Thursday, December 27, 2007

On the reliability of Wikipedia for prior art in patent analyses

A 20 Dec 07 post on patently o has some comments concerning the use of wikipedia at the USPTO:

Maybe you can show that a page was created on a certain date, but it could have been modified yesterday. Thus, proof that a page was "published" on a certain day, is not proof that it was published in its anticipatory form on that day. It could have easily been modified into an anticipatory form. Wikipedia is an example of an evolving "publication."


CaveMan is missing part of the puzzle, however. Wiki's tend to keep track of changes, and archives the way a page looked at a particular point in time. At least does this. So if you wanted to see what the Coca-Cola wiki page looked like on Feb 10 of this year, go to:


I know that the examiners are encouraged to use the Internet Wayback Machine to determine the date of an online reference, but citing Wikipedia in an Office action is forbidden due to unreliability. Has Wikipedia been cited in an appeal or in litigation?


I must say that Wikipedia seems extremely reliable to me on most scientific issues, at least as reliable as other references that are allowed (offline and online encylopedias). If anything, the problem is that wiki is not detailed enough.

I recently encountered a PTO patent disclosing a homeopathic administration that is vastly more unreliable than the wikipedia article re same.

Bottom line: the ban on Wiki citations is bogus. [post by Malcolm Mooney]


What is interesting to me is that while examiners may be able to rely on anything that may prove anticipation or obviousness, e.g., a wayback machine entry dated before the critical date, the same evidence is - at this point - unlikely to be admissible in court due to hearsay issues.

AFAIK, the wayback machine has overcome hearsay objections in only one case (Telewizja Polska United States v. Echostar Satellite Corp., 2004 U.S. Dist. LEXIS 17876 (N.D. Ill., 2004)). Granted, I last looked into this about two years ago, but at the time there were a number of cases vehemently deriding the Internet in general as being completely unreliable to prove something was published on a given date due to the potential for hacks, modifications, etc.

The discrepancy between what an examiner can use to determine pre-grant patentability and what is admissible in court to invalidate an issued patent is what I find truly interesting. It makes a case for re-exam in attacking a patent's validity.

Let me just say that Wikipedia is one of the most biased and in many casess plainly falsified information sources whenever patents come up

Sooooooo many little nice folks like you Mooney out there typing on little keyboards...

You are sitting at the kiddies table tonight, Malcolm Mooney !!!

Business Week (9/4/2006 Issue 3999, p12) reported that the United States Patent & Trademark Office will no longer accept Wikipedia entries as "accepted sources of information"

An interesting article title "Kicking Wiki Out Of The Patent Office" (see, quotes Patents Commissioner John Doll stating "[t]he problem with Wikipedia is that it's constantly changing . . . " The article further quotes a USPTO spokesperson, who state "We've taken Wikipedia off our list of accepted sources of information."

It is important to note that this article is over a year old.

[End patently o comments]

IPBiz further notes the following from patentlibrarian.blogspot:

According to the Sept. 4 issue of Business Week, the USPTO has recently banned Wikipedia as an acceptable source of information for determining the patentability of inventions. Patent examiners and applicants have been citing Wikipedia articles in patents since at least November 2003... and not very many of them at that. A quick search of the USPTO's patent database reveals that from January 2004 to the present only 74 patents cite Wikipedia articles. Compare that to 33,802 references to IEEE (Institute of Electrical and Electronics Engineers) publications. Other science and engineering publishers heavily favored by patent examiners and applicants include:

Association of Computing Machinery = 6,010
Elsevier = 4,493
Springer Verlag = 2,319
McGraw-Hill = 2,728
Chemical Abstracts Service = 2,274
Oxford Univ Press = 781
Society of Automotive Engineers = 968
American Society of Mechanical Engineers = 850
Society of Petroleum Engineers =706
Cambridge Univ Press = 639
Wiley Interscience = 545
MIT Press = 403
ASTM = 498
American Institute of Chemical Engineers = 207

Wikipedia did beat its arch rival Encyclopedia Britannica, which was cited in only 47 patents in the same time period.

Of the comments of "Malcolm Mooney", IPBiz notes text from a previous post on IPBiz:

Notice that Wikipedia writes of Edison's U.S. patent 223,898: Edison continued to improve this design and by Jan 20, 1880 had U.S. patent 223,898 for a lamp that could last over 1200 hours using a carbonized bamboo filament.

IPBiz to Wikipedia: try reading US 223,898 and see if you find the word "bamboo."

None of the commentary on patently o came to grips with issues presented in Internet Publishing: Online Today, But What About Tomorrow Or Where Have You Gone, 406,302?

Through the wonders of third-party copying on the internet, this text can also be found at,_But_What_About_Tomorrow_Or_Where_Have_You_Gone,_406,302_517620x1146036883.htm

***In passing, IPBiz notes that the key patent of the Wright Brothers, US 821,393, was filed March 23, 1903, about NINE months BEFORE the flight on December 17, 1903.

****Of Edison and US 223,898, see also a picture of Lamp #4 from the Edison Collection, one of the Tar-putty lamps made in accordance with the description in Edison's fundamental lamp patent. Also: John W. Howell.


Post a Comment

<< Home