While I agree the patent is garbage, it's also covers a highly specific configuration. The operative part of the patent is not: "Methods and systems are provided for connecting an electronic device to a network."

It's the claims, and to infringe you have to meet every limitation of the claim:

> A computer-implemented method for testing network connectivity for a network device, comprising: connecting to a network provider; pinging a first server with the aid of the network provider by directing a data packet to a static interne protocol (IP) address of said first server; pinging a second server with the aid of the network provider by directing a data packet to a dedicated uniform resource locator (URL) of said second server; and determining whether to maintain connectivity to said network provider based on whether a response was received by said network device from said first server and/or whether a response was received by said network device from said second server; and connecting to another network provider based on at least one criterion selected from the group consisting of: bandwidth of the another network provider, cost to maintain connectivity to the another network provider, cost to transmit information with the aid of the another network provider, a download rate of the another network provider, and an upload rate of the another network provider.

I have actually worked with some code that took in account all those aspects. The code was written in 2005.

While not a lot video games are open source. A lot of games do things like. Particularly FPS shooters do very similar things to select servers or game hosts. Mainly in trying to find the lowest ping to ensure responsive game-play.

Let's break the claims down. * "computer-implemented method for testing network connectivity for a network device, comprising: connecting to a network provider;"

  This is any computer connected to a network that tests connectivity.

* "pinging a first server with the aid of the network provider by directing a data packet to a static interne protocol (IP) address of said first server;"

  This is any application testing ping time/rate.

* "pinging a second server with the aid of the network provider by directing a data packet to a dedicated uniform resource locator (URL) of said second server;"

  Any application that makes a HTTP request to check that a URL is online.

* "determining whether to maintain connectivity to said network provider based on whether a response was received by said network device from said first server and/or whether a response was received by said network device from said second server;"

  If the tests failed or are high latency try a different connection or probably choose a closer device.

* "and connecting to another network provider based on at least one criterion selected from the group consisting of: bandwidth of the another network provider, cost to maintain connectivity to the another network provider, cost to transmit information with the aid of the another network provider, a download rate of the another network provider, and an upload rate of the another network provider."

  Choose the best connection from a list of connections/devices based off your tests and any prior knowledge like geography.

The claims in the referenced patent are drafted to be extremely limited and inflexible. It would not apply to your product. You already get around it by considering geography.

Pretty sure adding an additional test not in the list doesn't exempt them from the patent. It's "at least one of" not "at least one of and nothing not in this list".

This happens because the USPTO has no incentive not to rubber-stamp everything that comes across their desks. For some time now, their position has been to accept claims at face value and let the courts (read: the taxpayers) deal with the fallout.

Except look at vegetish's post. It was rejected as obvious with a notice of final rejection. And they fought it tooth and nail to get it pushed through regardless of that rejection.

Hopefully that at least means it will be easy to overturn if they try to enforce the patent.

“Final” rejections are poorly named. All they really indicate is that you will probably have to submit a request for continued examination—-and pay a hefty fee—-to continue. Ironically, a lengthy and contentious application process can make the patent more difficult to challenge later, in a more sophisticated setting, because the presumption that the patent is valid is the strongest with prior art that the examiner “considered.” (IAALBNYL)

(Shrug) They had the option to do the right thing and maintain their final rejection. So why didn't they?

It's hard to escape the conclusion that they just couldn't be bothered.

Agreed. The fact they were allowed to resumbit such a broad patent with obvious prior art is ridiculous.

They amended the claims which required the examiner to do another search.

As mentioned above, Final Rejection just means you have to pay more money to continue the patent prosecution. This is called a Request for Continued Examination. RCEs are common.

Generally, the initial filing fees provide for two rounds of examination -- a non-final rejection and then a final rejection.

If the applicant re-filed the same claims that were rejected in the final rejection (w/o amending them), the examiner can issue an immediate final rejection on a subsequent RCE if he disagrees with the argument the applicant might have supplied.

The last amendment to the claims narrowed them to the extent that the patent is useless. The examiner knew this and allowed the patent.

The last amendment to the claims exceptionally narrows them. The examiner recognized this and issued a notice of allowance.

This patent is worthless and harmless.