a production of The Federal Circuit Assessment Project | www.FedCir.org

A Coder's Guide to the Claim Construction Project


An Overview of the Process

The process of entering an coding opinions into the database proceeds as follows.

  1. First, a series of Lexis and Westlaw searches (in the Federal Circuit databases) are conducted using a variety of search strategies to caputure all the potential opinions meeting the criteria ("an observable claim construction analysis"). This step is accomplished by the project manager(s).
  2. The results of these searches are compiled, and duplicates are eliminated.
  3. Custom-built software extracts all relevant data from the search results and generates a CSV (comma-separated-value) file, which is then imported into the project's SQL database (at present, running on an Oracle product). At this point, the only data field which remains unpopulated is the "Methodological Form" field.
  4. Project staff members ('coders') individually code each opinion according to methodological form. Opinions which do not contain an observable claim construction analysis are eliminated from the database - they are coded "XX" by the coders, and the project manager(s) confirm this result and then delete the record.
  5. After at least two-thirds of the project staff has coded an opinion, the results are reviewed by the project manager(s). If the results are in satisfactory agreement, the opinion (with the 'methodological form' field populated by the most prevalent form chosen by the coders) is released into the public database. Note that, while inter-coder agreement is tracked, discrepancies are resolved prior to the record entering the public database.
  6. For data-integrity reasons, once an opinion is released into the public database, the form code cannot be changed.
  7. All analyses, etc. of the data are accomplished using the search and download tools from the public database.

Logistical Information for Coders

  1. All case-coding for this project is done via a web-based interface. Log in to the interface using the name and password you've been given at the following url: http://staff.claimconstruction.com/
  2. Upon successful login, you will be presented with the list of opinions in the database for which you have not yet entered a "Form" code.
  3. Each record has two active links to the opinion text: HTML/PDF (via FindLaw); and Lexis. Note that the FindLaw database, while free, is incomplete. The Lexis database is complete, but requires a login to the Lexis research system.
  4. Use the pull-down menu in the Form field to input your code for each opinion. To enter your results into the database, click the "Submit" buttons found at the top and bottom of the case list. Note that you needn't enter codes for the entire list -- any opinions not coded will be re-listed for you. Similiarly, you needn't code all opinions in one sitting: the system will always present you with a list of your uncoded opinions upon login.
  5. If you need to revise your codings, use the link at the top of the page. Note that you cannot revise codings for opinions that have been processed by the project manager(s) and relased into to the public database.

How to Code Opinions for the Claim Construction Project

For each opinion in the database, you categorize ("code") the opinion according to the following rules.

  1. There are six coding categories: two broad categories (Procedural, Holistic) and three "form" categories (strong, intermediate, weak). Thus the six categories are:
    - Procedural[strong] (Ps)
    - Procedural[intermediate] (Pi)
    - Procedural[weak] (Pw)

    - Holistic[strong] (Hs)
    - Holistic[intermediate] (Hi)
    - Holistic[weak] (Hw)
  2. Each coding category correcponds to specific characteristics of the claim construction analysis found in that opinion. (Note: for purposes of the claim construction project, we're only interested in the claim construction analysis; ignore other parts of the opinions.)
  3. All opinions that meet the screening criteria -- "an observable analysis of claim construction" -- must be categorized. If you find an opinion in the database that does not contain a claim construction analysis, choose code "XX". (These opinions will be deleted from the database.)
  4. When coding, the first choice is to determine the broad category in which the opinion falls (procedural or holistic):
    The Procedural Approach
    The procedural approach is characterized by adherence to a relatively strict rules-based hierarchy of interpretive sources, with a particular emphasis on the ordinary meaning of disputed patent claim language.

    The Holistic Approach
    The holistic approach is a far less-structured analysis, utilizing the array of possible interpretive information in a flexible, case-specific fashion. Most holistic opinions move directly into an analysis of the way that the text of the specification informs the meaning of the disputed terms.
  5. The next decision is to choose the strength of the methodological form. Generally, the more "pure" the analysis is in favor of one or another methodological approach, the 'stronger' it is. We generally reserve 'weak' classifications to those opinions where the analysis is so weak and obtuse as to be almost indecipherable. The following table associates the categories with their textual description as well as a few (clickable) example cases.

Form Description Examples (click for case)
Ps Expresses rigid process; Ordinary meaning controls, absent express definitions, etc. Johnson Worldwide Assocs. v. Zebco Corp., 175 F.3d 985(Fed. Cir. 1999)
Neomagic Corp. v. Trident Microsystems Inc., 87 F.3d 1062 (Fed. Cir. 2002)
CCS Fitness Inc. v. Brunswick Corp., 309 F.3d 1373 (Fed Cir. 2002)
Pi Framework of formal process established; Departure along some dimension: no talk of presumption, extensive discussion of specification. Insituform Techs. v. Cat Contr., 99 F.3d 1098 (Fed. Cir. 1996)
Micro Chem. Inc. v. Great Plains Chem. Co., 194 F.3d 1250 (Fed. Cir. 1999)
3M v. Chemque Inc., 303 F.3d 1294 (Fed. Cir. 2002)
Pw Difficult to discern the form of analysis. Some discussion/hint of process. Ethicon Endo-Surgery v. United States Surgical Corp., 93 F.3d 1572 (Fed. Cir. 1996)
Key Pharms. v. Hercon Lab. Corp., 161 F.3d 709 (Fed. Cir. 1998)
Brassica Protection Products LLC v. Sunrise Farms, 301 F.3d 1343 (Fed. Cir. 2002)
 
Hw Difficult to discern the form of analysis. Little or no discussion of process or ordinary meaning. Litton Sys. v. Honeywell Inc., 87 F.3d 1559 (Fed. Cir. 1996)
DSC Communs. Corp. v. Pulse Communs. Inc., 170 F.3d 1354 (Fed. Cir. 1999)
Viskase Corp. v. American Nat'l Can Co., 261 F.3d 1316 (Fed. Cir. 2001)
Hi Acknowledgement of process, but not used. Specification/prosecution history is clearly dispositive. Harris Corp. v. IXYS Corp., 114 F.3d 1149 (Fed. Cir. 1997)
Aqua-Aerobic Sys. v. Aerators Inc., 211 F.3d 1241 (Fed. Cir. 2000)
Smith &Nephew Inc. v. Ethicon Inc., 276 F.3d 1304 (Fed. Cir. 2002)
Hs No discussion of process. Immediate use of specification or prosecution history. O.I. Corp. v. Tekmar Co., 115 F.3d 1576 (Fed. Cir. 1997)
Wang Lab. Inc. v. America Online Inc., 197 F.3d 1377 (Fed. Cir. 1999).
Cultor Corp. v. A.E. Staley Manufacturing Co., 224 F.3d 1328 (Fed. Cir. 2000)

Hints and Tips for Fast and Efficient Case-Coding

  1. Code each opinion. Each opinion that fits the criteria is a different record in the database. Some cases will have multiple opinions (very rarely more than two): e.g., concurrences and dissents. Be sure to code all opinions in a case separately (if they meet the criteria).
  2. Beware the head fake. You will find that some opinions contain what might best be described as "boilerplate" language introducing the claim construction analysis, perhaps citing several cases, etc. At times, the actual methodological appraoch used to resolve the claim construction question will differ from the boilerplate introduction. (Most examples of this type will will describe the procedural method by way of introduction, but in fact move to a holistic methodological approach (by diving right into the specification, for example). What we code is the actual analysis used to resolve the question, not the boilerplate or introductory language. Don't fall for the head fake!