The combination of the new Google toolbar, the Toxic Sludge is Good for You book (EXCELLENT, by the way), and the Sourcewatch online encyclopedia inspired me to an idea. My kung-fu is not good enough yet to make it so on my own, but maybe there is somebody skilled enough in .xpi coding.

The idea is to make a browser toolbar for eg. Firefox (should be the easiest option, given the openness of the platform), that will allow the user to annotate the keywords on any page with the Sourcewatch (SW) links.

Could be especially useful for reading the online newspapers, quickly outing the authors, news suppliers, and front groups with links into the deceptive world of public relations.

Design considerations:

• Links on the pages to be annotated should not be changed. If the keyword is in a link, the added SW link should be accessible by right-click via menu, or by any other similarly easy mechanism. Alternative (and probably much better) method is adding a link icon after a detected keyword/keyphrase. The user should not be deprived of the option of following existing links, but options to follow alternate links should be added. The exact behavior of the link icons can be specified in an easy and configurable way, using stylesheets.
• The toolbar should keep its database of the keywords locally, to minimize load on the encyclopedia site. Client-side processing also protects the privacy of the users, as there are no centralized processing logs lawyers could attempt to subpoena. A synchronizing mechanism similar to rsync should be used for updates, to allow transferring only the differences between the local database and the on-site database. This will allow minimizing the load on the server while allowing rapid synchronization of the keyword links in the world where the PR front puppet groups may appear overnight. This could be further paired with Wiki-style community-driven database.
• There should be an option of a weak/fuzzy matching, and the keywords should be allowed to have synonyms, to cope better with real world where names can be spelled multiple ways and things can have more than one name.
• The approach with a link-icon after a detected keyword has one more advantage; a line of multiple icons can be used when there are multiple matches for a keyword/keyphrase or its part. A mouseover tooltip should appear over the icon, specifying the keyword, and if it was strong or weak match. The shape and/or color of the icon can furthermore indicate the database the keyword belongs to - agencies, names, anything else.
• The local database of the links should have approximately this structure:
  • definition of the database itself (name, where it should be refreshed from, time of last actualization, icon(s) to use for links, optionally wildcard URL for the queries (eg. http://www.sourcewatch.org/index.php?title=%s, where %s is replaced with the keyword-specific string)
  • list of synonymous keywords and keyphrases (different ways how to spell a name, fully spelled name or a name with first name as initial only, name with or without middle name, or acronym and spelled-out acronym, optionally clues for doing this automatically, what kind of fuzzy matching should be used)
  • link to the appropriate webpage (to save space may be just a string to substitute to a URL with a wildcard, specified at the definition of the keyword database)
  • type of icon to use for the link (again, as a reference to a list of icons defined at the definition of the database)
  • brief description for the mouseover tooltip for the icon
• For a dream-like project, the tool could have support for multiple independent keyword/link databases, to allow other organizations like PRwatch to offer the people their own knowledgebases.
• Optionally, for the future, consider the option of redesigning the system for fully decentralized data storage with massive redundancy and cryptographic authentication of the data, using a P2P-like approach, which will provide the system with lawyerproofness. But this is too much to ask for at this moment.

It shouldn't be a big problem for a moderately skilled programmer with Mozilla extensions experience. There is a potential problem with larger keyword lists and slower machines, as the processing may take way too long time then. In such case, there is still the possibility to write it in C and hook it right into the browser.

We have the right to know who is feeding us lies.

See also:

KnowMore.org Firefox extension