Prior Art On Verizon Patents

greenbird sends in word that Techdirt has up information from Daniel Berninger documenting prior art in the Verizon patents being used to destroy Vonage. "...due to the fun way the patent system works, introducing that kind of prior art to the USPTO for it to review the validity of Verizon's VoIP patents will take quite a bit of time and effort — much longer than Vonage has to fight Verizon in court." From Berninger's note: "In particular, the claims in both patents were anticipated by open standards assembled by the VoIP Forum (H.323) in 1996 and published in January 1997 with the participation of members from Cisco Systems, Microsoft, IBM, Nortel, Intel, Motorola, Lucent, and VocalTec Communications, among others... The Eric Voit patent applications reflect, in particular, contributions made by VocalTec Communication to the VoIP Forum during 1996 and formally published at the same time as a separate document."

Comments

So, the deal with patents and prior art ...

[joe_n_bloe]

The thing that people don't get about prior art and patents is that prior art is narrowly interpreted. A published discussion, even if broad and comprehensive, about voice over IP, is not necessarily likely to count against the issuance of a VoIP patent later on. In fact, the prior art may establish that the subject is actually of commercial interest.

Saying that a patent describes "just another way" to do some obvious thing, and is therefore trivial, is missing the point. It's exactly that "just another way" that a patent is intended to protect. Patents covering radical ideas are the exception, not the norm. If you have some obvious thing you need done, but you can't find "just another way" to do it that someone else hasn't already patented, and you don't have a license covering one of those other ways, you are S.O.L.

The excuse...

[EmbeddedJanitor]

We're all human and even patent examiners make mistakes!

Well I don't buy it! I think the USPTO is broken in a few ways.

First off, the examiners are likely working to some sort of quota: Gotta process 10 claims a week or whatever. They can spend a reasonable amount of time investigating the application, or they can process it quickly. If they find some prior art, they send it back to the applicant who will send in a more supporting paperwork resulting in more work for the examiner to clear the patent application. The shortest route to clearing the patent is to just grant it. Come Friday and you're a few behind for the week, well they get slipped through double-quick.

Secondly Uncle Sam makes a bundle out of the USPTO. Each examiner can crank out a few grand's worth of work a day. Make it harder to get patents and less people will apply (and the processing costs would increase). It is easier to just make patents as easy to get as one of those credit card college degrees.

Thirdly, the USPTO is not held accountable to any quality measures. USPTO does not wear the costs of bad patents. Heads don't roll if patents get overturned. The lawyers love it. All the patent applications bring in money. Bad patents == more work. Nobody is motivated to improve patent quality.

Basically everything is stacked to delivering poor patents. I have a few patents, more than half of which I think are crap. I recently searched one of my patents and was suprised to see that other patents were granted for the same idea, even though the application quoted my patent. This really sucks. A patent is supposed to be property, but here the USPTO have clearly sold the same property many times over.

Is there a solution to this all? Perhaps. Firstly, patent quality needs to improve. That can only happen if the USPTO is help accountable. For example, if they grant a patent that is later overturned, then the USPTO could be held accountable for costs and losses incurred.

Re:The excuse...

[Nullav]

They claimed they work on a quota system and all they had to do was a few quick Google searches and then grant the patent.

Well I know where I want to work now.

Re:Fun way?

[schon]

Clearly, prior art is not one of them - even if "the concepts in those patents were clearly discussed and published by others prior to the patent being filed" Come on, while VoIP concepts may have been discussed, it's quite clear that nobody would have thought of implementing them on the internet.

It's more basic then this

[hcmtnbiker]

The article links to a wikia article on the subject, which provides a very nice summary of the arguments. My question is how is this stuff even patentable?

Patent 6,282,574 clearly states that no one except Verizon can legally translate an IP to a telephone number and vice versa. The rest of the patents are basically saying Verizon owns the only right to transmit other various phone communications over TCP/IP.

WTF? How can someone be awarded a patent for their idea for an application layer protocol that depends on something like TCP/IP to even opperate?

Re:It's more basic then this

[Anonymous Coward]

Patent #1 - 6,282,574 - Method, server and telecommunications system for name translation on a conditional basis and/or to a telephone number

Upheld - claim 27 of the '574 patent.
26. A method comprising:

receiving a name translation request at a server coupled to a public packet data network;

translating a name included in the request into a destination telephone number associated with a name included in the request; and

transmitting a reply containing both the destination telephone number and a packet data network address of a telephone gateway coupled between the public packet data network and a telephone network through the public packet data network to a calling device.

27. A method as in claim 26, wherein the address is an Internet Protocol address.

Claim 26 is spurious because the phone companies had been doing network-based address translation over public packet networks for many years. This is how 800 number lookups worked, for example - the 800 number is mapped to an area code+local exchange number suitable for routing over the PSTN. The classic telecommunications packet networks used by the phone companies are different than the TCP/IP data networks we're familiar with - the links are synchronous and connections have to be set up and physically reserved in advance. But they're still packet-based.

As for Claim 27, that seems to be a good description of the Domain Name Service which existed for decades. Now, Verizon's lawyers may argue that it means something other than that. Tough, it's ambiguous because the terse way it was drafted ("A method as in claim 26, wherein...") can be interpreted in any number of different ways. They shouldn't be allowed to impose the most convenient interpretation for themselves many years after the fact.