USPTO Won't Accept Upside Down Faxes

bizwriter writes "This may seem like a joke, but it's not. The US Patent and Trademark Office will not accept patent filings faxed in if they arrive upside down. That's right, the home of innovation of the federal government is incapable of rotating an incoming fax file, whether electronically or on paper."

Comments

Idea

[Anonymous Coward]

Just send every single tax filing both ways. The right one gets filed, and wrong one gets rejected. Twice the work for the government.

Re:Idea

[ZachPruckowski]

And then finally they'll get the bright idea to implement software that recognizes whether it's upside down and only print out the ones that are right-side up!

They can't implement that software because a method for doing that has already been patented!

Much though I dislike software patents, that doesn't prevent using text to detect orientation. Someone upthread came up with a solution that wouldn't violate that patent, namely OCRing all orientations and the one with the most dictionary words is the correct orientation.

The posted patent compares letter width to letter height, and uses that to determine if the image is sideways. If the document is all capital letters or in Russian, it looks at the 'T's in the document, otherwise it uses 'i's. It then figures the ratio of what appear to be correctly oriented 'T's or 'i's to incorrectly oriented 'T's or 'i's and uses that to determine whether or not the document is upside down.

To circumvent that, you could test something different. If using different letters and the same overall formula don't evade the patent, you could still use factors like frequency analysis ('b' and 'd' are more common in English than 'q' and 'p') or attempting to detect different known incorrect characters (there's no English letter that looks like a sideways 'b', 'd', 'p', or 'q' or an upside-down 'k' or 'h' or 'y' (though an upside-down 'y' looks like a backwards 'h')

Well now

[zoomshorts]

It all depends on the definition of upside down. Back in the day it meant
the white side of the paper was oriented against the scanning device. All
pages were blank. Your mileage may vary.

Post ideas here.

[Kludge]

I work at a federal regulatory agency which is having the same issue. They were asking IT/tech/computer people if there was a solution around. Nobody knew of any software that auto rotates images based on text. Anybody? Reply here.

Re:Post ideas here.

[fuzzix]

I'm sure there's another way around, but gocr on the top or bottom section wouldn't provide enough data to "overrule" the header / footer, and doing the whole document would be pretty wasteful of computing time...

Well, I just did 2 gocr runs (with defaults) on a fax and its rotation, took about 4 seconds total on a VM sitting on a fairly over-subscribed box. The rotation itself took a negligible amount of time. Not implemented any automatic detection but what would be the overhead there?

There were about 5 dictionary words correctly recognised in the right-way-up version (with a lot of partial recognitions) and a lot of junk in the upside-down. I'm wondering now if there are easily recognised patterns in the junk without the overhead of naively running each "word" (even filtering out /[^a-zA-Z]/) through a cached subset of the language or performing some approx. string matching. Something like counting long, uninterrupted sequences of alpha chars perhaps?

It all falls apart on diagrams/handwritten contents... :)

Upside down or 180degrees?

[fiddlesticks]

Reading the FA, it could be that the faxer sent the fax the wrong way up/ down - so the office received a blank fax.

This would seem a perfectly valid reason to reject the submission

Re:Upside down could mean wrong sheet face up

[natehoy]

How would they differentiate that from just receiving a blank page (or a transmission error, or their own machine running out of toner or ink if it's a paper FAX machine).

Wouldn't the correct reply simply be "we got a blank page, so there's nothing to file, please resend"?

Re:I'm curious, who's the idiot?

[Skreems]

I came in here thinking the exact same thing. Looking at the comments on the page though, the author disputes that claim, although he doesn't provide any proof except that "the people involved said that's not what happened". Could just be covering his ass, or the speculation about using some software that intentionally does not allow modification of incoming documents could be correct -- I could imagine that being a legal requirement given the stuff they deal with.

Alteration?

[Known Nutter]

Could it be that the USPTO flipping the image constitutes altering the image?

Re:I'm curious, who's the idiot?

[clone53421]

Not only that, but the reason they say that it must have been rotated, not upside-down, is because otherwise the response would have been “your submission was blank”.

No, not if they’ve seen it a million times before, it wouldn’t...

First day on the job:

“Your submission was blank.”
“No, it wasn’t!”
“It was.”
“It wasn’t. I’m looking at it now!”
“Well, could you have possibly put the pages into the document feeder upside-down by mistake?”
“...”
“...”
“...oops. I’ll re-send it.”

2nd day on the job:

“The faxed submission was received upside down.”
“So rotate them 180 degrees, dumbass!”
“...”

Header orientation

[riverat1]

While I agree with most posters here that it's a silly rule I would point out that the fax header orientation in this case is opposite of the body orientation. If what is in the header is important to the USPTO (timestamp?) they may have a minor point.