copied from http://www.onebigweb.com/blog/index.php
Its been almost a year since my last post. Over this period, we have seen tons of action in the “web of data” space – the space which (as those four readers of this blog would know) constituted this blog’s staple diet.
As a precursor to some of the more detailed posts to follow over the next few weeks, I thought I could make a quick post on the apps I had written about earlier as well as some of the new entrants.
Ning : Offers a shared content store for social networking apps. It was the subject of my first post. The core issue that I saw with Ning was the stability of folkschema’s. Ning’s latest release fixes that by focussing on the standard social network data model which will provide the stable core from which derivatives can be created.
Googlebase : The only truly webscale semi-structured database. The concerns of openness have been addressed by a comprehensive read/write api. The issue of folkschema stability (again!) has been addressed by focussing on a small set of schema while retaining the generic capabilities.
Vast : Vast represents the perfect complement for Googlebase. While googlebase can only consume data, Vast extracts data from webpages. Vast clearly can play a usefull role while we transition from a web of data to a web of pages. But the future I suspect lies in the Googlebase approach.
Dapper : What Dapper could do for data is analogous to what youtube has done for video’s ie. force sites to open up their content and publish api so that they could regain control over their content. However, the future lies with those that offer commercial intermediation like Strikeiron/Mashery or those that offer aggregation and transformation capabilities like YahooPipes/OpenKapow.
Swivel : Allows users to upload small datasets which can then be visualized / compared with other datasets. But the tabular datamodel will make it tricky to mix data in rich ways. Further, what would have been nice is some high quality seed data.
Freebase : Offers a structured alternative to wikipedia. A semi-structured datamodel (rdf) along with a seed content from wikipedia provide the perfect foundation for growth. Exactly what swivel should do.
AdaptiveBlue : Identifies object type based on the text selected in a browser and allows users to perform relevant actions. One of the few that help to glue webservices together. Neat, but doest work consistently! Things should get better with more structure/ annotation in webpages!
Teqlo : Allows users to rapidly create simply mashup apps. In some ways, this represents a natural progression from being able to create simple dashboards using widgets. My first impressions were that it took too much effort to do very little. Need to take a closer look.
Radar Networks : Not yet released. Very secretive. Lots of datamining with initial focus on email and wiki’s. Think in terms of automatic tag creation for emails (person, topic etc). Very much looking forward to seeing this. A more detailed post to follow.