In Semantic thoughts, I talked a little bit about my Semantic Web epiphany. In this post, I want to discuss some additional thoughts as I get more familiar with concepts about linked data and the Semantic Web blogosphere in general (like I needed to find one more thing to get interested in).
In recent days, in the mainstream tech blogosphere and in media in general, I’ve been reading a whole bunch of posts, some hype, many others poorly written posts about what the semantic web is all about (in this post I am referring mostly to some of the more formal Semantic Web concepts like RDF). What really got me was Matthew Ingram’s post that the semantic web is boring, and therein lies my irritation. Yes, if one looks at the underpinnings of what makes up the Semantic Web, it looks somewhat hairy and academic (I was intimidated and unimpressed for the longest time). However, for some reason people’s expectations for the Semantic Web are either something glitzy or something life changing. Why is that? My gut feeling is that the semantic web is getting a lot of hype because some prominent blogs and even mainstream media have started talking about it, because Sir Tim Berners-Lee is Sir Tim Berners-Lee and the fact that he is involved gives it some degree of visibility, and because companies have started getting funding. But people forget one thing. The Semantic Web is essentially a backend framework for getting more information from the web of data. In a perfect world, consumers wouldn’t know or care that their application was powered by the Semantic Web, unless they were wondering why their new discovery service was so much more powerful. I don’t believe that the Semantic Web is going to be the panacea that some think, but I do agree with Nova Spivack that the next decade of the web is going to be all about how we make our data smarter. One driven as much by good data design, and leveraging metadata, as by brute force and cool algorithms for interrogating the current structure of the web (a la PageRank). The Semantic Web will only be one part of that smarter backend, which will include everything from new data sources (mobile, geolocation, etc), distributed computing, database architectures, etc. What the front end will look like, I don’t really know. What I do know is that the smarter backend will result in better applications, but just like most people don’t really know that Google Maps is powered by MapReduce or Yahoo Search is now powered by Hadoop, they don’t really know that the superior performance of their application in finding and presenting information is because the web is becoming a linked data web.
Yes, the fact remains that most people developing Semantic Web apps, e.g. BioDash, are not web designers, but more academic types, so their apps don’t look good, but that’s going to change. Just look at goPubmed. Perhaps that’s why so much rides on Twine. Twine is arguably, the first true, very visibly Semantic Web consumer app. Freebase, much as I love it, is a platform first and foremost and designed for the developer.
So where do we go? Personally, I don’t really care about better social network (although FOAF, etc do focus on connecting people), but I do care about “data finding
data“. In the life sciences, there is enough structured data that I believe a semantic life science web is inevitable. If we don’t go in that direction, the life science community is hurting itself. If we want to start making sense of all the complex relationships, best expressed as a graph, then RDF is the way to go. This will only happen when there is enough of a body of developers well versed in data models who work with those that know how to build usable websites that leverage new, smarter backend technology. That will take a few years, but I am confident it will happen out of sheer necessity.  One only has to look at a conference like CSHALS to see the kinds of problems being that bring together the Semantic Web and the life sciences.
I would like to give a shoutout to Talis here. If every Semantic Web evangelist could be Paul Miller, the world would be a better place. Talis is a platform developed by a company that uses it to build library systems. It also just happens to be a platform for building RESTful web services on top of a Semantic Web backend. It’s up to developers to grok it and use it for building the next generation of applications. Applications that really harness and leverage collective intelligence. That last part is why I believe that the Semantic Web is essential for the life sciences. We need to marry the collective intelligence both from the biological data, as well as a knowledgable community that really understands the data.
Semantic thoughts #2
In Semantic thoughts, I talked a little bit about my Semantic Web epiphany. In this post, I want to discuss some additional thoughts as I get more familiar with concepts about linked data and the Semantic Web blogosphere in general (like I needed to find one more thing to get interested in).
In recent days, in the mainstream tech blogosphere and in media in general, I’ve been reading a whole bunch of posts, some hype, many others poorly written posts about what the semantic web is all about (in this post I am referring mostly to some of the more formal Semantic Web concepts like RDF). What really got me was Matthew Ingram’s post that the semantic web is boring, and therein lies my irritation. Yes, if one looks at the underpinnings of what makes up the Semantic Web, it looks somewhat hairy and academic (I was intimidated and unimpressed for the longest time). However, for some reason people’s expectations for the Semantic Web are either something glitzy or something life changing. Why is that? My gut feeling is that the semantic web is getting a lot of hype because some prominent blogs and even mainstream media have started talking about it, because Sir Tim Berners-Lee is Sir Tim Berners-Lee and the fact that he is involved gives it some degree of visibility, and because companies have started getting funding. But people forget one thing. The Semantic Web is essentially a backend framework for getting more information from the web of data. In a perfect world, consumers wouldn’t know or care that their application was powered by the Semantic Web, unless they were wondering why their new discovery service was so much more powerful. I don’t believe that the Semantic Web is going to be the panacea that some think, but I do agree with Nova Spivack that the next decade of the web is going to be all about how we make our data smarter. One driven as much by good data design, and leveraging metadata, as by brute force and cool algorithms for interrogating the current structure of the web (a la PageRank). The Semantic Web will only be one part of that smarter backend, which will include everything from new data sources (mobile, geolocation, etc), distributed computing, database architectures, etc. What the front end will look like, I don’t really know. What I do know is that the smarter backend will result in better applications, but just like most people don’t really know that Google Maps is powered by MapReduce or Yahoo Search is now powered by Hadoop, they don’t really know that the superior performance of their application in finding and presenting information is because the web is becoming a linked data web.
Yes, the fact remains that most people developing Semantic Web apps, e.g. BioDash, are not web designers, but more academic types, so their apps don’t look good, but that’s going to change. Just look at goPubmed. Perhaps that’s why so much rides on Twine. Twine is arguably, the first true, very visibly Semantic Web consumer app. Freebase, much as I love it, is a platform first and foremost and designed for the developer.
So where do we go? Personally, I don’t really care about better social network (although FOAF, etc do focus on connecting people), but I do care about “data finding
data“. In the life sciences, there is enough structured data that I believe a semantic life science web is inevitable. If we don’t go in that direction, the life science community is hurting itself. If we want to start making sense of all the complex relationships, best expressed as a graph, then RDF is the way to go. This will only happen when there is enough of a body of developers well versed in data models who work with those that know how to build usable websites that leverage new, smarter backend technology. That will take a few years, but I am confident it will happen out of sheer necessity.  One only has to look at a conference like CSHALS to see the kinds of problems being that bring together the Semantic Web and the life sciences.
I would like to give a shoutout to Talis here. If every Semantic Web evangelist could be Paul Miller, the world would be a better place. Talis is a platform developed by a company that uses it to build library systems. It also just happens to be a platform for building RESTful web services on top of a Semantic Web backend. It’s up to developers to grok it and use it for building the next generation of applications. Applications that really harness and leverage collective intelligence. That last part is why I believe that the Semantic Web is essential for the life sciences. We need to marry the collective intelligence both from the biological data, as well as a knowledgable community that really understands the data.
Further reading
RealTech
Geospatial Semantic Web
Nova Spivack on the Meaning and Future of the Semantic Web
Nova Spivack on Collective Intelligence and Hyperdata
Web 3G
Paul Miller channeling Richard Waters
Technorati Tags: Semantic Web, Collective Intelligence