The Productive Scholar: Patents and Patent Searching from a Librarian’s Perspective

CC license, Steve Snod­grass, via Flickr.

David Hol­lan­der, Law and Legal Stud­ies Librar­ian at Prince­ton, and Wil­low Dres­sel, Assis­tant Librar­ian at Princeton’s Engi­neer­ing and Furth Libraries, gave a talk on Wednes­day, Feb­ru­ary 9th about the his­tory of patents, the com­plex­ity of the law behind them, and how one can use this knowl­edge to per­form bet­ter patent searches to find patents, and how to con­duct his­tor­i­cal research on past patents and patent applications.

Hol­lan­der began the talk by remark­ing that the con­cept of pro­tect­ing inven­tions and ideas had been a part of Eng­lish law for hun­dreds of years before the exis­tence of the United States. The first idea of some­thing like a patent can be found in laws relat­ing to Eng­lish craftsmen’s guilds. In the U.S., an inventor’s right to an ideal was writ­ten into the orig­i­nal ver­sion of the Con­sti­tu­tion, Arti­cle 1, Sec­tion 8, Cause 8, by the inclu­sion of the phrase allow­ing con­gress “[t]o pro­mote the Progress of the Use­ful Arts by secur­ing for lim­ited Times to Inven­tors the exclu­sive Right to their  Dis­cov­er­ies.“

This lan­guage autho­rized Con­gress from the very begin­ning to grant and pro­tect patent rights, and the idea of patent pro­tec­tion was passed into law in the first Patent Act of 1790. Since then, these con­cepts have under­gone only three major revi­sions (with many smaller amend­ments and changes): patent acts were passed in 1793, 1836, and 1952, the last of which is the ver­sion that is still cur­rent today.

The U.S. Patent and Trade­mark Office was formed by Con­gress to han­dle the appli­ca­tion process for, and grant­ing of, patents. Chal­lenges to patents are han­dled by the U.S. court sys­tem. What are the basic steps involved in secur­ing a patent and what are the results of a suc­cess­ful application?

First, the inven­tor must prove that he or she has come up with a new, use­ful and non-obvious process or product.

Sec­ond, if the patent is granted, the inven­tor is enti­tled to a 20-year period of patent pro­tec­tion. Dur­ing that time the inven­tor has the exclu­sive right to, make, use, or sell the invention.

Hol­lan­der con­tin­ued to say that there are two main views of what a patent rep­re­sents. First, some view the patent process as a means to ensure fur­ther progress and inven­tion by virtue of the fact that patents are made pub­lic for oth­ers to study. By secur­ing the inventor’s rights, dis­clo­sure of a new inven­tion can fur­ther the state of the art in that par­tic­u­lar area. Some oth­ers regard the patent process as largely pro­tec­tive of pri­vate prop­erty, because an inven­tor has the right to keep any new inven­tion or process secret. The patent process medi­ates in this case to make dis­clo­sure pos­si­ble through the government’s guar­an­tee of 20 years of exclu­sive rights to the inven­tor. In both cases, dis­clo­sure is a promi­nent part of the patent process. Patents are by nature public.

The spe­cific terms “new,” use­ful, ““non-obvious,”“process and “prod­uct.” are crit­i­cal to under­stand­ing patent law. The con­cept of “process,” or “prod­uct” is a descrip­tion of the actual thing being patented. There are sev­eral types of patents includ­ing “util­ity,” “design,” “plan,” and other patent types, how­ever util­ity is by far the most com­mon. A process, a machine, a man­u­fac­ture, or a com­po­si­tion of mat­ter can con­sti­tute the basis for a patent that describes a new process or prod­uct. A use­ful improve­ment on any of these aspects of inven­tion can also war­rant grant­ing a new patent that builds on an older patent. Noth­ing beyond this list of four cat­e­gories can be patented. Ideas, for exam­ple are not patentable.

Computer-related patents raise an area of ambi­gu­ity in patent law, because com­put­ers rely upon math and algo­rithms to oper­ate. Math is regarded as a “law of nature” by the patent process, and until recent decades, was not patentable. Because of this, early inven­tions in com­put­ing were not patentable, although that opin­ion has relaxed some­what in recent years, as the lack of pro­tec­tion for inventor’s rights became an obvi­ous detri­ment to progress in com­puter technology.

How then is a new patent appli­ca­tion exam­ined for via­bil­ity? Accord­ing to Hol­lan­der, there are three main points of con­sid­er­a­tion for a suc­cess­ful patent application:

First is nov­elty. In order to inves­ti­gate this qual­ity, events that occurred before the patent that might have antic­i­pated its devel­op­ment are exam­ined, and the cur­rent state of the art related to the patent in the U.S. and other coun­tries is examined.

Sec­ond, the patent must show some degree of min­i­mal use­ful­ness. If a new inven­tion has no per­ceived use, it can­not be patented.

• The scope of prior art
• The dif­fer­ence between the new inven­tion and prior art
• The level of “ordi­nary skill” required to have come up with the process or prod­uct rep­re­sented by the patent.

• The new patent’s claims are exam­ined against prior art
• The orig­i­nal patent’s is com­pared to the newly patented invention’s or process’s claims
• The courts decide whether or not an infringe­ment has taken place

Mul­ti­ple sim­i­lar patents that are filed at the same time, a process called “an inter­fer­ence pro­ce­dure,” try­ing to deter­mine which appli­ca­tion has priority.

Hol­lan­der cited an exam­ple of a com­plex patent chal­lenge in a cur­rent case of Microsoft vs.. a small com­pany, i4i, which suc­cess­fully sued Microsoft for patent infringe­ment. The courts decided in favor of i4i, and granted a hefty set­tle­ment, but Microsoft has twice chal­lenged that deci­sion, argu­ing that Microsoft know­ingly infringed the i4i patent, but that patent was wrongly granted in the first place, and so, was invalid. (The tech­nol­ogy in ques­tion was the use of cus­tom XML, patented by i4i, and used by Microsoft in Word 2007). Here is a link to i4i’s sum­mary of the case so far. An account of the same case from The New York Times can be found here.) The case is expected to go to the Supreme Court later this year.

A ques­tion from the audi­ence about inter­na­tional patents prompted Hol­lan­der to explain that patents can be filed in dif­fer­ent coun­tries simul­ta­ne­ously, a labor-intensive process. Alter­nately one can file a patent in any coun­try that abides by the Paris Con­ven­tion of Indus­trial Prop­erty—which gives the appli­cant a one-year grace period for fil­ing in other countries.

The third option for inter­na­tional patents, Hol­lan­der explained, is to file under the Patent Coop­er­a­tion Treaty, to which 125 coun­tries have agreed, before fil­ing in any other coun­try. This also gives the inven­tor a one-year grace period for fil­ing indi­vid­ual patents, but patent appli­ca­tions under this treaty must sub­se­quently be made indi­vid­u­ally in other countries.

Hol­lan­der con­cluded by show­ing a patent for “pet dis­play cloth­ing,” a wear­able struc­ture that allows a pet owner to carry small pets like ger­bils in vis­i­ble tun­nels on his or her body. Despite the exten­sive, and some­what ridicu­lous draw­ings of the “pet dis­play gar­ment,” the only part of the patent appli­ca­tion that mat­tered in secur­ing this odd patent are the list of claims at the end of the application.

Wil­low Dres­sel next explained how to use var­i­ous web-based searches to find exist­ing patents for research and dis­cov­ery. For those search­ing for patents because they want to file their own patent, Dres­sel rec­om­mended engag­ing a patent pro­fes­sional for the most accu­rate and com­pre­hen­sive results. Pro­fes­sion­als are best qual­i­fied to do these sort of searches of prior art in patents, whether they are a patent attor­ney, or–in the case of patents aris­ing from work-related inven­tions at Princeton–the Intel­lec­tual Prop­erty and Licens­ing Office, part of the Office of Research and Project Admin­is­tra­tion at the Uni­ver­sity. How­ever, a desire to file a patent is not the only rea­son for doing patent searches, and Dres­sel explained sev­eral resources that aid in doing com­pre­hen­sive patent searches.

Dres­sel showed a LibGuide she has made con­tain­ing a list of links to resources to aid in patent searches for schol­arly and his­tor­i­cal rea­sons. For legal advice on patents relat­ing to research, she referred the audi­ence to the web­site of the Intel­lec­tual Prop­erty and Licens­ing Office, cited above, which can advise inven­tors on patent­ing inven­tions and processes that arise from their work at Princeton.

There are many rea­sons for search­ing patent lit­er­a­ture, for his­tor­i­cal research, where as she remarked “patents are a great gate­way into the lit­er­a­ture of a par­tic­u­lar field or tech­nol­ogy,” and can sim­ply help to under­stand how things work. Patents also con­tain a wealth of infor­ma­tion about research, as many cor­po­ra­tions do not pub­li­cize their research inter­est, but are obliged to file patents for new dis­cov­er­ies. Each patent con­tains ref­er­ences to prior patents that relate to that tech­nol­ogy. Patents also con­tain a lot more infor­ma­tion about a process or inven­tion than is gen­er­ally avail­able oth­er­wise. Par­tic­u­larly since the patent office has records of patents since 1790, the lit­er­a­ture can pro­vide an unpar­al­leled his­tor­i­cal view as well as pro­vid­ing a greater under­stand­ing of how cer­tain inven­tions and processes devel­oped over time.

Dres­sel empha­sized the impor­tance of defin­ing syn­onyms for search terms describ­ing patents, by think­ing of the words that describe an inven­tion or process, and con­sult­ing a the­saurus to find sim­i­lar terms in order to con­duct the most pro­duc­tive and com­plete searches. A huge key­word set will help in mak­ing the most fruit­ful searches. Search­ing patents, said Dres­sel, is an iter­a­tive process that can be added to as new searches pro­vide more ideas for searching.

Patents can also be searched in a spe­cific field by using the class num­bers defined by the U.S. Patent and Trade­mark Office. Cita­tion searches con­tain ref­er­ences to patents that refer to other patents, as each patent cites ref­er­ences to other rel­e­vant patents.

A Google search for “free patent search tools” revealed sev­eral sites of more or less value. The U.S. Patent Trade­mark Office data­base is the author­i­ta­tive guide, with the most cur­rent doc­u­ments, but full-text searches are only avail­able for patents since 1976. Patents from 1790 to 1975 can only be searched by date, clas­si­fi­ca­tion num­ber, or patent number.

Free Patents Online com­bines key­word searches for Euro­pean Patents, and other World Patents, and com­bines both appli­ca­tions for patents as well as patents in their data­base, two sources of data that would require sep­a­rate searches at the U.S. Patent Trade­mark Office. This resource, how­ever, also has ads.

Google Patents attempts to pro­vide full search­ing of patent doc­u­ments from 1790 to the present, but the searches depend upon many doc­u­ments, some hand­writ­ten, that were scanned, and which may miss some terms that were not rec­og­nized by the opti­cal char­ac­ter recog­ni­tion used by the software.

Dres­sel used Free Patents Online to search for touch­screen tech­nol­ogy in mobile devices, using terms “mobile,” “com­puter,” “input,” “phone,” as search terms. The search revealed that each patent is filed under a code that describes the class of tech­nol­ogy that the patent belongs to, and the clas­si­fi­ca­tions can be nested by hier­ar­chy into other clas­si­fi­ca­tions. As an exam­ple, Dres­sel searched for inven­tions that might relate to touch-screen input tech­nol­ogy for mobile devices. She demon­strated how one needs to think about a vari­ety of search terms that might relate to a touch screen, and dis­cov­ered a fairly recent patent that con­tained a large amount of infor­ma­tion that could lead in other direc­tions for searches in this area. The search result turned up patent titles and a excerpt of each abstract, listed in order of rel­e­vance. Look­ing at a patent for a “touch screen for a mobile tele­phone,” revealed a sum­mary of the patent appli­ca­tion, its pri­mary class, its inven­tor, and a PDF of the actual application.

Fol­low­ing the pri­mary class list­ing, Dres­sel vis­ited a link to the U.S. Patent Trade­mark Clas­si­fi­ca­tion list­ing for the patent she found, in order to visit the clas­si­fi­ca­tion home­page to find out more about that clas­si­fi­ca­tion num­ber. (She pointed out that U.S. clas­si­fi­ca­tion num­bers can be trans­lated into Euro­pean clas­si­fi­ca­tion num­bers using tools on the Patents Online site, which extends searches to other coun­tries.) Drilling down on the patent clas­si­fi­ca­tion for “visual dis­play sys­tems” revealed a num­ber of entries in the hier­ar­chy of that clas­si­fi­ca­tion, as well sub-classes and parent-classes indi­cat­ing their level within the clas­si­fi­ca­tion hier­ar­chy. Both patent appli­ca­tions and patents were returned in the search at Free Patents Online.

Dres­sel con­cluded by men­tion­ing a cat­a­strophic fire in the U.S. Patent Office in 1836, when all the patents from 1790 to 1836 were destroyed. Only a few–less than 3000–of the lost patents were even­tu­ally recov­ered from other sources, for exam­ple patent hold­ers who had a copy of a his­toric patent. Because of the loss of infor­ma­tion, all pre-fire patents were renum­bered, begin­ning with an “X,” to show that they are part of this frag­men­tary period of record. There are other data­bases at Prince­ton, Dres­sel explained, that help to cover the miss­ing data from pre-fire patents, scans of sup­port­ing doc­u­men­ta­tion for lost patents. A spe­cial tab on Dressel’s LibGuide labeled “1836 Patent Fire” pro­vides infor­ma­tion about how to find infor­ma­tion about patents prior to 1836.