Many patent attorneys have long suspected that examiners during prosecution look only at the claims, and don't even bother with the specification. My experience of late is that examiners are now more regularly looking at the spec and handing out 112 rejections if there is a lack of support for claim elements, even on first office actions where no amendments have yet been made. (One recent action I received acknowledged that the original claims were technically part of the specification and thus not new matter, but nevertheless requested that the spec be amended to include the language of those claims.) But there is still an apocryphal story that I heard once again last weekend of a patent that had buried in its specification something like "Hey examiner, I bet you don't read this". I've never been able to verify that story (can you?), but I believe it.
When I heard it once again, though, some synapses fired to remind me of Raytheon's U.S. Patent No.
6,792,369 that I had come across a few years ago (thanks, bro). In this case, it wasn't the examiner who missed it, but the patent attorney. You're sitting there, reading the specification, when out of the blue, there is this conspicuous note to the inventor in all caps: "PETER, IS THIS CORRECT?" It's both funny and sad, and no doubt a source of embarrassment for the attorney and firm, who didn't take the few extra minutes to proofread their application before filing it.
And it made me wonder: just how many of these patents are there with stray notes to inventors? Well, thanks to Google patents, we have a clue. Googling just the term "Is this correct?" and sifting through the results yields, believe it or not, at least 23 issued patents, and 49 published applications with such notes. (By the way, it's tough to do these searches on the USPTO engine because of stopwords. Google's better on that, but has the drawback of not being current--it has a lag time of several months.) And that's just one particular phrase. There must be many others like, "Is this right?" or "Please explain" or "Do you really think this is distinguishable over the prior art, because it also has elements x, y and z."
Well, maybe not quite as bad as the last one. But not so far off. In addition to the general embarrassment, some of these leftovers clearly would be deposition fodder if the patents were litigated. Consider these, for example:
- U.S. Patent No. 6,331,957 (Intel Corp.) "<<Tsvika, is this correct, or why else is there no such register shown>>"
- U.S. Patent Pub. No. 2006/0212043 (Boston Scientific) "[Is this correct? What level of vacuum is required/preferred?]"
- U.S. Patent Pub. No. 2008/0052358 (Agility Management Partners) "/** Is this correct? Does it describe the way things were in the first version? If not, what is new and when was the change made? **/"
It's also noteworthy that these gaffes are (generally) found in the patents and applications of larger, more patent-experienced clients. That actually makes sense, since an attorney filing several applications at once for a client with a large portfolio may not give the same attention to each as he might to a smaller client. But it doesn't make it any more excusable.
What's worse is that the mistakes are rarely fixed: the issued patents with mistakes had those same mistakes in their respective published applications. So either nobody is proofreading the publications, or nobody wants to 'fess up that they missed it the first time. (There was at least one patent here, U.S. Patent No. 6,949,112 (Advanced Cardiovascular Systems, Inc.) that had a certificate of correction, but it seems to be the oddball.)
Worse still, in several of the patents, the mistakes persisted through multiple continuations and divisionals. For example, consider the following four issued patents to Nortel:
- U.S. Patent No. 6,650,803 Method and apparatus for optical to electrical to optical conversion in an optical cross-connect switch
- U.S. Patent No. 6,813,407 Method and apparatus for bridging optical signals in an optical network
- U.S. Patent No. 6,944,364 Method and apparatus for regenerating optical signals in an all-optical cross-connect switch
- U.S. Patent No. 6,947,623 Signals and methods for increasing reliability in optical network equipment
Each contain the following language in their specifications:
Full duplex is accomplished over a single optical fiber by transmitting and detecting signals in the single optical fiber at each end. [NOTE--IS THIS CORRECT TO SAY FOR FULL DUPLEX OVER A SINGLE FIBER. WE HAVE BEEN TRYING TO MOVE TOWARDS SAYING "TRANSPORT" SO WHEN AN OPTICAL RECEIVE AND TRANSMITTER ARE NOT PROVIDED. PLEASE COMMENT. WEA] The network 1304 also provides a bi-directional out-of-band signaling channel so that signals can be received and transmitted in each direction between the optical cross-connect switch and the attached network equipment 1402 and other network equipment coupled to the network 1304. [IN THIS CASE IT SHOULD BE OK TO SAY TRANSMIT AND RECEIVE BECAUSE IT'S THE SINGALING INTERFACE. CORRECT?]
One can't help but wonder how this could possibly be overlooked during a five year span covering at least three different publications and four issued patents. If the thought was that filing a certificate of correction would draw more attention, that logic died when the first divisional was filed.
As I said before, it's been common belief that examiners ignore the specification when reviewing an application. Why should we be surprised when we fail to look at the specifications when we are drafting them ourselves?
I'll spare some public shaming and hold off on listing all the applications and patents I've found. It's enough to know just from the examples above that the problem is more pervasive than any of us in the profession would like.
Recent Comments