SOME TIPS FOR COMMENTATORS TO KEEP CALM WHEN READING PATENT APPLICATIONS: After perusing through some technical journals and magazines recently, it would seem that many of the commentators (usually the open-source types) enjoy the sport of perusing patent publications via the USPTO web site, spotting a few applications with really broad claims, hope that the patent belongs to a Microsoft, Amazon, Google, etc. - and then roll the application into a cylinder with which the commentator beats the readers over the head with.

The latest example is Google's patent application for embedding advertisements into syndicated RSS and Atom feeds ("Embedding advertisements in syndicated content" - US Pub 20050165615) . This technology is apparently related to Google's efforts (at least since April of this year) to provide contextual ads for feeds. Of course, Google isn't the only one interested in this - companies like Pheedo and Kanoodle have been doing it for a while as well.

The problem here is that, when you read the claims in Google's application, you either begin laughing, or crying, depending on your circumstances and/or disposition:

Claim 1. A method of generating information including targeted ads, the method comprising:

receiving information in a syndicated format;

receiving a first targeted ad from an automated ad server system;

combining the first targeted ad with the received information to generate a set of response information which is also in a syndicated format.

Ridiculous, right? (Never mind that "and" was not inserted before the last limitation as is required in proper claim formatting).

Well, not really.

Just about every practitioner knows that when you are entering prosecution, you claim the subject matter as broad as you possibly can. Think of it as a bargain for the sale of a good: when you make your initial offer for sale, you don't start out negotiating with a price you think may be "fair" - you ask for the highest possible price that someone would reasonably pay. In all likelihood, a potential purchaser would respond to your offer with a lowball counter-offer, you come back with a lower price, and so on, until a final purchase price can be agreed upon, if at all.

Same thing goes for patent applications (at least in an ideal world): Applicant claims the most basic features (i.e., the broadest) of a core technology, while the Examiner uses the prior art to force the applicant to add additional limitations into the claim language. Ultimately, (remember, I'm still in the ideal world here) a reasonable compromise is made, and the application issues as a patent with circumspect features added into the claims.

Of course I know that mistakes are made, examiners don't properly search or analyze the prior art, and bad patents are ultimately issued. But contrary to some of the commentators out there, I don't believe this to be happening as often as people like to think (not in an egregious way, anyways).

Which brings me to my tips:

Tip #1: Expect the application to have a first set of claims that are on the border of the "laugh test." It's not the applicants job to initially limit the claims - that's what the Examiner is charged with doing.

Tip #2: Just because they are asking for it, doesn't mean they will get it.

Tip #3: If you really want to be productive, don't just say "here's another patent on something we've been doing for years," and leave it at that - if it's that obvious, find the documentation that supports your assertion and post it on your site. You will be astounded at how popular you will quickly become.