Re: changes made/thanks (re hardware designs)

[Ian Carr-de Avelon <ian@emit.pl>]

Graham wrote: 
>Re the OpenNDA: one (possibly) positive point and a whole load
>of doubts [but insert 'IANAL' everywhere]. 
>1) Some people really react badly to the name (eg Andy); how about
>cancelling the double-negative and simply calling it a 'Disclosure 
>Agreement'?
A rose by any other name.

>2) The OpenNDA might limit peoples freedom to combine designs. 
>Whatever the final conclusions on how effective
>the gpl is for hardware designs, gpl-ed designs already exist.
When you pass on your own design, there is no problem as you can demand
that they sign the OpenNDA before they get the package. The problem
comes in that they have the right to pass on GPL parts to others who
have not signed the OpenNDA. I don't think this is a problem in 
practice, because there will be a clear border between the 2. Eg if
I provide a CD with Linux and a processor design, you can pass on the 
Linux to anyone GPL, the processor specific Linux for the new processor
is probably also GPL, but the processor design can only be passed on
to those who have signed OpenNDA. These files sit in a seperate directory,
so it is clear.
The problem only really comes if in some sense the design is too integrated
to make destinctions, and I doubt that will happen. If you sell mother
boards with the GPL SPARC, do you have to provide the PCB layout?

>3) If someone took your design and closed off a modified version,
>you couldn't sue them unless they'd signed the NDA - you'd have to 
>prove who gave the design to them, and sue that person.
This is where we don't get 100% of what we would like, but it is not
quite as bad as you make out. There is the possibility to get a court
to accept that simply "someone" has broken the NDA and this is still
a trade secret. How often that will happen depends on country, but it
is there. Do you see any way of getting closer to that 100%? Don't 
forget that the OpenNDA(s) work on the basis of 1) defining the rights
and then 2) licensing them. The intention is that part 1) can be common
with propriatory NDAs, so there is big money available to support the
idea that the wording is sound and not easily avoided.

> So presumably any
>company with a subsidiary could sign the NDA, give the design to
>the subsidiary, let them do what they wanted with it, and you'd
>have a really hard time proving who'd passed the secret on. Or an
>employee could ask a friend to sign it, then give the design to the company.
>How could you possibly track that? And what good what it do anyone
>to sue the friend anyway? Its different from 'real' commercial
>secrets, where a company can be very picky about who they allow
>to sign an NDA in the first place.
The court may allow a claim when you can't find who passed it on. It is more
likely to allow a claim against the company you are after when you can show who
passed it on. (Here you are no different to Microsoft, AMD etc, so if they
find this useful and can't find anything better, maybe it really is the 
way to go). Maybe I can't track that, but maybe when you ftp the files
(having given a code which identifies you as someone who signed the OpenNDA)
my ftp server is modifying them. When the lawyers raid (like they do in NDA
cases) we can tell who passed them on.
A second point is that this unscrupulous indevidual or company has to never sign 
the NDA and create endlous subsidiaries to get away with it. The idea is that
you sign one NDA, which binds you to the common part1 and everything else
is on-line and open for you. Your competetors sign the NDA and they can
FTP everything, your guys have to ask the latest subsidiary, and the
subsidiaries keep getting tumbled and closed off. Once you accept that
part1, anything you are later found with which is released under it,
is bound by it. And it is the same part1  for Open as it is for propriatry,
so you are ostracised by the whole industry if you won't accept it. Part2
does not need to be signed because that is like the GPL, only giving away.
 

>4) Since NDAs depend on trade secrets, is a court likely to accept that
>you've made a reasonable effort to protect a secret if you've deliberately
>got very large numbers of people to sign the NDA?
My guess is that the court understands that there is a legal basis for 
licencing know-how and that this is a community in which each member has signed 
contract defining a common understanding. It is a fairly powerful statement
if that court says "No, you don't get your rights, because there are too
many of you." This could have serious carear limiting potential for the
judge and any politicians who then didn't want to do anything about it.

Yours
Ian