DISQUS

Community Page on DISQUS

what is this?

Jump to original thread »
Author

The Semantic Web’s biggest problem

Started by mathewi · 6 months ago

Paul Miller has a new column at ZDNet that’s all about the Semantic Web — or Web3.0, as some like to call it — and he’s got a post up about an interview he did with the Father of the Web, Sir Tim Berners-Lee, in which Sir Tim says that all of the ... Continue reading »

17 comments

  • "Let’s face it — the biggest problem with the Semantic Web is that it’s as boring as dry toast"

    Couldn't agree with you more on this Mathew, although I think you're being polite by comparing it to dry toast. No individual - outside of those who are active in the Semantic Web community - is interested in the plumbing. It's what the plumbing can enable and deliver that will be exciting. The 'magic', so to speak.

    I suspect that the magic isn't going to come from semantic search but will come from contextual browsing - understanding the users context and presenting them with (personalized) next steps to directly access relevant and meaningful information.
  • That's the kind of thing I'm hoping for, Fraser -- and I'm sure that someone, somewhere is building something magical.
  • I agree, although I would qualify that further as saying that the biggest problem is the lack of meaningful applications that leverage the awesome potential of structured data on the web.

    I've blogged a response here:
    http://blog.softwareabstractions.com/the_software_abstractions/
  • Mathew

    working to reach agreement on the underlying capabilities that enable innovation does tend to be as boring as dry toast. That's hardly a Semantic Web innovation. I doubt the world at large got terribly excited about the discussions over railway gauges that opened a continent to settlement, or the design decisions that resulted in cost-effective world-spanning clipper ships...

    I agree with both you and Fraser that the important piece is what we do next; it's the applications that people build on top of those underlying capabilities, now that they are in place.

    Some will be consumer plays like Twine, where any user may well be able to 'see', 'touch', and be excited by the Semantic Web.

    In most cases, though, Semantic Web technologies will be quietly implemented in the background... making Fraser's contextualisation better/smarter, exposing information from those big corporate data silos to other big corporate data silos, etc.

    Most of the work arising from the Semantic Web will be important. Most of it will make someone money. Most of it will deliver enhanced functionality or capabilities to the user of an application. Most of it won't be sexy or exciting to the end user, though. And surely that's only a problem if we think it's going to be otherwise?

    Paul - who has never owned (or even touched) a slide rule.
  • Thanks for the comment, Paul. You are quite right that most of the technical underpinnings of the things we take for granted (television, etc.) are quite boring to most -- that's why I'm hoping that things like Twine can pull people in and get them excited about the possibilities.
  • Definitely! :-)
  • Talking about the innards of the Semantic Web are boring (naturally). Looking at this whole thing through the lenses of Linked Data on the Web should provide a clearer insight into the virtues of being able to distill data from the information contained into Web Documents.

    For instance do you find the degree of serendipity in the following links boring?

    1. http://dbpedia.org/resource/Linked_Data
    2. http://dbpedia.org/resource/Semantic_Web
    3. http://dbpedia.org/resource/Web_2.0

    The real historic problem with this whole thing has be the lack of practical and comprehensible use cases. These shortcomings are vaporizing by the second :-)

    BTW - It would be nice if I could login using my OpenID when posting comments :-)
  • Kingsley, I agree about the OpenID thing -- as far as I know, Disqus is working on that.
  • Typo fixed version of my response:

    Talking about the innards of the Semantic Web are boring (naturally). Looking at this whole thing through the lenses of Linked Data on the Web should provide a clearer insight into the virtues of being able to distill data from the information contained in Web Documents.

    For instance do you find the degree of serendipity in the following links boring?

    1. http://dbpedia.org/resource/Linked_Data
    2. http://dbpedia.org/resource/Semantic_Web
    3. http://dbpedia.org/resource/Web_2.0

    The real historic problem with this whole thing has been the lack of practical and comprehensible use cases. These shortcomings are vaporizing by the second :-)

    You can also look at my Personal Profile Page (*go to the Linked Data Viewer tab*) to see practical usage of Linked Data to expose all of the data I've chosen to share with the public via my Linked Data Space:

    http://myopenlink.net/dataspace/person/kidehen
  • "Let’s face it — the biggest problem with the Semantic Web is that it’s as boring as dry toast..."

    Well, the only saving grace, then, is that it's not worse than your usual dry-as-toast posts. Perhaps you might do your own legwork for a change.
  • Thanks for stopping by, Steve. All finished torturing cats and
    stealing the kids' lunch money, are we?


    On Thu, Feb 28, 2008 at 8:54 AM, Disqus
  • Both of these posts, made me laugh. Mathew, great response.
  • I think when considering the problems of the semantic web one should ask oneself what an "irrelevant" search result is. When I get bad results for Google, for example, it is generally because they are misleading, or useless, not because they are syntactically or semantically different.

    Consider the NLP parse of the following two sentences:

    "If you suspect you have the Bill Gates virus you should visit 'www.symantec.com/bill_gates/' and download the patch immediately."

    vs

    "If you suspect you have the Bill Gates virus you should pull the power cable from your computer immediately."

    I know the difference, but my grandmother does not, for example. The underlying "semantic" meaning of a page will be the real hard problem of any semantic web.
  • Correct me if I'm wrong, but the semantic web depends upon the meta data needed to connect disparate bits of information. What you're saying is true - the infrastructure to share and store that meta data may well be ready to go - but where does the data itself come from?

    Unfortunately, the answer is "people" - people being notoriously inconsistent in naming, tagging, storing, and sharing their digital assets. FWIW, I don't think we'll have a truly semantic web app until there's good enough AI to understand digital assets and add consistent metadata.
  • Couldn't agree more about the importance of the work on the elements that enable interconnectedness by meaning. Also agree that the masses aren't that interested in the technology itself - as much they are in the experiences that products/tools, which leverage these technologies, will enable. I'd also written some about this here:
    http://www.secondintegral.com/axonomics/?p=33
  • You're right about the issue that no one knows what "semantic" means. Hell, my co-workers which has been woking professionally on the web far longer than me asks me all the time what I mean. It's very frustrating.

    Thank god there are interface developers which takes care of that part in a developer team. :)

Add New Comment

Returning? Login