Experimental IRC log sioc-2009-06-10

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.

00:21:41<Shepard>CaptSolo: sioc:topic or sioc:about? but I get the feeling I'm missing something about the problem :)
00:22:45<CaptSolo>Shepard: not necessarily (re missing)
00:22:55<CaptSolo>well, we don't know details
00:23:04<CaptSolo>sioc:topic 'd be my first pick
00:23:10<Shepard>charles2: welcome back :)
00:23:18<CaptSolo>have to figure when 'd we use sioc:about
00:23:27<CaptSolo>ah, charles2 is here again :)
00:23:49<charles2>howdy CaptSolo
00:25:48<CaptSolo>charles2: you might have received my email reply re. questions asked here
00:25:59<charles2>CaptSolo thanks for the message earlier. I am trying to extend tabulator firefox extension to support distributed microblogging and I was trying to figure out how I should represent a message that is intended for someone or about them
00:26:00<CaptSolo>provided i got the right email address for you
00:26:12<CaptSolo>wow, interesting!
00:27:24<Shepard>wasn't there a proposed "recipient" or something similar?
00:27:56<charles2>its pretty fun. The best I came up with was to use the reply_of to say that it is intended for a person and then i determine that if it is a microblogpost, it is a reply to a message, and if it is to a user it is a message intended to be for someone ( like the twitter @ mention)
00:28:01<Shepard>charles2: if it's really about a person then I'd go for sioc:about
00:29:14<CaptSolo>Shepard: might be good to discuss on the mailing list
00:29:20<CaptSolo>to see what other suggestions people have
00:29:35<CaptSolo>(unless you need just a "quick and dirty" solution)
00:30:17<charles2>I'd be interested in playing around with it :) Its for my undergrad research project so more learning is better.
00:30:49<CaptSolo>and you can experiment / look at different ways to do it then :)
00:31:17<CaptSolo>Shepard: it's just that i am uncertain now if sioc:about is a good solution for this
00:31:42<CaptSolo>Shepard: yes, there was a proposed recipient property, but i don't see it in the ontology
00:32:55<CaptSolo>charles2: the recipient property would be best fit for indicating whom the messages is intended for
00:33:12<CaptSolo>(if it existed. but we will need to add it sooner or later)
00:33:19<Shepard>I really love sioc:about btw, it's a very flexible property with many use cases and I've even created sub-properties of it for stuff in my uni project :)
00:33:33<charles2>:D
00:33:41<CaptSolo>here is some old discussion re recipient http://sioc-project.org/node/232
00:34:05<CaptSolo>(a copy of mailing list posts, should have same conversation on official SIOC-Dev archives too)
00:34:23<CaptSolo>addressed_to and has_recipient were proposed
00:34:52<CaptSolo>in fact, i wonder how can email and mailing list messages be fully represented in SIOC w/o such a property
00:35:05<CaptSolo>should ask Wikier what were they using (if anything)
00:35:30<CaptSolo>Shepard: sounds interesting re. subproperties and uses of sioc:about
00:35:44<CaptSolo>super-generic properties FTW?
00:36:05<Shepard>yep!
00:36:20<CaptSolo>ah, sioc:about is indeed a cool property
00:37:03<Shepard>I think this is an important principle in ontology design: if you see something that you think might be a problem in other areas as well, then make you solution as generic as possible so that it can be re-used elsewhere
00:37:17<CaptSolo>can be used to e.g. link together a post and RDF data extracted from it (from RDFa, etc)
00:37:53<CaptSolo>Shepard: an extreme would be to create an over-generic super-property that can be used for anything
00:38:04<Shepard>heh
00:38:09<CaptSolo>x:any_kind_of_relation
00:38:21<CaptSolo>heh, we've got sioc:related too :D
00:38:37<CaptSolo>and sioc:links_to (which is quite specific though, compared with the rest of properties mentioned here)
00:39:06<CaptSolo>generic properties make ontology more reusable
00:39:14<Shepard>well, an example: http://www.bbc.co.uk/ontologies/programmes/2009-04-17.shtml#genre - as much as I like the programmes ontology, properties like genre or tag are just general things that are neeed everywhere and so they shouldn't be restricted to programmes
00:39:29<CaptSolo>but - not sure what is the semantics of a microblog post
00:39:43<CaptSolo>from the message we can say for sure that a nick is mentioned in a message
00:40:14<CaptSolo>(perhaps SMOB has more detailed info on this, i am considering twitter / identi.ca for the moment)
00:40:30<Shepard>there's also sioc:links_to :)
00:40:31<CaptSolo>and we can also infer that a message is in a reply to someone
00:40:51<CaptSolo>though this inference i a bit shaky
00:40:57<Shepard>oh, you mentioned links_to already
00:41:24<CaptSolo>e.g. twitter broke replies (#fixreplies) and now some people work around it by putting some chars in front of the nich
00:41:29<CaptSolo>s/nich/nick
00:41:35<CaptSolo>such as ".@CaptSolo"
00:42:13<CaptSolo>but, ok, assuming we know something is a reply to somebody else, we would ideally have this has_recipient property
00:42:19<CaptSolo>might be a subproperty of sioc:about
00:42:40<charles2>so the problem i saw was that if the microblogs are truly distributed, I need to somehow embed the recipient. Right now I have reply_of then (once i get the uri deferencing to work without making my fan go crazy) figure out whether the reply is to a user or to a microblog post
00:42:42<CaptSolo>if we just know that a message mentions someone's nick, how would we express that in SIOC?
00:43:03<CaptSolo>it is not really about the person, strictly speaking it just mentions a person
00:44:04<CaptSolo>charles2: from domain/range of http://rdfs.org/sioc/spec/#term_reply_of you'd see it is a relation between items
00:44:20<CaptSolo>charles2: not entirely correct to use it b/w an item and a person or a user
00:44:39<Shepard>unless you see yourself as an item :)
00:44:54<Shepard>"post me, baby!"
00:45:00<CaptSolo>charles2: indeed, in the distributed scenario info about a recipient 'd be valuable to have
00:45:06<CaptSolo>Shepard: hehe :D
00:45:21<CaptSolo>Shepard: they might be defined as disjoint. can't remember now.
00:46:32<charles2>but is a User a person? A person holds a User right?
00:46:50<CaptSolo>charles2: btw, sioc uses inverse properties, that is, it has relations in both directions
00:47:23<CaptSolo>which might be seen as a bit redundant as you could express the same knowledge using just one of them
00:47:50<charles2>actually i found it rather useful–at least when i was trying to figure out what was going on in tabulator
00:47:53<CaptSolo>charles2: in the pair has_reply / reply_of the former is used more
00:48:18<CaptSolo>charles2: useful to have inverses?
00:48:44<CaptSolo>how do they help debug / track what is going on in tabulator?
00:48:56<CaptSolo>charles2: a person hold a user account
00:49:34<CaptSolo>Shepard++ /me still laughing about the "post me, baby!" :D
00:50:22<CaptSolo>charles2: see if you want to use has_reply instead
00:50:55<Shepard>;)
00:51:35<charles2>CaptSolo: is has_reply appropriate? I have a MicroblogPost which contains the text itself and links to the user / microblogpost it replies to
00:56:03<charles2>if i understand correctly has_reply points to something which has replied to the item in question?
01:09:36<CaptSolo>charles2: that is the tricky bit, and a reason why we have inverses
01:10:05<CaptSolo>because people think of different names for relation depending on its direction
01:10:24<CaptSolo>while you could use the same property and just reverse its operands
01:11:34<CaptSolo>if X is a mblogpost that links to another microblogpost Y it replies to, then
01:12:07<CaptSolo>how do you currently express that relation?
01:13:58<charles2>CaptSolo: currently i have a microblog X and in it i replies to another post Y
01:28:29<CaptSolo>charles2: how would that relation look in RDF?
01:31:13<charles2>CaptSolo: oh sorry, unexpected visit from a friend. one sec
01:35:58<charles2>http://pastebin.com/d2c98c3eb
01:36:52<charles2>all of the users and posts are in one file at the moment for simplicity
01:37:08<charles2>so #p1 is another post
01:48:57<CaptSolo>the <#Michael> part is not correct re what we talked about before
01:49:04<CaptSolo>but putting that aside...
01:50:03<CaptSolo>so there is #s1 sioc:reply_of #p1
01:50:25<CaptSolo>which translates as "s1 is a reply to p1"
01:50:55<CaptSolo>now, it will look different in the "tree-like" structure that you used
01:51:08<CaptSolo>but you could express the same relation also as
01:51:21<CaptSolo>#p1 sioc:has_reply #s1
02:03:06<charles2>CaptSolo: cool :) thanks for explaining
02:04:05<charles2>oh so, if i wanted to put the #p1 sioc:has_reply #s1 I would have put that under #p1, correct?
02:15:08<CaptSolo>yes
02:15:27<CaptSolo>the relation is inverse so object and subject swap places
02:16:09<CaptSolo>ok, later. time to sleep.
02:19:45<charles2>alright
02:19:53<charles2>thanks for all the help! :)
15:34:47<charles2>CaptSolo: I just realized something. Currently I have a User that has a microblog post. What i can, and probably should do, is have a Microblog have the Microblog posts, then use the reply_of to point to the microblog.
15:40:32<charles2>that relation could represent the equivalent of @ replies
15:41:30<charles2>It would be interesting to know how SMOB does it

Back to channel and daily index: content-negotiated html turtle