- From: Steve Harris <steve.harris@garlik.com>
- Date: Wed, 29 Apr 2009 15:06:50 +0100
- To: Kingsley Idehen <kidehen@openlinksw.com>
- Cc: Semantic Web <semantic-web@w3.org>, foaf-dev Friend of a <foaf-dev@lists.foaf-project.org>
On 29 Apr 2009, at 14:57, Kingsley Idehen wrote: > Steve Harris wrote: >> On 29 Apr 2009, at 14:06, Kingsley Idehen wrote: >>> Steve, >>> >>> If we isolate the "FOAF Profiles" bubble of the LOD-Cloud >>> pictorial, would you say these sources are representative: >>> >>> 1. http://esw.w3.org/topic/FoafSites >>> 2. http://pingthesemanticweb.com (PTSW) >>> 3. http://sindice.com >> >> It seems highly unlikely. >> >> The only way to get a representative sample is to select some of >> the data randomly. > Okay. > > So I end this thread by asking: isn't that basically what we have in > our instance? Its data comes from the sources above plus others. I think I wasn't clear enough, I meant "...is to select some of the data [from the whole population] randomly". Given that the 3 sources above don't have 100% coverage, and are biased in their own ways that can't be the case. Anyway, that's really offtopic, my point was that if you have a large enough slice of the FOAF data you can see that the majority of the data has bNodes in it. >> ESW links a human-curated selection of sites, PTSW gets fed >> similarly similarly and Sindice crawled, IIUC. >> >> I don't think anyone even has a good idea of how many FOAF files >> are out there, to know if they have a good selection or not. I >> think we have 12 million or so unique ones, but we know there's an >> awful lot more out there. >> >> Ontop of that, "FOAF" is especially vague, eg. do qdos.com profiles >> (eg. http://qdos.com/user/Steve-Harris/18b6f60b41e05aaa418565ebfe901d6b/turtle) >> count as FOAF profiles? They have foaf:People in them, and use one >> or two foaf properties, but foaf: is not the most common prefix. >> >> What about DOAP files with lots of FOAF in them? Some use foaf: >> more than doap:, and so on. > DOAP files are picked up from PTSW and a few other data sets that > use FOAF. Some of them are, yes. > Maybe we chat by phone of private IM (IRC, Twitter, Identi.ca etc > about this) ? If you like, I'm on #swig, but I'm AFK quite a bit this afternoon, so it might be patchy. - Steve
Received on Wednesday, 29 April 2009 14:07:28 UTC