The slow adoption of semantic technologies
- Posted on [April 8, 2008 ] @ 3:12 pm
- Write comment
One blog I enjoy reading regularly is AVC I enjoy how Fred intersperses his business and personal interests in his posts. A recent post from Fred that caught my attention on his site highlighted a Semantic web tutorial on ReadWriteWeb another one of my regular reads. Alex Iskold does a good job describing the basic technologies people are using for Semantic applications. As someone who works very closely with publishers, I’ve spent a significant amount of time encouraging companies to provide metadata with their content, so we can better index/rank their articles. I’ve recognized a number of roadblocks that have prevented widespread adoption of these semantic technologies (what many people have started to call web 3.0), including a lack of standards, a lack of adherence to proposed standards, and a lack of interest/use/understanding in general.
Semantic applications will benefit from a universal set of standards for identifying and recognizing semantic information, but the lack of a clearly defined standard has prevented ubiquitous adoption. Alex Iskold highlights a number of competing systems, including RDF and good old meta headers. Meta headers have been around forever but haven’t receveived as much use as they should have.
Another barrier is the fact that publishers don’t often notice the benefits and value-add from including semantic data in webapps because it hasn’t had an impact on their bottom line yet. The reason for this is because Google is generally good-enough for most types of queries, and those apps looking to be the next big thing by leveraging semantic data are going to find it very hard to differentiate their service and demonstrate their real value proposition if they aren’t very good at web-search and automatic identification/extraction of data (unless it’s a complementary service/feature).
Lastly, the apps on the web today don’t automatically generate/disseminate semantic data, so it’s hard to get widespread use. TechCrunch writes about this topic from time to time, but it has yet to implement semantics in their site (unless they are only showing them to webcrawlers). Sure, you might argue that blogs aren’t the best example (because RSS feeds provide some structured data), and you can call me a hypocrite because I don’t have that data in my blog either. However, I work very closely with a number of large commercial academic, scientific, and news publishers who are also slow and/or reluctant to provide detailed metadata about their articles. Some publishers even think that people should pay a fee to access their metadata! It will only be a matter of time before these publishers come around to find that is a close-minded approach which reduces their overall visibility. If you don’t make your content and metadata available online, soon you’ll be finding yourself left out of the conversation and playing catchup with everyone else.