The Third Man and Patent Searching

I recently had the good fortune to be in Vienna for the Information Retrieval Facility Symposium for 2010 (IRFS2010). PIUG members Stephen Adams and I are on the program committee for this meeting so we have been involved for more than a year in helping to put this meeting together and lining up speakers for the sessions.  The symposium went very well and many participants commented on how much valuable material was presented during the course of the three day event.

The IRF is an organization that I hope more PIUG members will support in the future since the goals of the organization represent items of importance to the patent retrieval community and PIUG members in particular.  For those of you who haven't heard of the IRF it is a non-profit research organization that looks to bring together the patent information and information retrieval communities to work together on the problem of information retrieval in the area of patent information. At the end of the day the group is trying to apply the advances made in search (think Google) to making the job of patent searching more efficient and comprehensive than it is today. This is a tall order, as all of you can imagine and the IRF has been at it for almost four years now.

As the organization was getting established they focused primarily on bringing the two communities together and helping them to understand one another and some of the nuances of what professional patent searchers do. Over the past year the group has sponsored a number of research projects which have focused on evaluating existing search systems and how well they can be applied to the corpus and problems associated with patent searching.  As you can imagine the initial results have not been particularly encouraging, the work we do after all is complicated and there are important reasons many of us use value added abstract and indexing databases but there is a need to establish a baseline with respect to existing tools before new ones can be designed.

As part of the evaluation exercise there was also a need to better understand how the patent information professional does their job and to gather some basic demographics on those of us who work in the field.  This was accomplished by way of a survey and the results were delivered on the first day of the symposium by Wim Vanderhauwhede of the University of Glascow. I am happy to report that PIUG members were one of the largest groups represented in the survey. The IRF has posted the presentation given (IRF Link) and I encourage all of you to have a look at the results. It is critical that PIUG members participate in exercises like this one so we can ensure that the people who work on our behalf understand us and know our needs. I would love to see us as a group engage in more activities like this in the future. There was also a posting today in the discussion forum from Marie-Pierre Garnier on the survey results (http://wiki.piug.org/display/PIUG/Survey+on+Patent+Search+Behavior).  From her post you can download the full report entitled Survey on Patent Users Search Behavior, Search Functionality and System Requirements.

As I mentioned, there were a lot of great presentations but in this post I would like to point out one other area of interest which is also an area where PIUG members could assist in the future. There was a section of the program on new interfaces which was subtitled Breaking the Cycle of Command Line Use.  As most of you are aware in many cases the primary interface used by patent searchers to query our databases of choice is the command line. Command line was an old technology 25 years ago when it was being phased out by the graphical user interfaces that came along in the mid-80's. It serves an important role for patent searchers but with all of the advances over the years the question really should be asked if there isn't something better that can be used instead. One alternative that is making it's way through the IRF and has been used for many years in the text mining, cheminformatics and bioinformatics communities is the use of visual workflows or pipelining methods.  Visual workflows have similar characteristics compared to the command line but the interface involves the use of visual elements that represent databases and transformations (commands) as opposed to what essentially amounts to writing code (which at the end of the day is what command line really represents where a search is really a small program to accomplish a data transformation) with a command line. The IRF is putting together a search interface that they call Leonardo which will be an attempt to bring the idea of visual workflows to the world of patent searching. It is an exciting idea and could truly change the way searchers conduct their searches in the future.

So where does PIUG come in? In the not too distant future the developers of Leonardo will be looking for searchers to provide feedback and help with testing of the new interface. I hope PIUG members will take the opportunity to step forward and provide guidance and perspective on how this new interface will help us do our day to day jobs.  We do important work and our perspectives are valued so the Board is looking for ways to make sure that we have opportunities to share our thoughts and knowledge.

Of course this will not be limited to surveys or software development.  We have been approached to help the patent offices of the world as well and I will talk about these opportunities in a future post.

Thanks for your time,
Tony