This actually happened a long time ago, June 1998, but what the hell... I haven't written it anywhere before.

I was at work, at Marimba at the time, and I was scheduled to interview this fellow for a software engineering position.  His name was Santosh Doss.  And I'm checking over his resume... and lookee here, he's got a patent.

I was reading a lot of patents around that time, and I was probably writing my ARP Patents and my Moog Patents articles then.  And also at that time, the IBM Patent Server was the tool to use, so I used that to track down his specific patent so I could look it over.

US Patent 5,659,164:
Method of and System for Apparatus for Two-Way Automatically Creating, Identifying, Routing and Storing Digitally Scanned Documents
Inventors: Edward Schmid, Meridith NH; Santosh Doss, Newton, MA
Filed: Oct. 15, 1996
Granted: August 19, 1997

It's a scheme for doing a batch operation, scanning a stack of paper documents on a document scanner with a feed mechanism, and distributing the resulting images to local files, remote files, email, and so forth. Each document can be one or more pages long.  The patented part is a bar code sticker that is placed in the corner of the first page of each document in the stack, and so after scanning each page, some software detects the presence of the bar code sticker, knows that this is the first page of a new document in the stack, and that it has finished with the previous document, and by reading the bar code it can identify the document and route the resulting scanned image bits to an appropriate place.

Okay, fine.  It sounds pretty obvious, tagging the beginning of each item in a stack has been around forever in some form or another, and bar code stickers have always been designed to be recognizable by computer.  So as such, this sounds like just another one of those obvious patents that should never have been granted.

But... notice anything funny?  Take a closer look at the scan from the US Patent and Trademark Office (USPTO):

The image of the scanned patent page shows (and who can't see this coming...) a bar code sticker in the upper right hand corner.

Apparently the patent office has been using this very scheme for years.  That would make sense, being that they clearly have a need to scan stacks of multi-page documents and deposit the resulting image bits into files.

During the interview I asked Santosh about this.  I wanted to make sure that I had correctly interpretted the patent-speak of the text of the patent.  Yep, that's exactly what the patent is for.  I asked him if he had ever heard of the patent being used by anybody.  He said that he hadn't.  I then asked if he ever checked out the image of the patent from the USPTO?  No, he hadn't.  I told him what I found and we had a chuckle over it.

So, not only is the patent office eager to patent the obvious, they're eager to patent stuff that they already use.  I mean, I can imagine walking into the patent office, pointing to a desk, and saying, "I'd like to patent that stapler there".

Now, I suppose it's possible that the bar code procedure could have been put into place after the patent was granted, and maybe Santosh could sue the patent office for the use of his invention that the patent office granted to him before using his invention.  Or maybe the patent office uses the bar code stickers for some other, unrelated, application.  It's hard to say.  But still...

Well, if it is prior art, then isn't the patent no good anymore? I'm not a patent attorney, and I haven't invented something yet (working on it...), so I wouldn't know.
I'll bet that the barcode is a standard inventory thing.

Hey Matt,

(Well, if I have to explain the joke... :-) )

That's right, if there is prior art then the patent has little or no value. But it's rare to find that the prior art was already in use at the patent office. (Or at least when the page was scanned.)

-- Don

Hi Matt,

Hold the phone. Stop the presses. Your analysis bespeaks the interplay of omnibus claiming and peripheral claiming. Over 100 years ago, the U.S. Patent and Trademark Office allowed, and even preferred, omnibus claiming. According to that scheme of protection, an invention was defined by "what was described in the patent document." A statement in the patent document would say, "I claim as my invention everything that is described in this document." The court system was literally blogged down with competing claims of invention. Thus, the U.S. Patent and Trademark Office required an examination system coupled with peripheral claiming .... to separate the wheat from the chaffe.

Peripheral claiming is a form of expression wherein an inventor describes the best mode to practice his invention in the text and figures of the patent document, and separately describes the legal aspects of his invention in a separate section called the "claims." (because historically, the inventor was "claiming this part to be his invention.") This standard of practice is called "peripheral claiming" to denote that the metes and bounds of the inventor's exclusive right to practice his invention are limited to words in the "claims" section of the patent document. (there are limited exceptions to this rule that are beyond the scope of this posting).

Turning now to Schmid et al., U.S Patent No. 5,659,164, the legal protection is set forth in claim 1, which states, inter alia, "...routing the digitized document information scanned therefrom in accordance with the recorded cover sheet identification-routing information directly to said user..." Thus, according to '164, the "digitized bar code information" MUST contain the routing information for a particular user. The USPTO issued "cover" of the '164 patent does not contain such routing information. Thus, the '164 patent is not "obvious" in view of technology practiced by the U.S. Patent and Trademark Office.

I speculate that Mr. Schmid was far-removed from patenting process, and did not fully understand peripheral claiming during your interview. It makes for a funny story though.

--Todd

Hey Todd, thanks for posting...

The digitized bar code on the scanned patent contains the patent number, a number uniquely identifying the client, or "user", of this document in the batch scanning operation.

Ie., the word "user" might mean a process or task, as well as a human. Right?

Or, and I admit this is pushing it a little, the patent number can be considered "routing information" for a particular user because that's the identifying address where the user wants the result of the scan placed.

Anyway, at the very worst case, this patent is ridiculously close to what the patent office appears to be doing.