Re: [OpenBIOS] Intel holding back information

[Timothy J. Massey]

On Sun, 08 Aug 1999 22:38:23 -0400, Ed Brinker wrote:

<description of clean room techniques snipped>
> How might we implement such an approach.  It would probably require two
> separate mail lists.  One for the group to reverse engineer the other BIOS
> and a second for the group creating the new BIOS.  The could be no
> communication between the two groups about the BIOS's other than the
> "identified functions."

Now here's the catch:  how do you prove a negative, that the clean room
group did not indeed disassemble a BIOS, did not sneak a look at that other
mailing list, etc?

If you're going to do a clean room technique, you need that strictly
enforced.  That means being able to control **EVERY** last action of both
the clean and dirty room people.  The information generated by the dirty
room group and the actions of the clean room group need to be strictly
controlled.  You need to say that there is **ABSOLUTELY** no way that the
clean room people could have gotten information from the dirty room group. 
That means monitoring all means of communication.

In our day and age, on an OpenSource project, that is not going to happen. 
A single e-mail from the dirty group to the clean group ruins the entire
effort.  We don't exactly have the resources to sequester the two groups,
now do we?  Compaq and Phoenix did.
> The first group could make no comments about the work in progress of the
> second group as that could be intrepreted as providing "guidance."  The
> second group    
> could not ask if the function was really necessary as that would be
> intrepreted as asking for "guidance."
>
> If we are to take this route, then perhaps it is best we split soon to
> avoid the process of legal "contamination."

Again, it just can't happen.  There are too many ways for the arrangement to
be broken, and even if every member of each team did in no way break the
barrier, how could it be **PROVED** that they didn't?  Again, proving a
negative.  You need a *tremendous* amount of evidence to do this...

Tim Massey

-
To unsubscribe: send mail to [e-mail]
with 'unsubscribe openbios' in the body of the message

Intel holding back information

[William Meyer]

> Now here's the catch:  how do you prove a negative, that the clean room
> group did not indeed disassemble a BIOS, did not sneak a look at that

other
> mailing list, etc?

You don't prove it. You assert it. And you document every move a clean room
member makes. Then it's up to a claimant to prove otherwise. At least in the
US.

William Meyer

-
To unsubscribe: send mail to [e-mail]
with 'unsubscribe openbios' in the body of the message

Intel holding back information

[Gustav Foseid]

William Meyer:
> You don't prove it. You assert it. And you document every move a clean
> room member makes. Then it's up to a claimant to prove otherwise. At
> least in the US.

I believe there are some experience with "clean room" approaches in the
free software community, especially when it comes to Java implementations.
Som information could probably be found on the websites for projects like
Japhar (http://www.japhar.org/), Kaffe (http://www.kaffe.org/) and GNU
Classpath (http://www.gnu.org/software/classpath/).

Consulting these projects on how they have dealt with "clean room", free
software and diversing laws in different countries seems like a good idea
to me.