Hello Everyone,

Anyone seen the updated Google Patents page, very sexy and a bit of a surprise that the Big G is still investing in the project with the seemingly perpetual beta tag.  I have mentioned to some of you that my organization finds Google Patents valuable as a ready source for patent drawings and claims.  The site is very reliable and I can copy and paste from it into our client's wiki and all of the impedded links for the drawings, claims and full-text remain active.  Why store collections of patents as PDFs when you can draw upon the Google servers at will?

Well, I found myself on the site during a routine patent review meeting this morning and imagine my surprise when the entire site was completely redone.  There are now large prominant buttons allowing for a PDF download and the ability to read the patent.  They have also changed the layout so that the claims and citations are beneath one another instead of side by side.  This makes the page a little less crowded and personally I enjoyed the enhanced readability of the content.

But what might have been my favorite new feature is how they have enhanced the search within this patent box.  Instead of hiding this box in the upper left hand corner of the page it is now right below the page imagine.  Previously when you typed something in this box the system took you to the first occurrance and showed a list of the other occurances along the right side of the window.  Now the user is provided with a list of the number of times the term occures with the page number and the term in context (a lot like KWIC - key word in context).  When you click on one of the results you are again taken to the page image with the term highlighted but they have added a navagation bar along the right side of the window that shows you where else the term appears within the length of the document.  This is a feature that is common to other patent viewers on the market but it is the first time I have seen it from Google.

I have not had an opportunity to look into previous issues with patent coverage or general searching so I can't report on those items.  I can say that I like the formatting and layout changes they have made and they make a convenient site one that is even more useful and easy to use.

I encourage interested individuals to have a look and post your impressions in the comments.




  1. Hi Tony,
    Thanks for pointing out the recent improvements to Google Patents.  They really add value to the product and I agree the search within the patent feature is great to have on Google Patents. Thanks again and I hope you have a wonderful July 4th weekend!

  2. Hello Again,

    Over the weekend and early this morning there were several posting to this thread on the Chemical Information Mailing list.  The commentators rightfully pointed out that perpetual betas are part of Google's business model, the system is not up to date and that there are strange typos in the system.  All of these items were previously pointed out by PIUG members when this service was first rolled out and the situation surrounding those items don't seem to have changed much with the facelift.  I responded to these comments and added a few more thoughts of my own in an additional posting which I thought would be of interest here as well:

    I think we can clearly establish that Google is not all that interested in creating documentation around this service.  All of the comments that are coming in are certainly my experience as well and the reason why I don't rely on Google Patents as my primary searching source.  I use it primarily for document review and for sharing patents with my clients.  The new features support that particular use and are a welcome addition in my opinion (not that anyone is disagreeing with that assessment by the way, I am just saying).

    With regards to Brian Gray's analysis on Google's business practices, I couldn't agree more but Gmail gets much more frequent updates than Google Patents did.  I was just a little surprised to see that they are continuing to invest in this service since very little had occurred (besides adding some US pre-grant applications and adding some additional US patent content) since they launched and attempts to reach Google with comments and suggestions have almost always fell on deaf ears.

    While we are talking about Google Patents I thought I would add my $0.02 about one thing I like and one thing I dislike about searching with it.

    On the not so great side - it was good that they added US pre-grant application to the database but instead of making them searchable by the standard publication number (which everyone else in the world does) they instead made them searchable by series and serial number (otherwise known as the application number).  This is really counter-intuitive and frankly a pain since other database system use years instead of series numbers when displaying application numbers to help avoid confusion between documents that have the same serial numbers but different series.  You can see where the uninitiated would hear the term application number and think, well sure, this must be how you search for applications but of course, you still want to use the handy publication number that the USPTO provides.

    On the this has some potential side - I often find relevant answers using the Google search algorithms that I don't find with a traditional Boolean search.  Boolean is my first love and will always (or likely always) be my first choice for searching but after I have done that I will usually take a quick look at Google (using a form of PageRank for searching I presume) and a Latent Semantic Analysis based service to see if there are some documents that I pickup based on the different approaches to searching that these represent.  Invariably I do find a few additional documents and it helps me develop my strategy to discover why these were missed by my Boolean query.  The end result is a more comprehensive result which is what we all strive for.


    1. PIUGers should note that the SIPO and KIPI databases use the application number. In KIPI, you can search on the unexamined, examined and registered publication numbers, but the results are listed by application number. In SIPO, you can search on the unexamined publication number, but not the registered patent number.

      The practice of searching on publication numbers is an artifact of the old U.S. patent system in which an application went unrevealed until the patent was granted.

      The unexamined publication is a biproduct of a legal obligation to publish after 18 months if the application has not already been published; it is not a legal step on way to a patent. In the case of a Korean B1 (many) or a Japanese B1 (few) registered patents, there is no unexamined publication because the applications are published as registered patents before 18 months.

      The application number is the only field that is guaranteed to be populated for all applications. In other words, there are application numbers with no corresponding unexamined publication number and there are applications with no corresponding registered patent numbers. But there are no unexamined publication numbers without a corresponding application number and no registered patents without a corresponding application number.

      From an information science viewpoint, the application number is the most rigorously sound number.

  3. All points well taken, but these are public databases. All documents contained within them will have a unique publication number; unpublished applications by definition remain inaccessible to the public and will not appear in these databases. I suppose you could argue that in the future the US system will be harmonized with the rest of the world, but I suspect at that point both numbers would be issued simultaneously and the USPTO AppFT server would continue to use publication numbers as the primary means of identifying applications. The legacy issues here are immense. This points to one of the weaknesses of Googles products, that they tend to make decisions based on technical merits without much concern for social and cultural issues. Google has a lot of good products but besides their search and gmail, how many others do you use or even know exist? Best example of this is their video service. It was available before Youtube really got going and is better than Youtube in many technical aspects. Yet the service went nowhere and they eventually paid some big $$$ to buy out the competition ( sounds like someone else we know cough*microsoft*cough ).

    I like their patent search engine, but find it useful only for searching legacy (pre 76) patents. Other public servers provide better search and Googles famous ranking system is weakened by the fact that many of the metrics available for website ranking don't apply to these documents. The obvious one that does ought to give users pause for thought. If you select a patent with their system the ranking of that patent will improve a bit. A quick look at Googles privacy policy shows that they collect information from users, that this infomation can be shared with other parts of the company and its affiliates, and can be used for advertizing and marketing. All national public servers are quite different in this regard, specifically rejecting the collection and dissemination of user information. I would be a little leary of using their servers for this reason for the same reasons that we all had concerning IBMs womplex ( now Delphion ).  Google has a _very_ active R&D department, holds many patents, and is expanding into as many markets as it can. I was surprised that a special privacy amendment wasn't made for this products, as many of their other offering have amended policies. I don't think you need a tinfoil chapeau to see some problems here...

    Keith Nagel
    KPN Consulting: Maker of IP-Discover patent search software