ReadWriteWeb has an interesting overview of the types of semantic web applications. I will not try to explain the various types, as that is handled better in the article itself. However, I would like to look at some of the applications that were mentioned in the article. I will not be discussing the actual formats of the information, as those standards will eventually emerge instead of people trying to support several technologies at once.
One of the applications that uses a “bottom-up” approach is Semantify from Dapper.net. I did try to use Semantify on this site, but I had some issues. The interface for Semantify is fairly nice, but non-obvious. I consider this a very big problem as I am very technical, and many web publishers may have more issues than I did. Otherwise, you go through a few steps to analyze the data, and create fields that can be used for the semantic web. Nothing against the service, but if the semantic web is going to become mainstream, it has to be more automated and part of the infrastructure of the web. Some services are trying to automate the semantic discovery process. SemanticHacker is the home of the TextWise API and is one of the discovery services. I submitted some text from my previous posts and had mixed results. If the content is fairly specific, it does a decent job of finding the semantic categories. For my career oriented posts, it did not fair well. These tests were using the demo version on the home page which limits the input to 2000 characters. Another discovery service is the Open Calais API being offered by Reuters. I have not had the opportunity to use this yet, but it definitely looks interesting.
The other side of the equation is how do we use the semantic information? Semantic search tools would use the information populated by the various semantic authoring tools. Two search tools highlighted are Hakia and PowerSet. These are interesting tools, but to really win the search game it needs to be outrageously better than Google. I am not sure if that just means when searching for Paris, France you get various pages, maps, travel guides and images for Paris. However this feels more like a meta-engine for Google. The contextual data is critical and hopefully these tools will find a better way to have semantic information used in searches.
There are other things appearing for the semantic web as well. There are Wikipedia-like warehouses of information like Freebase and not quite databases like Twine. The issue is that these tools are looking for a problem to solve. This is also part of the problem with the semantic web. Regular people really do not know what it will be and whether they should care. Some academics and really smart people have some good ideas, but I still think we are far from a semantic web reality.
I will reiterate what I have said once before, if tools and services don’t converge on a problem and solution it will be dead on arrival. The tools are getting closer with the APIs, but we need services wrapped around the APIs, like a WordPress semantic auto-discovery plugin. I do believe that semantic search will be the killer app, but I think we are still a year away from that reality. So, we wait.