Hi Eric, thanks for the feedback on my web services ideas. I have to admit that when I made my first post it had been quite a while since I had first read (cough... browsed) the project docs at http://www.communitypatent.org/project_docs/. The process described here is fairly narrow in scope and the ideas I have described assume a much wider scope, where any patent application or granted patent (from any jurisdiction) could be commented on with the aim of not only trying to provide prior art to invalidate claims but for a range of other useful purposes. So getting back on topic for the p2patent spec, I would at least like to see the comments submitted to the p2patent project being available as a web service to other parties to access and "include" in their sites. With the aim of linking back to the p2patent site to attract more comments by "experts" who have come via those third party sites. I haven't given it too much thought to the technology choices at this point but I agree that using an existing framework would be preferred. The TOPAZ project looks interesting as does Kowari. I don't have that much experience with web services and have no vested interest in any particular technology so I would defer to others with more knowledge when it came to making such technology decisions. Having built a simple prototype commenting system already (in-house), my experience is that designing the user interface and deciding on the data storage formats is the hardest part. My prototype used a "repository server" from a content managament system called Daisy (http://cocoondev.org/daisy/) to store & retrieve comments (combination of XML "parts" and database fields). The user interface (comment input) was a simple Cocoon flowscript and my patent search webapp (Perl code) called a REST web service from the repository server to display the comments. The actual coding part was done in a day. Deciding on the UI aspects and what data fields to store to make the commenting actual useful... we're still working on that ;-) The reason for choosing Daisy, was that we were already using it for our web sites as a CMS and I knew how it worked. It is limiting in some respects and I wouldn't recommend it for the p2patent project. I like some of the features of Fedora (http://www.fedora.info/), which TOPAZ is built in top of, in particular the "Extensible metadata management" and "Expressive inter-object relationships" aspects. Another thought I had was to use the idea of a mashup (http:// en.wikipedia.org/wiki/Mashup_%28web_application_hybrid%29) to combine the data from a number of web services to create an AJAX-based web app, where most of the "mashing" is done in the browser. GWT is an obvious choice for this sort of thing. E.g. the "find a patent application" use case, could call a patent search web service (we are planning such a thing for our patent lens search service), which would then give the user a list of hits. When the user chooses a patent to view (browse a patent application), the browser then merges the patent data web service with a p2patent comment web service to produce a view with both data (patent text + existing user comments). etc. Cheers Nick On 31/10/2006, at 9:50 AM, Eric Hestenes wrote: > Hi Nick, > Thanks for the response to this inquiry and your suggestions. > > What I hear described is a web service approach that would handle > the storage of comments and metadata in a way that is modular and > logically separated from the application that uses the data. One > could imagine that our application is a subscriber to that service, > and that other applications such as 3rd party software could also > subscribe to the service. If well-designed it could be extensible > to other types of metadata, but our initial focus would be on > comments, claim rankings and prior art references. It does seem > that if we are capturing comments and prior art, one could open it > up so that others can subscribe to that information. > > It sounds powerful, but I am wondering whether this is something > that has to be built from scratch, or we can build off some > framework to get this functionality. My understanding is that > Kowari is designed for this. Other areas we would have to look at > are user identity management and how the ratings and rankings > mechanism is related to the metadata service. > > The Open Source As Prior Art has also been exploring how metadata > can be organized (see http://osapa.org/wiki/index.php/OSSTag). > > The PLOS TOPAZ effort may soon have some of these features also but > I am wondering if there are other frameworks that we could build from. > > Is there any code framework that covers this type of functionality? > > Is using Kowari an approach of interest for the purposes you describe? > > Sincerely, > > Eric Hestenes > Technical Lead > NYLS Community Patent Review project > > Nick dos Remedios <nick at cambia.org> wrote: > On 30/10/2006, at 3:56 PM, Eric Hestenes wrote: >> It would be interesting to hear if anyone in the developer >> community has ideas or thoughts on this subject. > > One idea I have, is for sites such as patent database providers (I > work at CAMBIA, which provides a free full text patent search - > http://www.patentlens.net/) to collect comments and peer reviews of > patents and then share/publish these "comments" with any other > interested parties (other patent informatics sites, etc) via a > centrally administered web services server (OSDL?). > > So, instead of relying on a single site to be the point of contact > for submission of patent reviews and prior art, etc, many sites, > with different user foci, could contribute to the effort. Bloggers > could incorporate the web service just as easily as patent database > providers. > > For this to work, the web services schema would have to be flexible > enough to allow comments from a range of user perspectives > (software patents, life science patents, engineering patents, etc), > while still enforcing a degree of strictness necessary for proper > prior art documentation. We have ideas about also gathering > information that falls outside the specs of the P2P project, such > as users sharing information on licensing of particular patents as > well as general comments about the technical issues relating to the > practice a given patented invention. > > It has the drawback adding complexity and thus slowing the process. > If too many players try to influence the schema/functionality then > it could degrade into a lowest common denominator spec that is > impractical to actually implement. > > It would be up to each site as to how to collect and display this > data. However I would expect there to be a collaboratively > developed prototype web client that could be used as a starting > point as well as providing a way for bug fixes to be shared. > > This would not be limited to patent database sites as it is now > fairly easy to access the EPO OPS web service (http:// > ops.espacenet.com/) to GET patent data for a given patent number > (their DocDB database has all jurisdictions: front page data, > family data and legal status) using a SOAP client. The above > mentioned web service could even provide this data itself! > > Another idea is for the comment collection and publishing to come > from a single source but for other sites to be able to provide > access to this service via a simple iframe. The iframe would > display existing comments for that patent (by passing a patent doc > ID) and would also allow the user to add comments to that patent. > There may be technical problems with this idea due to > authentication, etc. and its really only an idea off the top of my > head right now. > > Anyway, thats my Monday morning babble... I hope it makes sense. > > Nick > ---- > Nick dos Remedios, PhD > CAMBIA > <http://www.cambia.org/> >
