Revyu Revyu / Items

Ideas

Get Feed
 It would be great to be able to bring in Revyu data to Twine. Something to think about.

Comments

  • Public Comments

    • 21 months ago


      We are working on the ability to start consuming external data. Are you guys using SIOC now? Just curious. It's still something we are thinking about, but I'm interested in exploring this.
      Revyu
    • 20 months ago


      OK, general thoughts on Linked Data in Twine now blogged at http://www.twine.com/item/110wgmfc4-35y/twine-and-linked-data-tom-heath-s-displacement-activities-by-tomheath. There are a few points there about consumption of external data; guess there'll be more discuss as time goes on; sounds like Radar are on top of it!
      Revyu
    • 21 months ago


      No SIOC right now unfortunately. We've considered it, and spoken to Uldis Bojars and the rest of the SIOC guys about this, but not implemented anything yet. Our initial issues were around redundancy (i.e. does using SIOC to markup reviews in addition to the Review Vocab and FOAF just create redundant data without really adding any value), although this view may change. There's also a chance that these issues will be resolved at the schema-level. Danny Ayers and I have an action outstanding on firming up the latest version of the Review Vocab, so that may lead to some kind of mappings. We'll see.

      On the subject of consuming external data, I'd love to talk more about this. Revyu currently consumes FOAF from registered users who have supplied the URI of their FOAF file, and we also consume external data from DBpedia, the RDF Book Mashup, and the Open Guides. There's more info in our paper from ISWC2007 at http://iswc2007.semanticweb.org/papers/889.pdf . These connections are a little hard-coded right now, but that will be changing shortly to be far more dynamic. I'd really like to see more external data consumed by Twine, and intelligently so - that's the feature that will really start to challenge my use of del.icio.us+Facebook+LinkedIn, but was what I missed most when I first signed up to use the app. I'll write a longer blog post on this and other impressions shortly. Do you have any particular types of data in mind at the top of your list for Twine consumption?
      Revyu
      • 21 months ago


        Actually your comments would be helpful within Twine itself. I will add you to another Twine for that shortly.

        Obviously FOAF and DBpedia are on the list. RDF Book and Open Guides are interesting suggestions.

        We also struggled with the question of ontology redundancy, versus mapping, or what else to do. We haven't implemented SIOC yet for the same reason. We already have an ontology covering many concepts. What is the solution you have settled on?
        Revyu
    • 21 months ago


      Hi Nova,

      Yeah, totally agreed! We have Reviews, People and Tags (plus the bundles in which they're created) all available as RDF in Revyu. Can the tags and people be consumed by Twine as-is? What else would we need to do to get the reveiws in there? We're planning to improve the handling of locations in Revyu, so would be cool to discuss how this might tie in with Twine's handling of places.

      Cheers, Tom.
      Revyu
    • 21 months ago


      Hi. Haven't forgotten this thread, just been busy with other things the last few days. Blog analysis on its way. Re an ontology to cover all the things that people may want to Revyu, we've always held off doing this as our intention was to be completely open world and therefore the modelling exercise/effort seemed futile. Instead the aim was to reuse classes from existing ontologies on the Web, deriving type information from a combination of user tags and background knowledge on the Web. In many cases this works fine, and will become more effective as more data is published in RDF and interlinked, and we add a little bit of reasoning.

      The biggest issue right now however is a lack of stable URIs for many common concepts, and this is a frustration. The work on class hierarchy within DBpedia based on YAGO looked promising, and I got fairly involved in this last summer, however the reliance on wordnet has introduced many issues/inconsistencies leading me to doubt that as the solution to our issue in Revyu. Semi-automatic selection/use of classes from the Web (and the associated mapping) will be feasible in the next few years I imagine, but for now I think we need a step-increase in publication of lightweight ontologies for common concepts that people frequently want to talk about, and the associated social processes to build communities around these. Richard Cyganiak and I have discussed the idea of a Vocab Hackfest in the autumn, perhaps around the time of ISWC2008. Is this something you'd be interested in supporting/getting involved in?
      Revyu
      • 20 months ago


        It would be potentially. We should definitely send a few of our people -- I'm looking into that. Actually we should really be speaking there about our platform...
        Revyu
    Add a Comment
Report This

Twine is about discovering, collecting and sharing the content that interests you. Learn More

Join Twine

Stats

First Posted By

First Comment By

Tags

Community Tags

Who's Interested In This?

Forgot your password?