SparkIP, intellectual property exchange

If you have ever filed a patent, you will know that after writing it up in your engineering notebook one of the next steps is to do some preliminary research on prior art. The problem is that searching for prior art is not exactly the easiest thing in the world. The United States Patent and Trademark Office (USPTO) provides a very up to date database of patents and disclosures but the search functionality and presentation is mediocre at best. Google has also ventured into the patent search field and scientific publications with Google Patent Search and Google Scholar. Both tools are pretty good, for example, Google Patent Search, unlike the USPTO, allows one to view all the related images without the need of a plugin. Additionally, the full patent document can be downloaded as a PDF for future reference. On the downside, on occasion I have found it to be a bit out of sync with the USPTO and also not listing all the related citations. Further, I would also be concerned about whether Google stores the search terms and if they might get published.

Google though, is not alone in the intellectual property search and exchange game. Today, SparkIP launched SparkIP.com, which is billed as an intellectual property exchange for the scientific community. Ed Trimble, the Chief Executive Officer of SparkIP, will be familiar to many since prior to SparkIP he was the CEO and founder of EzGov Inc., before it was acquired by ChoicePoint. In essence, SparkIP intends to bring universities, innovators, corporations, attorneys, and researchers together and make the research on emerging technologies as well as the brokerage of IP easier. For now, the SparkIP database contains over 3.5 million US patents, with plans to add patent applications and international patents in the upcoming weeks. They are also seeking to partner with universities and labs from around the world and make their technology listings available on the site.

What is unique to SparkIP, as opposed to the other sites mentioned before, is their concept of SparkCluster maps. Basically, any search on the site will return a list of SparkClusters, each of which is a map of contextually related items which are arranged in clusters. Each cluster is represented as a node which is connected to other related nodes via edges. The size of the nodes represents the number of items that are contained in that cluster. Clicking on a node brings up another screen that shows filtered items based on the query. One also receives additional options such as that ability to filter by assignee name and date range.  The visual presentation is very appealing and the navigation is quite intuitive, except for the required double-click in the SparkCluster view. On the other hand, either the database might still be incomplete or the search functionality might need some improvements, since queries for recent patents that yielded results on other sites yielded none here. Overall though I think that this clustered approach and visualization has great potential in terms of IP search, and the dreaded search for prior art might become a lot less time consuming soon.