Barbara McGlamery, a taxonomist at Martha Stewart Living Omnimedia gave the ending keynote for the LITA National Forum 2011.
In her engaging presentation she described the concept of the semantic web as web pages that can both be processed meaning fully by humans and by computers that then index and present the data from those pages in new and interesting ways. She argued that if web 1.0 was about making connections online, and web 2.0 was about online collaboration, web 3.0 will be about intelligence and making the web smart.
She then went on to describe the process she went through when she was working for Time on trying to implement the semantic web, in what she called the "Big S" method of making a website completely readable by a computer. In such a system a computer is given enough information that it can draw its own conclusions. If you have a dataset that has enough information about the relationship between things, a computer can make inferences. In such a system you can do some powerful things, however it is time intensive to implement, processor intensive to search, and laden with many other problems. Despite this, McGlamery described several attempts to use these methods at Time, that ultimately did not lead to a practical project.
After this, McGlamery described how she has instead been using "little s" semantic web principles in her work at Martha Stewart Living. This uses lighter-weight standards and is much easier to implement, but you wind up needing people to filter and adjust the conclusions a computer may come to based on its understanding of data, because of the limitations of this kind of approach. Despite that, this seems to be the way of the future, and I found myself quite interested in learning more about what is involved, particularly in implementing Google's microdata semantic standard in order to make our own website smarter without the immense work required with a heavier "Big S" solution.
In her engaging presentation she described the concept of the semantic web as web pages that can both be processed meaning fully by humans and by computers that then index and present the data from those pages in new and interesting ways. She argued that if web 1.0 was about making connections online, and web 2.0 was about online collaboration, web 3.0 will be about intelligence and making the web smart.
She then went on to describe the process she went through when she was working for Time on trying to implement the semantic web, in what she called the "Big S" method of making a website completely readable by a computer. In such a system a computer is given enough information that it can draw its own conclusions. If you have a dataset that has enough information about the relationship between things, a computer can make inferences. In such a system you can do some powerful things, however it is time intensive to implement, processor intensive to search, and laden with many other problems. Despite this, McGlamery described several attempts to use these methods at Time, that ultimately did not lead to a practical project.
After this, McGlamery described how she has instead been using "little s" semantic web principles in her work at Martha Stewart Living. This uses lighter-weight standards and is much easier to implement, but you wind up needing people to filter and adjust the conclusions a computer may come to based on its understanding of data, because of the limitations of this kind of approach. Despite that, this seems to be the way of the future, and I found myself quite interested in learning more about what is involved, particularly in implementing Google's microdata semantic standard in order to make our own website smarter without the immense work required with a heavier "Big S" solution.
No comments:
Post a Comment