[tweetmeme source=”Intellogist” only_single=false]
Today we’re proud to bring you an extensive and fascinating interview with Marla Grossman, the Executive Director of the Coalition for Patent and Trademark Information Dissemination. We discuss a range of topics related to patent information and the US patent system that we know you’ll enjoy.
Marla Grossman joined American Continental Group in 2010 and brings a reputation for being one of the nation’s preeminent IP and technology lobbyists. Through her wide network of contacts in the Washington community and across the country, she has worked with a diverse portfolio of clients, including associations and Fortune 500® technology, health care, financial and traditional manufacturing companies. She brings ACG’s clients expertise in strategic public policy planning and representation before federal agencies and Congress.
Before joining ACG, Ms. Grossman was a partner at PCT Government relations, a consulting firm focused on IP and technology government relations matters. Prior to that, Ms. Grossman was a partner at the law firm of Verner, Liipfert, Bernhard, McPherson and Hand (now part of DLA Piper US LLP), where she headed the Internet Ventures Practice Group.
Ms. Grossman has lectured extensively on IP, technology and entertainment issues and has served as a guest lecturer at Harvard University’s Kennedy School of Government from 1995-2000. Ms. Grossman received her J.D., cum laude, from Harvard Law School, and a B.A., summa cum laude, from Yale University (Phi Beta Kappa). She is admitted to the Maryland and District of Columbia bars.
Why might the USPTO be headed down a dangerous path with regards to data quality? What’s the best path forward for patent office collaboration? Will the USPTO get the funding it needs? Read on to find out!
Chris Jagalla, Intellogist: Hi Marla, thanks for agreeing to come on the Intellogist Blog and share your patent industry knowledge and expertise with our readers! The reason your name caught my attention was that a colleague of mine, Kristin Whitman, had mentioned some of the fascinating things you had to say at this year’s PIUG Conference in Cincinnati with regards to the future of USPTO patent data.
Could you enlighten our readers about the background of how the USPTO handles patent data and why the way the USPTO handles patent application data, which gets very little attention, is actually important to the end-users of that data (patent holders, legal counsel, and searchers)?
Marla Grossman: There are about 35,000 registered patent agents in the US who file documents with the USPTO today. 90 percent of their filings are now electronic. Every document that comes in to the office, even if it is electronic, comes in as page images. All of that material requires analysis, capture and a lot of detail work. Currently, that analysis and detail work is done at the pre-grant or grant stage.
The way the USPTO handles patent application data is important because the value of the data is, in part, determined by how usable it is. If the data is accurate and consistent, then it is usable. If the data has errors or there is little or no consistency, then the data becomes untrustworthy and difficult to utilize.
An end-user can get by with inferior data with one application if that end-user has a good search database. The problem arises when there are thousands, hundreds of thousands or millions of bits of sloppy data. Then the system becomes untrustworthy and unusable.
Patent applicants often get patent information wrong on their application and in subsequent filings. In fact, applicants make millions of such errors each year – a fact that can be validated by analyzing the documents publicly available on the USPTO’s Public PAIR system. In part, this is because of the 35,000 registered patent agents who file patent applications; the vast majority files fewer than 10 applications a year. So filing patent applications is not the mainstay of their work, and, they are human, so mistakes get made. But most end-users don’t realize that so many mistakes are made because the USPTO corrects these mistakes during the data capture and examination processes. Imagine if these millions of errors were not corrected. When an end-user searched for something, that user would not know whether he or she was getting full and complete searches.
Now – one last word about sloppy data. I mentioned that the end-user can get by with inferior data with one application. But no patent owner wants their patent to be the sloppy one. Moreover, the Constitutional dictate of the grant of the patent presents a bargain amongst the inventor, the government and the public. The government grants the inventor the right of exclusivity in exchange for the disclosure of a novel invention. For that disclosure to be meaningful to the public, it needs to be accurate and usable. Thus, I would argue that the USPTO needs to disseminate high quality patent data to fulfill its end of the constitutionally mandated patent bargain.
I: There have been rumblings that the USPTO might try to switch to an all XML-based patent application system. Certainly on the surface, having documents delivered in easily-to-manipulate form such as XML instead of having to take the extra step of optical character recognition sounds enticing. From hearing what you and some others have said, it may not be that easy to make such a transition. Why is that? What advantages/disadvantages do different data processes hold with regards to the patent system specifically?
MG: There is a short term and long term component to the answer:
1) In the short term, moving to a new system means that the USPTO will need to convert all of the “in process” applications into the data format that’s required by that new system. These applications all exist in image form in the current system and need to be converted to eXtensible Markup Language (XML)-tagged text. There is a risk that in order to keep the costs of that conversion down, the agency might accept a lower quality conversion than is currently used for publicly disseminated patent data. This would impact quality of the published content for this set of approximately 1 million applications. The agency could also decide it doesn’t want to go to the expense of converting some data types like chemical structures to a searchable format. This would have a big impact on the scientists who rely on that information.
2) The longer term issue pertains to any move by the USPTO to require XML from patent applicants. The USPTO hopes that the agency will be able to get well-structured, accurate XML content from this wide range of filers and load the content directly into their new system, use it for examination, and ultimately publish from it. But, as I just discussed, even in today’s PDF-based e-filing system, these filers make millions of errors each year in coding documents, getting the wrong information into fields on forms and all sorts of other mistakes. The errors are currently corrected in various parts of the data conversion process, which the USPTO hopes to scale back substantially or eliminate completely. But the idea that this vast array of filers will do better with a much more complex filing standard than PDF stretches credulity to the breaking point. Rather, the likelihood is that many filers will struggle to submit clean, well-formed, compliant XML content, and there will be more – not fewer – errors entering the system. So if the USPTO imposes an XML filing requirement but skimps on the process to identify errors and correct them, the quality of the content feeding patent examination and publication would deteriorate materially.
It is important to note that I’m not arguing against using XML as a data standard for patent information, nor am I arguing against getting text from applicants. Rather, I’m saying the USPTO should strive to get as much clean text from applicants as possible in a small, defined set of formats that the applicants already use (word processing formats, PDF, spreadsheet, etc.) to reduce the need for optical character recognition. It is also essential that the office retain a data conversion step in the process at the USPTO using existing XML expertise to “funnel” these documents in a few standard formats through in order to produce consistent, reliable, well-structured XML content for examination and publication from this wide range of filers. Taking this approach would minimize total costs, it would provide the simplest solution for patent applicants by eliminating all the churn that will occur if applicants have to find and implement effective XML solutions, and it would maximize the quality of the content for all downstream users, including patent examiners and the users of published patent content.
I: Could you share with us your ideal data distribution structure in terms of applicants, patent offices, and third party entities?
MG: Patent examiners today work largely in an image-based environment using the USPTO’s Image File Wrapper (IFW) system. While this is far more efficient than a paper-based environment, IFW places significant constraints on examiners’ productivity that would not exist in a text-based environment. Thus, patent examiners need text in order to be maximally productive.
The USPTO recognizes this, and in its 2010 – 2015 Strategic Plan, Objective 6 toward reaching Goal 1 of optimizing patent quality and timeliness is for the office to develop and implement the patent end-to-end process system, including the development and implementation of XML for all data from application to publication.
However, merely providing text in XML format to examiners as it comes from applicants will not maximize examiner productivity because, as I’ve just discussed, there is an enormous number of applicant-generated errors and anomalies that flow into the examination process. This bogs down examiners’ work, making it much more difficult for examiners to produce a quality product and contributing materially to the current backlog of applications awaiting examination.
That being said, I think the ideal data distribution structure would be this:
First, the USPTO would get text from applicants. There would then be a harmonization step where errors would be corrected right after filing and high quality, fully compliant XML-tagging would be applied. This higher quality data would flow into patent examination process with automated analysis that would be updated in real time to update the examiner’s work.
The publication steps then become easy because the USPTO would have a data set that is accurate and ready for publication. The USPTO would publish in XML, as it does today.
The real change in the data distribution structure would be in the time at which the errors would be corrected and the XML tagging occurs; the USPTO would take steps to ensure the quality of inputs to the examination process right at the front end – both for new applications and for documents that enter the process across the full life cycle. You know the saying, “Garbage in, garbage out”? Well, in my ideal world, the USPTO would use such a process to get the data right at the front end and would apply adequate controls throughout the lifecycle so the data is properly maintained. The result would be “high quality data in, high quality data out”.
I: What’s the most important ingredient to providing a clean and reliable data system for patent documents?
MG: A crucial ingredient to providing a clean and reliable data system for patent documents – and something the USPTO may be missing in its current thinking about the new system – is the ability to bring consistency and accuracy to the diverse content that the USPTO receives from the wide range of patent agents who file documents with the agency. This ability requires domain expertise in the USPTO’s business rules, in XML, and a broad view of patent information. The agency applies such expertise today, but not until the publication stage. In its new system the agency needs to retain that step, but move it to the front end and funnel all documents through it. Failure to do this could bog down the new system very quickly and will almost certainly lead to a significant deterioration in the quality of published patent data.
I: Moving on to the issue of USPTO funding…it seems like this is one issue that everyone in the intellectual property industry agrees on—the USPTO should be allowed to set and keep its fees so that it may be sufficiently supported. Do you think that the provisions to hold the funds as designated by the House of Representatives sufficiently accomplishes the retainer of fees?
MG: First, let me note that not all stakeholders agree that the USPTO should be allowed to set its own fees. Significant stakeholders, such as the patent examiners union and the Intellectual Property Owners Association (IPO), believe that Congress should continue setting the fees. (*See footnote.)
Now, I absolutely agree that most stakeholders believe that the USPTO should be allowed to keep and utilize the user fees that it collects. I strongly believe, and I think most stakeholders believe, that anything less than that is a tax on innovation.
As to whether the House patent reform bill’s provisions sufficiently guarantee that the USPTO can retain its own user fees – the answer is “no”.
As Douglas Holtz-Eakin, former Director of the Congressional Budget Office, has recently commented, the House bill’s anti-fee diversion provisions would not guarantee that fees paid by UPSPTO users would be dedicated for use by USPTO. In the House bill, the USPTO would not be authorized to spend the fees it collects. Although the House bill sets up a Patent and Trademark Fee Reserve Fund at the Treasury Department, the Reserve Fund would not hold actual monies, but would simply hold IOUs. As with other “Reserve” and “Trust” funds in the Federal Government, the funds could be used for general government spending. USPTO would not have access to the “excess fees” deposited into the Fund apart from the annual Congressional appropriations process. Hence, despite the budgetary apparatus, there would be no effective difference from current law.
In comparison, the Senate bill creates a USPTO Public Enterprise Fund at the Treasury Department, which would receive all fees paid to USPTO and which would be available to USPTO when needed. The funds would not be funneled through the Congressional appropriations process and would, therefore, not be exposed to the threat of being diverted to general government spending.
I: There are a lot of things the USPTO could do with additional funding, including paying for overtime, hiring additional examiners, and improving facilities among other priorities. What do you think is the most important part of the equation to improve patent quality and reduce the patent backlog?
MG: I think the December 27, 2010 recommendations on patent quality submitted to the USPTO, the European Patent Office, and the Japan Patent Office — by IPO, the American Intellectual Property Law Association, BusinessEurope, and the Japan Intellectual Property Association – are very instructive. Their recommendations provided a number of nuts and bolts suggestions for applicants preparing a quality patent application, for patent offices and courts, and for members of the public submitting information to patent offices and assisting in training patent examiners.
One new USPTO program that those groups endorse, and I think might be the most important part of the equation to improve both quality and reducing the backlog, is the hiring of additional examiners with previous industry or law firm experience. Traditionally, the USPTO had hired mainly new engineering graduates. The USPTO would spend valuable time and resources training these graduates, and then when they finally got enough training to be productive, they would leave the agency. There are many reasons for the attrition, but in today’s day and age, few people stay in their first job out of school. I think that examiners with previous, relevant work experience are likely to produce quality work at the USPTO more quickly and are more likely to stay with the USPTO so the agency can reap the benefit of the time and resources it puts into training examiners. Retention of examiners is essential for quality and production.
Also, properly funding the IT infrastructure upgrades, especially but not limited to the PE2E [Patents End-to-End] process, will be essential. The benefits of any recruitment will be diluted if the examiners have to continue with the current slow and stand-alone series of workflow modules that they have at present. Investment in the IT infrastructure of PE2E is essential to the efficiency of the examination process and the quality of output at USPTO.
I: What do you think about work-sharing between different national/regional patent offices?
MG: The patent offices of many developing countries lack the ability to examine patents as effectively as the world’s major patent offices, like the USPTO, the European Patent Office, the Japan Patent Office, and the Korean Intellectual Property Office. This can result in the issuing of weak patents by the patent offices of these developing countries and the duplication of work by the larger offices of countries in which the patents are also filed. I believe that one way developing countries can combat this problem and reduce the burden of patent examinations is to rely on the assistance of more advanced patent offices to conduct their examinations, a form of “work sharing.” By reducing the examination expenses of the patent offices in developing countries, these offices can make better use of national patent fees to fund improved patent filing systems, general education campaigns, and policies supporting technology transfer from universities. This issue is ripe for study, and I hope that the USPTO assesses the possibilities for expanding work sharing for developing countries under the patent prosecution highway (PPH) and the effects such work sharing would likely have on the quality of the international patent system.
I: Does there need to be a more unified base of shared law to accomplish some of the efficiency that’s being sought?
MG: A unified base of shared law would likely increase patent office efficiencies. Chances are, it would lead to a reduction of examination costs, and the exchange of patent results for similar applications between offices could relieve many inefficiencies. With a unified base, an applicant seeking a patent in a foreign jurisdiction would be able to pursue the patent application process with greater knowledge and readiness. As a result, an examiner reviewing a patent from a foreign applicant would most likely come across fewer errors, and this might result in fewer “low quality” patents. Importantly, as Robert Willis argues in a Spring 2009 article about whether the US and foreign patent laws should be uniform, shared law would likely make worldwide patent protection more effective.
Whether a unified base of shared law in the near future is realistic or not is a different question. As a recent PPAC report on USPTO global efforts report points out, the global efforts of late have largely involved trilateral communication between Asia/Pacific, the US and Europe. These harmonization efforts have mostly been about identifying processes that have worked in one office and discussing why they would be beneficial to implement across all offices. For example, despite earlier resentment in Japan to the open ended grace period, Japan has recently used the US system as an influence to revise their patent laws to adopt a 6 month open-ended grace period; many in the patent community feel that they will soon adopt the 12 month open-ended grace period, which would align them with US law. However, some EU countries still reject the idea of the grace period. Many ideological differences toward the patent process still exist between the USPTO, EPO and WIPO, and developing countries have at large continued to refuse participating in many harmonization efforts
I: What do you think of the IP5 collaboration, especially with regards to creating a unified patent classification?
MG: Like harmonization efforts, any effort of patent offices around the world to collaborate through learning from each other’s best practices is a step in the right direction toward a more efficient global patent process. The current IP5 project, aimed at creating a unified patent classification system, is the combination of the best practices and traits from the US Classification system, the European Classification system, and the International Patent Classification system. These elements alone make it a seemingly beneficial collaboration and something that has the potential to make the patent process more efficient globally. The emergence of a partnership between the USPTO and the EPO to align their classification systems with the IPC, which is administered by WIPO, still leaves hope for the emergence of a unified base of shared laws in international patent community.
While the jointly developed classification system will be more detailed than the IPC to improve patent searching, and it will surely allow the two offices to start eliminating the unnecessary duplication of work between the two offices, it also is a system that is different from the current system used by the USPTO. While I don’t believe that the changes will result in the same difficulties that I suggested for the transition to an exclusively XML application format, this still means the stakeholder community and the USPTO will have to learn a whole new system for searching US patents and the USPTO patent database.
The USPTO has started this education process with presentations at the Patent Information User Group 2011 annual conference, and it will hopefully continue its outreach and education to other stakeholder groups.
I: Do you have any closing thoughts with regards to the future of patent data and/or informatics?
MG: It would be great if the USPTO would take steps in the future to ensure the quality of inputs to the examination process right at the front end – both for new applications and for documents that enter the process across the full life cycle. The biggest risk to the USPTO’s new end to end patent system is the wide range of inconsistent, incomplete and inaccurate filings by applicants. It’s a very big problem today and one that will probably become even greater if the USPTO expects applicants to file in XML. So ensuring the accuracy, consistency and completeness of the content provided to patent examiners by identifying and correcting the wide range of errors, missing information and other anomalies right at the beginning of the process would not only measurably improve examiner efficiency and effectiveness, it will also be absolutely crucial to the success of the USPTO’s new system. The technology and expertise is available today that would allow the errors to be identified in such a way that examiners wouldn’t have to search for them, thus reducing the effort expended in examination while at the same time improving accuracy. If the USPTO included a step like this immediately following the receipt of new documents, it would likely pay for itself many times over by positioning examiners to focus on the substantive issues of patent examination, and eliminating all the downstream costs and lost cycle time that the USPTO incurs when errors are allowed to persist through the patent prosecution process – not to mention helping to ensure the success of the new system.
I: Thanks so much for your time, Marla! More info about Marla Grossman can be found at The American Continental Group website.
*MG: Footnote: The reasons for some stakeholder’s opposition to the USPTO setting its own fees are several: First, unless user fee diversion is ended once and for all, if the agency can raise fees, there will be a heightened temptation for congressional appropriators and the Administration to divert funds because the USPTO will be able to raise fees the following year to cover the losses. Second, there is a concern that placing fee-setting authority with the USPTO will lead to declines in productivity and increases in fees in the longer term. Currently, Congress’s ability to set the fees serves as a “check and balance,” and if the agency can set the fees, that check and balance would be eliminated. Finally, there are important public policy considerations that are implicated with fee-setting – for instance, setting the application fee low serves the public policy objective of encouraging patent application filing. Many stakeholders feel comfortable with Congress overseeing that public policy role rather than the agency.
Thanks for reading!
This post was contributed by Intellogist Team member Chris Jagalla. The Intellogist blog is provided for free by Intellogist’s parent company Landon IP, a major provider of patent searches, trademark searches, technical translations, and information retrieval services.