USPTO Adopts PDF for Patent Documents

The USPTO has adopted PDF as the default format for displaying patent documents from its databases, making it one of the last (if not last) of the major patent offices to switch to the world's most popular document image format. Since 1998, viewing US patent documents obtained from the USPTO website required the use of a TIFF plugin such as alternaTIFF or interneTIFF.

USPTO Launches Global Patent Search Network

The USPTO has launched a new patent search system called the Global Patent Search Network (GPSN). Initial coverage includes Chinese published applications, patents, utility models from 2008 through 2011. Users can search patent documents in the English or Chinese language and retrieve full-text Chinese patents and machine translations.Additional Chinese patents and other national collections will be added over time.

PatentScope Mobile

PatentScope now offers a mobile search interface. As far as I know, this is the first patent office database optimized for mobile devices.

Also read that WIPO is planning to add "the bulk" of patent data from the IP5 patent offices, e.g. EPO, USPTO, JPO, SIPO and KIPO, to PatentScope in 2011. The EPO data was added a couple of months ago.

This will be a great addition to an already excellent tool.

2011 Patent Update @ Chemistry Librarians' Workshop

A couple of weeks ago I gave a brief update on patent search tool developments at the Chemistry Librarians' Workshop at the University of Toronto. One of the topics I covered was the recent addition of Markush searching in SciFinder. I'm still trying to figure out how to promote this powerful tool to faculty and grad students, most of whom are not too keen on reading patent documents.

FPO Adds Non-Patent Literature Collection

FreePatentsOnline (FPO) has added a full-text collection of non-patent literature to its suite of patent databases. Not much is stated about the collection, but according to the copyright notice that appears in each record it appears to be sourced from Gale Cengage Learning, a producer of full-text and bibliographic databases. The NPL collection can be searched alone or with the patent collections.

NPL records include bibliographic data, full-text and tables, but not images. Date coverage appears to start as far back as the early 1980s up through March 2011. In addition to full-text searching, it is possible to limit searches to specific fields such as author name (use IN), article title (use TTL) , and full text (use SPEC). Searching by patent classification is not possible as none of the articles appear to have USPC or IPC codes.

PatentScope Adds EPO Collection

The EPO patent document backfile, which includes bibliographic data from 1978 through mid-2010 and full-text data from 1996 through mid-2010, is now available in PatentScope. With this addition, the WIPO's patent search system has 7.7 million patent documents from 21 regional and national patent offices.

Earlier this year PatentScope added the national patent collections of Colombia, Costa Rica, Guatemala, Panama, Peru and Uruguay. Search interfaces in Korea, Russian and Spanish have also been deployed.

AusPat beta Launched

IP Australia recently released a new version of its patent database, AusPat beta. The new version allows users to retrieve patent specifications from 1904 to the present. Approximately half of the 1.4 million AU patent specs have been loaded. The complete collection is expected to be available in March 2011. Pre-1980 documents appear to be searchable only by patent number. Not sure if or when other bibliographic data will be added. See AusPat beta users guide for search tips and other.

How Complete is the USPTO Patent Database?

There was an interesting discussion last week on the Intellogist blog about the number of allegedly missing patent documents in the USPTO's PatFT database. Of course, this is an important question for anyone who uses the database, but especially for anyone who is doing legal or business research. (PatFT is by default the public patent database of record, although the USPTO does not make this claim.)

Determining the number of records that should be in the PatFT database is relatively easy. The USPTO assigns patent numbers in sequential order, as it has done since 1836. Let's take a closer look at utility patents issued from 1976 to the present. We know that the number of the first utility patent issued in 1976 is 3,930,271 and the highest patent number issued to date (as of Oct. 12, 2010) is 7,814,566. Subtract the latter from the former and add one and you get a total of 3,884,296. So the full-text collection in PatFT should contain 3,884,296 utility patent documents.

However, some of the numbers in the 3,930,271-7,814,566 range are unused because allowed applications (applications that are on the verge of being issued and have been assigned numbers) may be withdrawn from issue by the USPTO or the applicant. These numbers are withdrawn permanently and not reassigned to different applications. (The USPTO publishes lists of these withdrawn patent numbers each week in the Official Gazette.)

How many withdrawn patent numbers are there in our time frame? That's also easy to determine because the USPTO publishes an up-to-date list of withdrawn patent numbers. According to the list, there are 19,753 withdrawn patent numbers in the range 3,930,271-7,814,566. So we must subtract this number from the number above to get the total number of utility patents issued after Jan. 1, 1976 in the PatFT database.

"Potentially assigned patent numbers" - "withdrawn patent numbers" = "total issued patents"

3,884,296 - 19,753 = 3,864,543

We can check this number in PatFT by searching the "Application Type" field (APT) for patents coded "1" (utility patent applications).

apt/1 = 3,864,555

This search retrieves 3,864,555 hits, which is 12 *more* than the number we expected to see based on the calculation above. For a collection of almost 4 million documents, this is a very, very small discrepancy. I would expect similar results for other types of patent documents in the database, e.g. plants, designs, etc.

The reasonable conclusion is that there are no significant gaps in the USPTO's PatFT database, at least for the period after 1975. Of course, no database is perfect and there could be a few missing records in PatFT, but they are probably extremely rare.

Google Hosts Bulk USPTO Patent Data

The USPTO and Google have entered a two-year agreement to allow Google to provide bulk patent and trademark data to the public, according to a USPTO press release dated June 2.

The agreement requires Google to host the data at no charge and without modification. It appears that the data is being offered in weekly segments, at least for the last ten years or so. Older data (pre-1996) is offered in yearly files.

This is a great day for public access to patent information, especially for people who want to mine patent data or build large patent databases. I'm not sure how useful it will be for average users who might want to download a much smaller, targeted set of data.

Reporting Data Errors in Patent Databases

Kudos to esp@cenet for providing an easy and convenient way for users to report data errors. Each record includes a link to an error reporting form; all you have to do is click and fill in a few lines. This week I reported my first error (an incorrect application number and date) and within two days the record was updated.

Why don't more patent offices do this? Patent databases are teeming with errors. You'd think they would take advantage of the tens of thousands of users who use their databases every day.

The USPTO provides an e-mail for reporting data errors but it's located on a page outside the database, which means that you have to stop your search, open up a new window, find the page and then compose an e-mail. Very cumbersome.

I wasn't able to find any way to report an error in the WIPO's PatentScope. The CIPO's Canadian Patent Database includes the e-mail and telephone number for the Client Service Centre in each record, but it doesn't make it clear that you can or should report errors.

Bird Airbags?

Last year a large wind farm went up on an island near Kingston, Ontario, where I live. The installation consists of 86 turbines scattered around the island, dozens of which are visible from the city’s waterfront. The decision to build the farm was controversial. Although everyone likes the idea of clean renewable energy, many people feared the impact of the farm on the natural landscape, human health and wildlife habitat. Kingston sits astride a major North American migration route and some residents expressed concerns that untold numbers of migrating birds, as well as local bats and birds, would perish on the rotating blades. This has been a problem for wind farms all over the world.

One Californian inventor has come up with a uniquely novel solution: airbags for birds. In May 2008, John Silwa of Los Altos Hills filed an application for a "Method and Apparatus for Reducing Bird and Fish Injuries and Deaths at Wind and Water-Turbine Power-Generation Sites” (US 2008/0298962). The invention consists of airbags installed on turbine blades that would be activated by a motion-sensing or proximity sensing system. In theory, birds approaching the blades would trigger the airbags and bounce off unharmed or with “temporary stunning, blackout or non-critical injury.” It’s hard to believe that an airbag mounted on a spinning turbine weighing several tons would save a tiny bird.

Other inventors have taken more conventional approaches. Dr. William Hodos, professor emeritus of psychology at the University of Maryland, patented a system of painted patterns on turbine blades that make them more visible to birds (US 6,623,243). Melvin Kreithen of the University of Pittsburgh proposed using pulses of microwave energy to scare birds away from turbines and other structures (US 5,774,088). A published application filed by Keith Lima of Massachusetts discloses a system of ultrasonic or sonic wave generators that would deter birds from flying into wind turbines (US 2005/0162978). Canadian Angus Tocher has patented a “habitat friendly” wind energy extraction system based on small turbines encased in shrouds (CA 2,501,025 and US 7,220,096). German inventor H. Schlueter proposes mounting an ultrasonic sound producing device on rotor blades to warn bats (DE 102007025314).

The interdisciplinary nature of this problem makes it an interesting patent search challenge. The obvious starting point would be IPC class F03D, which covers wind motors and turbines. One possible search strategy is:

IPC = F03D AND Title/Abstract Keyword = (birds or bats or animals or fowl)

Other appropriate IPC codes include A01M29, which covers devices for scaring animals and birds, and H04B1 for electric signaling systems and G08B3 for audible signaling devices that could be used to frighten or warn animals.

Patent Database Review

There were a lot of developments in patent information last year. Below are some of the highlights from my favorite public patent databases and related websites.

Canadian Patents Database (CIPO)

The Canadian Patents Database, which is maintained by the Canadian Intellectual Property Office, contains more than two million Canadian patents and published applications from 1869 to the present. Full-text images are available from 1920 forward. Recent improvements include a few aesthetic changes to the search interface and the inclusion of a representative drawing (if available) displayed in the bibliographic record. In addition, as of January 29, 2010, abstracts in both English and French are available for applications filed under the PCT. (Approximately 75 percent of patent applications received by the CIPO are filed via the PCT system.)

PatentScope (WIPO)

PatentScope is the public database of record for PCT international patent applications published by the World Intellectual Property Organization. It contains approximately 1.7 million international applications published from 1978 forward. In 2009, WIPO extended PatentScope to include national patent collections from the African Regional Intellectual Property Organization, Cuba, Israel, Korea, Mexico, Singapore, South Africa and Vietnam. The largest of these are Korea (1.3 million documents from 1973-2007), Mexico (180,000 documents from 1991-2009) and Israel (144,000 documents from 1900-1999). WIPO also introduced a new search interface with simple, structured and browse functions, and the option to display search results as tables or graphs. The “classic” PatentScope search interface is still available for searching PCT international applications.

Esp@cenet (EPO)

Esp@cenet is a collection of free international and national patent databases hosted by the European Patent Office (EPO). The worldwide database contains approximately 60 million patent documents from more than 80 countries and over one million non-patent literature references.

Early in the year, the EPO extended esp@cenet’s coverage of Latin American countries, adding several thousand patent documents from Chile (2005-2008), Ecuador (2005-2006), Nicaragua (2006-2008) and Panama (1999-2006). During the summer, more than six million U.S. patent assignment records dating back to 1981 were reloaded. Legal status data was added or reloaded for Russian patent and utility models, Polish patents, and Chinese patents and utility models dating back to October 1985.

At the end of the year, the EPO announced a number of enhancements that were implemented in early 2010. These include highlighted search terms in titles, abstracts and full-text; full-text searching for EP and WO documents in all three official languages; and the ability to sort search results by date, inventor, applicant and ECLA code. Users can now enter more search terms in any one field (the previous limit was four).

In related news, the French National Institute of Industrial Property (INPI) replaced its old fee-based patent and trademark search systems with free web-based services. These include legal status information for granted French patents (B documents) and European patents designating France. (French patents are not available in the esp@cenet worldwide database but may be accessed via the French esp@cenet gateway.) Other services include databases of French designs and models from 1910 forward and international designs and models from 1979 forward.

USPTO

The USPTO hosts several public databases, including the two main patent databases, PatFT, containing issued patents from 1976 forward and AppFT, covering published applications from 2001 forward. Other databases include the Patent Application Information Retrieval (PAIR) system and the Patent Assignment Database, which contains recorded patent assignment information from 1980 to present. The USPTO website underwent a major reorganization in mid-2009 but there were no significant changes to the patent databases.

U.S. patent documents reached several notable milestones in 2009. In February, the USPTO published the two millionth application. The USPTO published the first application (A document) on March 15, 2001. Prior to that date, applications remained confidential until a patent issued. On March 3^rd, patent no. 7,500,000 was issued and on May 19^th plant patent no. 20,000 was granted.

President Obama’s new initiative to expand public access to government information and data could have a big impact on the USPTO. In September, the USPTO posted an RFI called the USPTO’s Data Dissemination Solution. The proposal seeks input from public or private sector parties interested in helping the USPTO make virtually all its public information freely accessible on the internet. The USPTO estimates that all of its data sets total about two petabytes. In exchange, the parties will be able to retain and use the data for their own purposes. It will be interesting to see if the USPTO can find any partners willing to accept these terms.

Last fall, David Kappos, the new Director of the USPTO, launched a blog called the Director’s Forum. In a recent post he expressed a desire to update the Manual of Patent Examining Procedure (MPEP), the USPTO’s handbook of patent rules and regulations. He described the current system used to produce the MPEP as an embarrassment. It takes too long to update and is cumbersome to produce. There are good document management systems that could solve these problems. Kappos also suggested that a reengineered MPEP could include wiki-style content contributed by patent professionals.

Patent Lens

Patent Lens is a free full-text patent database maintained by Cambia, an independent, non-profit research institute based in Australia. It contains approximately 10 million full-text patent documents published by the Australian patent office (1998+), USPTO (1976+), EPO (1978+) and WIPO (1978+). One of its unique features is the ability to search for gene sequences in patent documents using NCBI’s Blast software. Recent improvements to Patent Lens include a new patent search interface in Chinese and French and the ability to search PCT applications in the language of filing or publication, e.g. Chinese, English, French, German, Japanese, Korean, Russian and Spanish. Patent family trees now include a key explaining the family member colour-coding system.

FreePatentsOnline

FreePatentsOnline (FPO) is a free full-text patent database launched in 2005 by James Ryley. It covers full-text patent documents published by the EPO, USPTO and WIPO, and Japanese patent abstracts. Early in 2009, FPO added titles for patents cited on the front page of U.S. patent documents. (The actual printed patent lists only the number, date and inventor name.) Around mid-year, FPO launched a spin-off site called LocalPatents that maps U.S. patent data to geographic location. Patent clusters display on a map of the U.S. Users can zoom from the state level down to individual towns and cities and retrieve patents granted to residents of that location.

Boliven Patents

In January 2009, Boliven, a small start-up based in NYC, launched a website designed for professional entrepreneurs, inventors, researchers and patent attorneys. One of the resources offered was a patent database that covered U.S. patents from 1976 forward, EP documents from 1978 forward, PCT applications from 1989 forward and Japanese patent abstracts. Korean patents and INPADOC data were added later. The search interface included several innovative features, including a “Quick Flip” display option for rapidly viewing patent documents, faceted filtering and analytical tools for displaying search results as charts and graphs. Registered users could save searches, create search alerts and download data. Although the search engine performed well for keyword searches, it struggled with USPC and IPC classifications, producing results that were unreliable and inaccurate.

From January to April 2009, access to the patent database was free. On May 5, Boliven’s management announced that it would begin charging users $60 per month. Within a few weeks, however, it changed course and announced that the patent database, among other services, would remain free and that it would attempt to generate revenue from other services. Additional collections of public documents and records were added over the course of the year. However, the website’s future is uncertain. On January 15, 2010, Boliven’s management suddenly announced that it would cease operations as of January 22, citing its failure to meet operational and financial goals. As of February 23, the website and patent database are still online, although no new patent data appears to have been loaded since December.

PatSnap

PatSnap is a new fee-based patent search and analytics service launched in 2009 by a Singapore-based company. It covers full-text patent documents from the U.S. (1971+), Europe (1978+) and PCT (1978+) and Chinese patent abstracts from 1985 forward. A free trial was offered to members of ELD. Users can register for a free basic account that includes searching and viewing documents from the U.S., European and PCT collections. A day pass costs $39 and includes Chinese patent abstracts, basic analysis, 100 PDF downloads and 500 bibliographic data exports per day. The full plan costs $199 per month. PatSnap has a well-designed user interface, good search options and powerful analytic tools. Keyword searches performed well compared with other public patent databases, but classification searches did not produce the expected results.

Related News & Websites

Intellogist is a new website and online community for professional patent searchers. It is sponsored by Landon IP, a firm located in Alexandria, Virginia specializing in patent and trademark searches and patent analytics, and supported in part by advertisements. Both novice and experienced patent searchers will find it very useful. Resources include profiles of commercial and public patent databases, comparisons of patent search system capabilities, national patent coverage, best practices in prior art searching and a glossary. Registered users can contribute and revise content.

International Patent Classification (IPC)

There are major changes afoot for the IPC in 2010. The division between core and advanced levels, which was introduced in 2006, will be removed. Patent offices currently classifying documents using the core level will now use the main groups instead. New versions of the IPC will be published once a year in electronic format only. The integration of local classification systems (USPC, ECLA and JPO FI/F Term) will be accelerated.

IP5 Initiatives

The world’s five major patent offices (USPTO, EPO, JPO, SIPO and KIPO) known as the IP5 moved forward on series of joint projects announced at the end of 2008. The EPO is the lead office tasked with developing a common documentation database and common approach to patent classification. The USPTO is leading the development of a common approach to sharing and document search strategies and common search and examination support tools. The JPO will develop a common application format and access to search and examination results. KIPO is responsible for training policy and mutual machine translation. And the SIPO is working on a common set of rules for examination practice and quality control.

PCT Coverage in Public Patent Databases

How the data in a database or search engine is structured is as important (or more so) as the data itself.

I was reminded of this old rule-of-thumb a couple of weeks ago when I decided to compare coverage of PCT applications in public patent databases, namely PatentScope, FreePatentsOnline, esp@cenet, and Patent Lens. The results were fairly consistent until I got to Patent Lens, where my benchmark searches retrieved far more documents than the other three databases. (See table below.)

Most surprising was that my date of publication searches retrieved many more documents for five of the six dates I had selected. If PatentScope, which is the official record of the WIPO, says that 3,280 PCT applications were published on March 19, 2009, why did Patent Lens tell me it found 4,364? Obviously, this has serious implications for anyone using Patent Lens to do competitive intelligence, market research or simply track the number of PCTs filed by their organization.

Fortunately, the friendly folks at Patent Lens provided the explanation: Patent Lens indexes all versions of published PCTs, which inflates the number of retrieved documents. This includes subsequently published international search reports (A3 or A9), amended (A4) and corrected versions (A9). PatentScope and espacenet link these documents to the record for the initial publication (A1 or A2). FreePatentsOnline's PCT coverage apparently includes only the first published application (A1 or A2), although I haven't yet confirmed it. This explains the consistency in search results in FPO, PatentScope and esp@cenet.

So what are the practical implications for non-IP professionals who use Patent Lens? Well, if a researcher or tenure-track professor searches his or her name or university, they may get an inflated document count.

Espacenet Enhancements

This morning I discovered several nice enhancements in esp@cenet:

1. Expanded search terms
For years the maximum number of search terms you could input in any esp@cenet field was four. This limitation appears to have been lifted. I was able to successfully run searches using five or more terms.

2. Highlighted search terms
Search terms are now highlighted in retrieved titles and abstracts.

3. Sorting search results
Search results can be sorted by priority date, inventor, applicant, ECLA code and upload date.

The latest issue of the EPO's Patent Information News hints at more improvements to come later this year.

Boliven Withdraws from the Field

According to an e-mail notice sent to members last week, Boliven, a social networking service aimed at inventors, researchers and IP professionals will cease operations today. It was still online as of 5:00pm EST today. (Jan. 22)

The service was launched just a year ago. In addition to networking services, it offered access to millions of public documents, including patent documents from the U.S., Europe, Japan, Korea and the WIPO. The patent database had a nice interface and excellent tools for analyzing search results, but the search engine had significant problems understanding patent classification codes, which greatly limited its usefulness.

Boliven wasn't the first free patent database to offer analytical tools. WIPO's PatentScope database introduced them in 2006. Last year FreePatentsOnline launched a cool map mash-up called www.localpatents.com that maps U.S. patents by geographic location.

Although I really enjoyed using the analytical tools, Boliven's business model never made sense to me. After all, there are already plenty of free and commercial patent databases and networking sites. In May, after allowing users to test the site for free for a few months, it announced that it would start charging for access to its databases. A few weeks later it did an about-face and announced that members would continue to have free access.

Patent Info News Supplement / EPO Discontinues DVD Products

The EPO is now publishing a supplement with its Patent Information News newsletter that covers data and technical topics formerly covered in the INPADOC Patent News. e.g. Legal status codes, country coverage, etc. The first issue contains a nice overview of patent procedure in the U.S.

The latest issue of PIN also states that the EPO will cease production of the ESPACE WORLD DVD at the end of 2010; this product contained digital copies of PCT applications. Three other DVD products, ESPACE ACCESS, FIRST and ACCESS-EPC, are also slated to disapper. The data contained in them is now be available in a new online database called the Global Patent Index (GPI).

Thus marks another step in the long retreat from DVD as a media for disseminating patent info.

AusPat v1.3 Enhancements

IP Australia has announced the release of AusPat v1.3, the new web-based Australian patent database that replaced the PatSearch system in early 2008. AusPat v1.3 enhancements include (quoting from the release memo):

1. "e-Journal functionality has been incorporated into AusPat with the introduction of a new search field (called “Publication”) allowing users to search by Publication Action and Journal* Range.
2. The “Acceptance Published Date” search field has been removed. To search for applications by Acceptance Published Date use the new “Publication” search field.
3. Search and display of limited publication history for PATADMIN applications.
4. Ability to launch electronic Journal.*
5. Link to add AusPat to Instant Search Box for users with IE 7."

* The Australian Official Journal of Patents

Boliven Patent Goes Commercial

Boliven Patents is now a "professional-grade" fee-based service, according to an e-mail sent yesterday to current members of the Boliven Network. The service was launched in January as a free beta patent search engine with integrated analytical tools, search alerts and data export functions. It now includes more than 60 million patent documents from the US, Europe, Japan, Korea and others. Recently added data includes INPADOC/DOCDB data and US patent assignments.

Current members will have complimentary access for three months, after which they must pay $60 per month on a month-to-month basis. New users will be eligible for a free two-week trial period.

This isn't surprising news, given that Boliven is a private firm with a pretty obvious business plan. But I was hoping that the beta period would last longer or that part of the service would remain available to the public. This development underscores the danger of relying on third-party patent database providers to provide access to public patent information. They can disappear at any moment. Will FreePatentsOnline or Patent Lens be next?

New Website Compares Patent Search Systems

Landon IP, a private firm specializing in patent and trademark searches and patent analytics, has launched Intellogist, a free website that aims to help patent searchers locate sources of patent information, evaluate public and commercial patent search systems and exchange best practices in prior art searching. This service will be useful to both novice and experienced patent searchers. The site is supported in part by advertising.

New Patent Databse from Boliven

Boliven, a New York-based company, has launched a free patent database called Boliven Patents Beta.

The database contains data and images for US patents from 1976 to present, EP documents from 1978 forward, WO applications from 1989 forward and JP abstracts from 1976 forward. Search options include Basic, Advanced, Expert and Patent Number.

Search results can be filtered by source, assignee, date, and document type and sorted by relevance or date. A "QuickFlip" display option allows searchers to flip through displays of front pages very quickly.

Users who register for personal accounts (by invitation only) can take advantage of analytical tools, search histories, alerts and lists. For more information, see the press release.