| Article Index |
|---|
| Collaborative Searching |
| Page 2 |
| All Pages |
Patent Search Collaboration & Tracking System
In 1999 while consulting for a company performing a lot of patent searching, it became obvious that their existing process was very inefficient. The problem with patent searching is that refining a search will retrieve a number of the same patents again. Without tracking the results you are likely to start re-reading patents until you realise that you have already read it (although after reading 100 patents you can never really be sure). Another problem is that a potentially important patent needs to be read quite thoroughly, which may have already been done by other team members. Information needed to be stored about searches done as well as their results, which was manually copied to word, excel or paper, and generally wasn't easily available for others to view. There was a whole range of improvements that could be made to enhance the process that weren't being considered at that time.
A web-based system was designed to improve the process. It was a relatively simple concept, consider a decorator pattern, but for the web. A HTTP proxy was written and would parse the retrieved search results page. It would then mark the patent links based on whether they were already seen and whether they were considered important or not. This then gave the user a view that the patent site had enhanced functionality and was tracking their searching.
The user could view the patents by clicking on the links, which would send a request to the patent site, through the proxy. The proxy would store that the user was viewing a particular patent. The patent returned from the request was parsed while the proxy was adding comment boxes and categorizing field comboboxes to the page. The user could then view the patent and add comments and categorize the patent as significant or not on the actual patent web page, even though the originating patent web site did not provide this functionality. The user then clicked the added save button on the patent page which then POSTed the data back towards the patent site, but when it was passed to the proxy, the proxy would store the results in a database and not forward the information to the site. The proxy made a basic patent site appear as though it had lots of useful functionality, including user registration.
One significantly useful feature of the system was that it allowed users within the same group to share search results and search strings. By setting up a topic to be searched, any group member could perform new searches and add metadata to various patents within that topic. This work could be continued by other group members at the same or later time. The system enabled group members to work as if they were one.
