AN OVERBROAD PATENT ON CUSTOMIZING A WEBSITE - This issued patent seeks to patent the idea of... building a user profile by watching user behavior on a website! 10 minutes of your time can help narrow US patent applications before they become patents. Follow @askpatents on Twitter to help.

Patent: U.S. Patent No. 6,330,592

Question: Have you seen anything published before Dec. 5, 1998 that discusses:

  1. Tracking visitor interests according to what content they access
  2. Categorizing dynamically-generated content according to categories of interests
  3. Caching dynamically-generated content and serving cached versions to subsequent users who exhibit those interests.

If so, please submit evidence of prior art as an answer to this question. We welcome multiple answers from the same individual.

Claim 1

Claim 1 is representative and claims "A method of customizing a web site, said method comprising":

  1. labeling content of the web site;

  2. when at least one visitor accesses the content of a web site, registering the labeled accessed content in a personalized data file;

  3. storing the data file for the at least one visitor;

  4. generating at least one pre-customized display for a first visitor;

  5. caching the at least one pre-customized displays on the server computer;

  6. displaying the at least one pre-customized display to the first visitor;

  7. analyzing the data file of a second visitor and associating the second visitor with the at least one pre-customized display, wherein analyzing is performed after generating; and

  8. displaying the at least one pre-customized display to the second visitor, wherein the at least one pre-customized display is not regenerated before displaying the at least one pre-customized display to the second visitor."

SPECIAL NOTE: "Pre-customized display" means pre-associated with some category or other means of tracking visitor interests. The rest is all fairly close to plain English.

Claim 11

Claim 11 is probably the broadest; and so the easiest to find prior art for. It claims "A computer readable memory that can direct a web site server computer to function in a specified manner, comprising":

  1. visitor files stored in said computer memory of said web site server computer;

  2. pre-customized web content items stored in said computer memory of said web site server computer; and

  3. executable instructions stored in said computer memory of said web site server computer, said executable instructions including

    a. instructions to access an existing visitor file for a visitor;

    b. instructions to review data in existing visitor file to determine visitor preferences; and

    c. instructions, based on said visitor preferences, to provide pre-customized files to visitor."

SPECIAL NOTE Term that need clarification:

  • visitor file: A file on the server that tracks visitor preferences and interests.
  • web content item: The spec has defined this very broadly, encompassing "Any item that can be accessed and viewed by a visitor, such as an entire web page, a component of a web page, an insertion into a web page or a graphic link. Web Content Items can be defined through a markup language, including, but not limited to, HTML."
  • pre-customized: means the same thing it did in Claim 1.

Ideally, prior art will embody each of the limitations (list items) for one of the claims above.

What is good prior art? Please see our FAQ.

Want to help? Please vote or comment on submissions below. We welcome you to post your own request for prior art on other questionable US Patent Applications.

  • This patent was originally granted to Vignette in 1998 for technology that was developed by CNET in 1995. Would any of CNET's work be considered prior art?
    – Roger
    Jul 23, 2013 at 15:43

6 Answers 6


"Improving Web Server Performance by Caching Dynamic Data" (with Jim Challenger). In Proceedings of the USENIX 1997 Symposium on Internet Technologies and Systems (USITS '97), Monterey, CA, December 1997.


HTTP State Management Mechanism by David M. Kristol of AT&T Bell Laboratories and Lou Montulli of Netscape Communications Corp from February 1997 http://www.ietf.org/rfc/rfc2109.txt

This memo explains the creation of cookies to track users and gives examples of usage including "a magazine browsing system, in which a user's previous reading affects which offerings are presented."

  • I don't see anything in there about caching dynamically created content? It really seems directed more to tracking state.
    – taenric
    Jul 23, 2013 at 21:30

Slashdot was doing all of this in June/September 1997.

1) you could post to a thread

2) comment threads are organized by the article. the article into groups

3) not sure of the cacheing strategy. there could have been caching in apache, in the apache CGI module, in CGI.pm or the web app code itself. Would have to go into the code to see that.


EDIT: Sorry, it looks like I go this answer in before the Request was complete

  • Pre-customized display: Pre-rendered web page. pages are generated by perl on demand. For Pt.8 the common way to cache these pre-rendered pages is to put a web-cacheing proxy in front of them. Ari Luotonen, Web Proxy Servers (Prentice Hall, 1997) ISBN 0-13-680612-0 Jul 22, 2013 at 23:26

The Squid open-source HTTP proxy seems to offer a similar functionality...

"Squid is a fully-featured HTTP/1.0 proxy which is almost (but not quite - we're getting there!) a fully-featured HTTP/1.1 proxy. Squid offers a rich access control, authorization and logging environment to develop web proxy and content serving applications. Squid offers a rich set of traffic optimization options, most of which are enabled by default for simpler installation and high performance.

Where did Squid come from? Squid is based on the Harvest Cache Daemon developed in the early 1990's." http://www.squid-cache.org/Intro/

The Harvest Cache A Heirarchical Internet Object Cache suited to HTTP http://grahammitchell.com/harvest/

...As a result, there has been an attempt to retrofit caches for internet information system servers like the popular CERN proxy-http cache. One recently developed cache which deserves further examination is the Harvest cache, which boasts improved performance of an order of magnitude over the CERN cache and even over popular http daemons like Netscape's Netsite and NCSA's 1.4 httpd.

The Harvest cache is a heirarchical "internet object" proxy-cache developed jointly by researchers at USC and the University of Colorado in Boulder in the latter part of 1995.


I know this is far newer than the patent but Cloudflare's Railgun does dynamic content caching. Maybe deeper researching of their topic or asking them could yield some of the inspiration for the system.


AOLserver allows caching of dynamic content...

This appears to be from 1996.



CGI::Cache - Perl extension to help cache output of time-intensive CGI scripts so that subsequent visits to such scripts will not cost as much time.

from 1998/06/18


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .