisations are better suited for certain purposes than others. That's one of the reasons why we have more than one serialisation. |
| 14:35:43 | <shellac> | Anchakor, I wouldn't guarantee that it's disproportionately slower to parse. And there are cases like SVG where RDF/XML may be hidden in the depths of a document |
| 14:35:44 | <kidehen> | LeeF: might be worth a callout to other SPARQL players re. RDFa hooks etc.. |
| 14:37:27 | <Anchakor> | tobyink: well then the question is, what do we do when the number of serializations starts to grow really large? IMHO using one serialization as standard and export to it the other serializations isn't bad (though I would pick something else then rdf/xml :)) |
| 14:38:43 | <LeeF> | kidehen, I need to focus my efforts on the new SPARQL work instead :) |
| 14:39:15 | <kidehen> | LeeF: callout means: just asking right here: anyone else using RDFa with their SPARQL processor :-) |
| 14:39:48 | <shellac> | Anchakor: I think the real concern is that the simple function model.read(url) used to complain quickly about most urls, but now it fails silently |
| 14:39:52 | <tobyink> | Anchakor: for dumps, probably N-Triples, but I imagine there are even more efficient storage mechanisms if you throw off the ASCII shackles. |
| 14:40:25 | <tobyink> | Following query works nicely in Virtuoso: ASK FROM <http://ontologi.es/rail/stations/gb/LWS.rdf> WHERE { <http://ontologi.es/rail/stations/gb/LWS> a <http://ontologi.es/rail/vocab#Station> . } |
| 14:40:30 | <kwijibo> | Anchakor: i guess it's kind of like with vocabularies etc - the useful ones survive and are supported, the others fall by the way side |
| 14:41:46 | <kwijibo> | tobyink: is N Triples ascii only ? |
| 14:42:06 | <tobyink> | Indeed it is, but that's not exactly what I meant. |
| 14:42:25 | <kwijibo> | yep www.w3.org/2001/sw/RDFCore/ntriples/ |
| 14:42:28 | <kwijibo> | US-ASCII |
| 14:42:29 | <Anchakor> | kwijibo: yeah, but rdf provides means to create mapping between the ontologies |
| 14:42:50 | <kwijibo> | GRDDL kind of lets you map from XML serialisations at least |
| 14:43:33 | <tobyink> | I was rather referring to things like hash tables and linked lists and all those other internal representations that RDBMS developers have nightmares about. |
| 14:46:50 | <kwijibo> | in publishing RDF, I'm moving more turtle as first priority |
| 14:46:58 | <kwijibo> | *more towards |
| 14:48:28 | <kwijibo> | or RDFa |
| 14:49:15 | <Anchakor> | readability: turtle > rdf/xml > rdfa |
| 14:50:37 | <tobyink> | Anchakor: in terms of looking at the source code, RDFa is not especially readable compared to some other serialisations, but in terms of what you see in-browser is often the best. |
| 14:50:40 | <kwijibo> | yep - actually no hybrid HTML + RDF approaches are very readable - too much going on |
| 14:52:28 | <Anchakor> | tobyink: well sometimes yes, but sometimes you want to want to use the publishers Model and Controller but make your own View |
| 14:52:54 | <Anchakor> | s/want to want/want/ |
| 14:53:28 | <idmclean> | Greetings, swig. |
| 14:55:15 | <idmclean> | Anchakor, Check this out http://web.cs.mun.ca/~ulf/pld/index.html |
| 14:56:42 | <Anchakor> | hi |
| 14:57:31 | <idmclean> | Anchakor, I decided to drop my normal screenname "KickAssClown". |
| 14:57:53 | <idmclean> | At least while I'm lurking around the swig channel. |
| 14:58:14 | <kwijibo> | tobyink: what is the state of play with your cascading semantic sheets ? |
| 14:58:19 | <Anchakor> | aah :) Im looking at http://web.cs.mun.ca/~ulf/pld/structure.html#Top really interesting |
| 14:59:12 | <idmclean> | I was thinking the work of Ulf would be a good resource for designing and developing a programming language schematic from. |
| 15:00:27 | <tobyink> | kwijibo: My lastest draft is at <http://buzzword.org.uk/2008/rdf-ease/spec-20090102.html>. I don't have any immediate plans for a new draft - not because I don't think it's an interesting/worthwhile project, but rather because I can't think of anything that needs desperately adding to it. |
| 15:00:53 | <kwijibo> | and implementations ? |
| 15:01:08 | <idmclean> | logger, pointer? |
| 15:01:22 | <idmclean> | logger, pointer. |
| 15:01:33 | <tobyink> | Swignition implements it <http://buzzword.org.uk/swignition/try>. |
| 15:01:42 | <kwijibo> | ah - should work with any GRDDL client ? |
| 15:01:45 | <kwijibo> | or not ? |
| 15:02:28 | <tobyink> | kwijibo: GRDDL theoretically works with any "transformation language", but in practice, most GRDDL clients only do XSLT. |
| 15:03:04 | <tobyink> | RDF-EASE essentially just defines a new transformation language and says how it works with GRDDL. |
| 15:03:19 | <Anchakor> | idmclean: yes, will take a lot of time to absorb such material for me :) |
| 15:05:00 | <logger> | See http://chatlogs.planetrdf.com/swig/2009-05-18#T15-05-09-1 |
| 15:05:00 | <logger> | See http://chatlogs.planetrdf.com/swig/2009-05-18#T15-05-09-4 |
| 15:06:17 | <mattl> | tobyink: i'd like to export libre.fm data in RDF.. is that useful? |
| 15:08:12 | <idmclean> | Anchakor, I'm walking through the primer which mhausenblas linked me to. My aim at the moment is to define the basic shape of the five core dialects: declarative (RDF schema and rdf-representation of code), imperative, functional, semantic, and automata. |
| 15:08:18 | <tobyink> | mattl: http://alpha.libre.fm/rdf.php?page=/users/tobyink/recent-tracks |
| 15:08:53 | <tobyink> | Oops - http://alpha.libre.fm/rdf.php?page=/user/tobyink/recent-tracks |
| 15:09:55 | <mattl> | tobyink: nice. so, could you adapt the existing users-dump script to export in rdf as well? |
| 15:10:38 | <idmclean> | Anchakor, I need to define the class of thing "dialect" and the relationship of the dialects to one another. The relationship is represented as a set of triples linking all five dialects to each other. |
| 15:10:57 | <tobyink> | mattl: As a shell script? I'm not that much of a masochist. |
| 15:11:32 | <tobyink> | It's doable, but I'd think Perl, Python or PHP would be easier. |
| 15:11:43 | <mattl> | timbl: what was your csv2n3 thing? |
| 15:11:57 | <mattl> | tobyink: if we can get hold of the script tim has, we could just use that, i guess. |
| 15:12:03 | <tobyink> | PHP of course has the advantage that we can re-use existing nixtape code. |
| 15:12:23 | <mattl> | i'm not sure how people feel about php as shell script |
| 15:12:35 | <Anchakor> | idmclean: I don't know if this would be possible to describe this formally precisely, since the paradigms basically are different abstractions |
| 15:12:54 | <tobyink> | Anyone running nixtape and gnukebox is bound to have PHP installed anyway. |
| 15:17:45 | <tobyink> | idmclean: another parse of KickAssClown could be a clown that is being ordered to kick a donkey. |
| 15:19:04 | <idmclean> | Anchakor, for now all I need to describe is a "dialect" independent of the specific details of the different classifications of languages. A dialect is an abstraction representing a set of grammar, syntax, and semantics. I don't think we need to go beyond simply defining that there is a class of thing "dialect" and leaving it at that for now. Different dialects will have their own abstractions specific to their purpose, but that wil |
| 15:19:30 | <tobyink> | ... but that will <truncated>. |
| 15:24:55 | <idmclean> | tobyink, what will be <truncated>? I'm unsure as to whether you are refering to something mattl or yourself said earlier, or to what I just said. |
| 15:25:38 | <tobyink> | idmclean: Your last message was truncated. See: http://chatlogs.planetrdf.com/swig/2009-05-18.html#T15-19-14 |
| 15:25:44 | <kidehen> | mattl: what are u trying to achieve? A dump of MySQL data in RDF form? |
| 15:25:58 | <idmclean> | but that will be defined latter as we know more about those dialects. |
| 15:35:05 | <idmclean> | In the ontology tlol, tlol:dialectic a tlol:language? |
| 15:35:40 | <idmclean> | err dialect is a subclass of some language. |
| 15:37:12 | <idmclean> | so to define a language as a class we would declare "tlol:Language a rdfs:Class"? |
| 15:38:29 | <Anchakor> | yep |
| 15:39:32 | <idmclean> | "tlol:Language a rdfs:Class; rdfs:subClassOf :Dialect ."? |
| 15:40:12 | <Anchakor> | yes, thought you probabaly meant tlol:Dialect |
| 15:40:14 | <idmclean> | err ""tlol:Dialect a rdfs:Class; rdfs:subClassOf tlol:Language ."? |
| 15:41:11 | <Anchakor> | hmm I don't understand what you mean by dialect really - you mean like that Scheme is dialect of Lisp? |
| 15:42:28 | <idmclean> | http://en.wikipedia.org/wiki/Programming_language_dialect |
| 15:49:22 | <idmclean> | Anchakor, in this case, a dialect is a sub-grammar to the langauge of languages. In the model I'm using, we don't strictly define an overarching language, we instead define small specific grammars and semantics and design them to be mappable to one another. |
| 15:51:03 | <Anchakor> | idmclean: yes, that is good approach imho |
| 15:59:27 | <idmclean> | I'm a little confused. The primer I'm looking at is for N3. The rdf and rdfs are not in N3 notation, correct? |
| 16:01:39 | <Anchakor> | they are, n3 covers more than rdf |
| 16:06:35 | <idmclean> | Anchakor, I need the primer for RDF-XML or Turtle. The more expressive a dialect is beyond it's purpose, the less appropriate it is for TLoL. |
| 16:08:18 | <Anchakor> | idmclean: http://www.w3.org/TeamSubmission/turtle/ |
| 16:08:34 | <tobyink> | N3 allows a superset of RDF. RDF/XML only supports a subset. |
| 16:09:02 | <tobyink> | Turtle, N-Triples and RDFa should support all of RDF. |
| 16:11:39 | <Shepard> | does n-triples define any unicode escaping? |
| 16:12:16 | <Anchakor> | tobyink: charsets are defined on level of serializations, or rdf mentions that it serializations should use unicode to be fully compliant? |
| 16:12:22 | <kidehen> | tobyink: I think you're a version behind re. mapper ontology for googlevocab |
| 16:12:37 | <kidehen> | tobyink: re. what's listed at: http://ontologi.es/ |
| 16:13:07 | <tobyink> | kidehen: it looks like I am. Just a moment... |
| 16:14:17 | <kidehen> | tobyink: there should be owl:equivalenClass mappings from each potential google URI at class and property levels. Covering "/", #, and '/#' terminated URIs. |
| 16:16:06 | <kidehen> | tobyink: so you should have google, google2, and google3 prefixes if up to date |
| 16:17:06 | <idmclean> | tobyink, I orignally started the design of tLoL with XML as the declarative dialect because it's easy to parse the xml from the data/instructions. The declarative dialect is going to be used to declare the definition of the other dialects. |
| 16:18:29 | <kidehen> | tobyink: example: http://lod.openlinksw.com/describe/?url=http%3A%2F%2Fpurl.org%2FNET%2Fgooglevocab%23Person&sid=60252 |
| 16:18:44 | <kidehen> | tobyink: the whole thing: http://lod.openlinksw.com/describe/?url=http%3A%2F%2Fpurl.org%2FNET%2Fgooglevocab%23&sid=60252 |
| 16:20:09 | <bengee> | ACTION waves from nyc, checking if hotel wifi is stable enough for irc |
| 16:20:33 | <tobyink> | kidehen: http://purl.org/NET/googlevocab updated |
| 16:21:00 | <kidehen> | tobyink: great! |
| 16:21:03 | <PovAddict> | ACTION waves back at bengee to acknowledge his messages are getting thru |
| 16:21:17 | <bengee> | heya, thx |
| 16:21:22 | <tobyink> | bengee: I will probably have more RSS 1.0 patches for ARC2 soon. |
| 16:21:29 | <bengee> | cool |
| 16:22:10 | <bengee> | guess it's time for a public code repository |
| 16:22:13 | <tobyink> | The aim being to take subjects which aren't rss:items or rss:channels and nest them in the XML. |
| 16:23:54 | <idmclean> | Anchakor, the declarative dialect has three main concerns as I see it: declaring schema(grammar, semantic, object, state, function, and operator definitions),... |
| 16:24:05 | <idmclean> | Anchakor, ....declaring problem/solution implementations/instances (source code in potentially two forms: Machine-readable and Human-readable), declaring relationship mappings between nodes. |
| 16:27:39 | <idmclean> | Anchakor, the question that arises is what declarative language would perform these functions without clashing with the other dialects? XML was my first choice because it simply declared a node. RDF-XML I expect would serve similar purpose, declaring nodes, and relationships between nodes. |
| 16:28:44 | <idmclean> | Anchakor, I'm unfamiliar with Turtle, and N3. What would be the pros and cons of XML, RDF-XML, Turtle, N3, and others? |
| 16:30:48 | <Anchakor> | idmclean: these are some pretty deep and complex problems... but if I were you I would just learn and use rdf - serialization doesn't matter |
| 16:35:22 | <idmclean> | Anchakor, would you care to elaborate on what you mean by "serialization doesn't matter"? I fail to understand. |
| 16:36:48 | <kwijibo> | RDF is a model that can be serialised in different ways |
| 16:36:58 | <Anchakor> | idmclean: rdf isn't a file format, rdf is a way to express information as directed graphs, which can be serialized into a file format or stored in triple store |
| 16:38:05 | <kwijibo> | N3 is turtle + rules + path language + some other stuff |
| 16:38:13 | <kwijibo> | i'd go with turtle |
| 16:38:51 | <Anchakor> | yeah, me too |
| 16:39:01 | <kwijibo> | or maybe start with n triples |
| 16:39:11 | <kwijibo> | or rdf/json |
| 16:39:32 | <kwijibo> | since there is no syntactic variation to worry about until you understand the model |
| 16:44:56 | <idmclean> | The way I concieve of the format of a expression in tLoL is thus <object>main <expression> if (<condition><variable>k</variable> == false</condition>)<expression>return false</expression></expression></object> |
| 16:46:56 | <Anchakor> | idmclean: no idea what you mean by that now :) |
| 16:49:31 | <idmclean> | Say we have a function: main(){print("Hello, World!"} |
| 16:51:02 | <idmclean> | How would you mark that up in RDF/XML so that you could pass it through a parser to arrive at the same expression in a different language? |
| 16:52:25 | <idmclean> | Meh, I'll just learn RDF-XML for now. |
| 16:53:57 | <idmclean> | I'm pretty sure my concept will work in RDF-XML, I have no idea how the other forms of RDF will effect the form of my concept. |
| 16:54:13 | <Anchakor> | idmclean: wait a few weeks, I hopefully have holidays soon and will work on this :) |
| 16:54:53 | <Anchakor> | idmclean: if I were you I would forget all about xml... graphs > trees :) |
| 16:55:18 | <kwijibo> | idmclean: the point is that the other forms of RDF will not affect your concepts |
| 16:55:32 | <kwijibo> | because they all express the same model |
| 16:55:43 | <idmclean> | Anchakor, I agree. Graphs are awesome. tLoLs is designed based on a 5-cell pentachron graph. |
| 16:56:10 | <idmclean> | kwijibo, unaltered. Yes. |
| 16:56:56 | <idmclean> | http://en.wikipedia.org/wiki/Polyglot_(computing) |
| 16:57:52 | <Anchakor> | idmclean: polyglots are merely a curiosity, nothing useful really :) |
| 16:58:11 | <idmclean> | This is one of the concepts underlying tLoL. A given program in tLoL will end up being a polyglot of dialects. |
| 16:59:56 | <Anchakor> | that reminds me I have name for my project too - Trish (triple shell) - executable rdf :) |
| 17:00:49 | <Anchakor> | well not really executable rdf, but rather rdf based code generating system |
| 17:00:55 | <PovAddict> | idmclean: http://mauke.ath.cx/stuff/poly.poly |
| 17:01:51 | <idmclean> | When the langauge performs analysis, breeding, mutation, and selection, it will perform transformations on various families of polyglots to derive new families of polyglots. I am hoping that through these transformations, I can get the langauge to find the differences between it's dialects to develop new dialects for use. |
| 17:03:15 | <idmclean> | PovAddict, that has brainfuck in it. |
| 17:03:29 | <PovAddict> | it has 15 languages |
| 17:05:48 | <idmclean> | I'm concerned with syntax/semantic colissions that might result from using alternative declarative dialects. |
| 17:39:25 | <mhausenblas> | http://dret.typepad.com/dretblog/2009/05/rest-and-rdf-granularity.html |
| 17:39:38 | <mhausenblas> | L: REST and RDF Granularity |
| 17:40:15 | <mhausenblas> | L:| REST and RDF Granularity |
| 17:40:24 | <PovAddict> | L1:"" |
| 17:40:46 | <mhausenblas> | thanks, but I was just about to replace it with something meaningful :D |
| 17:41:28 | <mhausenblas> | L1: blog post contemplating about the tension between RDF-based semantic web approaches and RESTful architectures |
| 17:41:34 | <mhausenblas> | L: blog post contemplating about the tension between RDF-based semantic web approaches and RESTful architectures |
| 17:42:00 | <alex275> | Tim Berners-Lee on the next Web http://www.ted.com/talks/tim_berners_lee_on_the_next_web.html |
| 17:42:16 | <mhausenblas> | L: see also his REST tutorial from WWW09, http://dret.net/netdret/docs/soa-rest-www2009/ |
| 17:43:26 | <mhausenblas> | cygri - this is the chap I was talking about and who I'd like to invite here to repeat his tutorial ... |
| 17:43:51 | <PovAddict> | argh @ that blog having comments the wrong way (newest first) |
| 17:55:45 | <idmclean> | mhausenblas, thank you for sharing that link. I think I figured out where I've been stumbling for the past couple of hours. |
| 18:00:02 | <mhausenblas> | idmclean: yw ;) |
| 18:00:10 | <mhausenblas> | btw, I like your new nick ;) |
| 18:00:21 | <mhausenblas> | not that the old one was not good, but ... :D |
| 18:00:34 | <mhausenblas> | anyway, time to run ... cya all tomorrow |
| 18:00:46 | <mhausenblas> | good nite Web of Data |
| 18:03:04 | <Anchakor> | idmclean: how should this nick be parsed? :) |
| 18:03:24 | <idmclean> | I DM |
| 18:03:34 | <PovAddict> | Ian Mclean |
| 18:03:39 | <idmclean> | Otherwise, it's just Ian Douglas Mclean. |
| 18:04:17 | <Anchakor> | ah, I thought it wasn't Dungeon Master :) |
| 18:09:57 | <idmclean> | What is the difference between Turtle and N3? |
| 18:16:17 | <Anchakor> | turtle is made just for rdf, n3 has some things added like rules |
| 18:17:35 | <mischat> | idmclean: http://www.w3.org/DesignIssues/diagrams/n3/venn |
| 18:18:55 | <idmclean> | What are ?var and {graph literals}? |
| 18:20:15 | <mischat> | ?var as in a variable in sparql, and {graph literals}, i think are when in n3 you have literal values as the subject of a triple, but i dont do much n3 |
| 18:20:54 | <idmclean> | rules has a lot of meanings for me. I deal in games, programming, meta/hyper rules, linguistics, etc. Could you specify the context of these rules, what do they do? |
| 18:22:06 | <mischat> | nope, read http://www.w3.org/DesignIssues/Notation3.html |
| 18:22:27 | <idmclean> | Pardon, mischat. I intended that question for Anchakor. |
| 18:27:24 | <Anchakor> | idmclean: I think rules are important too, but I'd rather express them _in_ rdf |
| 18:29:42 | <idmclean> | the design issues for N3 gives {?x} => {?x} as syntax of a rule. |
| 18:30:42 | <Anchakor> | yep, I haven't really ever learned n3 so I probably won't be able to help you understand though |
| 18:31:17 | <Anchakor> | idmclean: http://www.w3.org/2000/10/swap/doc/Rules |
| 18:43:35 | <idmclean> | Anchakor, I think rule in this context means the same think as a rule in BNF. |
| 19:03:07 | <gsnedders> | http://philip.html5.org/docs/rdfa/ |
| 19:03:47 | <gsnedders> | M: Unfinished proposal (with processing rules for both text/html and application/xhtml+xml) for RDFa in HTML 5. |
| 19:03:56 | <Philip`> | gsnedders: Hmm, looks like a load of rubbish to me |
| 19:04:05 | <Philip`> | (but I may be biased) |
| 19:04:06 | <gsnedders> | Philip`: Since when were you here? |
| 19:04:14 | <Philip`> | gsnedders: Since I joined |
| 19:04:21 | <gsnedders> | Philip`: When was that? :P |
| 19:04:37 | <Philip`> | gsnedders: 18:05 BST |
| 19:04:54 | <gsnedders> | Philip`: n00b. |
| 19:06:55 | <PovAddict> | Philip`: load of rubbish? |
| 19:07:20 | <PovAddict> | like html5? *runs* |
| 19:07:39 | <gsnedders> | PovAddict: I think Philip` would say any document he authors is rubbish, though |
| 20:09:57 | <PovAddict> | can anyone recommend wiki software? |
| 20:13:21 | <PovAddict> | a more specific question: does anyone know of wiki software supporting Atom Publishing Protocol? |
| 22:03:07 | <snail> | PovAddict: see http://semanticweb.org/wiki/Semantic_MediaWiki |
| 22:19:17 | <libby> | http://www.triplescape.com/doapamine/ |
| 22:19:36 | <libby> | N:|DOAPamine is set of Java 5 Annotations that can be used to insert DOAP data into Java classes. Also included is an Ant task to reflect on the annotations and produce DOAP in RDF/XML. |
Back to channel and daily index: content-negotiated html turtle