Schneier on Security

is there information available about how many NSA patents get through?

just found 275 NSA patents since 1790 with this URL:
http://164.195.100.11/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2Fsearch-bool.html&r=0&f=S&l=50&TERM1=%22national+security+agency%22&FIELD1=&co1=AND&TERM2=&FIELD2=&d=pall
don't know whether the query catches them all though.

Off-topic:

Does the NSA just hire Math gods or what?

P.S. Thanks, Bruce, for your most informative and interesting blog.

This is not so informative in terms of evaluating the trends. If the NSA filed four times more patents this year than last, nine would mean a much smaller percentage are getting blocked.

Absolute numbers without context don't tell much.

I have to wonder if this type of thing will motivate people to to publish/sell their ideas before applying for patents if they fear that they may be censored by the government...

Too bad the good old Soviet Union is no more, which leaves those terrorists with bombs (but then again, everyone's a potential terrorist so there's still plenty of enemies to watch).

I find patents these days only protect patent lawers jobs. Not inventors.

"Does the NSA just hire Math gods or what?"

No, they hire a wide variety of professions. Mathematicians are popular, of course, but physicists, chemists, electrical and computer engineers, and even musicians (who can sometimes pick out background patterns more easily than others by virtue of their training) are common hires as well.

Let's not forget the linguists that work there, either (not that I used to work there as a Navy linguist many moons ago).

They also have their normal share of employees with no apparent talent other than finding a way to get from the start of their shift to the end without performing any discernible work, every day, day after day.

I mean, besides the linguists. ;)

My brother once had a patent application that was initially held up in this way. He envisioned a field of small explosives that would be ignited in such a way that their small shock waves would converge on a specific point, creating a large shock wave. Perfect for an insurgency that wants to knock aircraft down. Just needs radar, a computer, TNT, and some software.

They eventually let it through. I wonder if they figured it would not work?

-Stiennon

This item from the Wired News blog provides a different take on the story:

http://wiredblogs.tripod.com/elsewhere/index.blog?entry_id=1264879

Possibilities:

The NSA is doing more patentable research: the probability of any given patent being blocked has not increased, but more patents -> more blockage.

The NSA is opening up: the internal secrecy threshold has decreased, so they are trying to publish things they would not have published in years past.

The DOD is clamping down: the DOD is stopping the NSA from publishing things that they would have let slide in years past.

Hard to know which of these is the case without more info. Given the political climate, I suspect the third though.

If a patent application is blocked in such a manner as described, what happens if a third party (either within the US, or, since the USPTO accepts global submissions, from another country) submits a similar claim?

Can the USPTO reject an application based on a conflict with an unpublished patent? Or would the third party then receive royalties for a US Government-owned invention?

Seems to me it might be better to grant the patent but mark it Secret instead of Public. And, of course, any third party submissions could also be classified as "Secret" (which could then impose additional security measures on that third party).

Patents are for the rich.

For an individual or small business, suing someone for violating your rights can cost so much that you are financially ruined.

Alternatively, for a rich company, suing someone for supposedly violating the company's rights, can ruin the person being sued - simply becaus ethey cannot afford the cost of opposing the litigation.

One wonders if the NSA disguises the amount of breakthrough research that it does just by sending out a random number of patents every year, most of which range between somebody's shopping list and very obscure nonsense.
I mean how often do significant advancements in either cryptography or signal analysis really happen? Not often I suspect. The low-hanging fruit in this area has already been picked.