[time-nuts] (OT) Patents...

[Jim Lux]

On 10/1/11 3:13 PM, k6rtm at comcast.net wrote:
>> From a previous (now hijacked) thread--
>
>>> I wouldnt try to read patents. These are convoluted beasts, written by lawyers
>>> for lawyers. If you want to understand the technology, read the papers. Most
>>> of them are eithe freely available or for a small fee. If you have a good
>>> university nearby you can find them for sure in the physics library.
>>
>> Having read some of the patents I know that they can be quite
>> descriptive. There is a new tradition in patent-writing which you are
>> refering to, but that is rare with older patents.
>
> As a patent attorney, I resemble that remark!
>
> My undergrad work was physics and computer science. Yes there are plenty of obfuscatory patents out there.
>
> However... My job when I write a patent application is to describe the problem and the solution in as clear and concise a manner as possible.
>
> U.S. patent applications have two broad requirements, enablement and best-mode. Enablement means someone of average skill in the art should be able to read the patent and go build one without undue experimentation. Best-mode means you're telling people the best way you know how to solve the problem at the time you filed the patent. That's what I want to do -- describe the problem, and then teach someone how to build the solution, to build one that works, keeping things crisp, clear, and concise.
>
> As to going and reading the papers... Some times, yes -- but when I worked with Labs (and with Rick) I'd get disclosures from some folks that would start out with a 12 page derivation of PLL stability criteria -- Danny, we don't need that in the patent! Yes, we can cite Gardner's book, but we don't have to teach people how PLLs work when we're trying to patent a particular aspect of PLL operation! Or the disclosures that start out "Recall the retarded solution to Maxwell's Equations as shown in..." That's not going to make it into the patent application!
>
> Clear and concise, that's what I do.
>
> I'll be happy to carry on further discussion off-list, and return to trying to figure out if I really want to pick up a used Datum Tymserve 2100...
>

since it sort of fits with the idea of this list (figuring out how to do 
stuff with time)..

I think there is huge variety in patents, readability, and usefulness. 
As Bob points out there's the whole best mode disclosure part and some 
are wonderful and some not so useful.  There's also ones that are pretty 
good, but leave out all the key numbers you need.  Nice block diagram, 
description of function, but none of the details.

A patent attorney commented to me that there is a whole art to this 
(echoed by Bob) and a strategy.  It used to be that you had to be pretty 
good with your description, because the examiner wanted to really make 
sure that you had "reduced it to practice" (perhaps not actually built 
it, but had gotten close enough to know that it would work).  His 
experience was that applications with good descriptions and specific 
data (" in tests, a type 7474 dual-D flip flop, clocked with a 10 MHz 
clock, allows determination of the uncertainties to 124.2 nanoseconds" 
was a sure-fire attention grabber)

On the other hand, you don't want to give away the store on your trade 
secrets.  So you put in specific data, for a good embodiment of the 
invention, but maybe not the best you can possibly do.

There also used to be an art in titles of patents, back when searches 
were done manually.  An acquaintance of mine was enamored of patents 
with Catalyst in the title, because there are thousands and thousands of 
them to search through.

These days, though, there are fast electronic search engines.  AND, the 
patent office is overwhelmed and seems to have gotten into a mode of "if 
the form looks reasonably ok, we'll grant the patent, and let the first 
would-be infringer do the real testing"

There's also different styles depending on who's paying for the patent. 
  If you're a government lab, you're not really expecting to use the 
patent to litigate or as an asset to bargain with (both uses which favor 
some vagueness), so you might be a bit more straightforward in your 
disclosures: no reason not to.  On the other hand, if you're a 
government lab, you probably also don't have a big budget for drafting 
patents (either for the inventor to spend their time on it, or the 
patent attorney drafting and prosecuting the application).


At a guess, patents from before, say, 1970, tend to be MUCH easier to 
use than those cranked out in the last 10-20 years.

(I'll also point out that back in the days when applications were secret 
til the patent issued, and you could keep a patent app going for many 
years with CIPs and such, that was an effective negotiating tool, 
because the other side didn't know if what they were contemplating might 
all of a sudden be infringing your newly issued patent, after they had 
invested millions of dollars in it)