Linked by Thom Holwerda on Thu 5th Oct 2006 20:44 UTC, submitted by jasper
Red Hat Red Hat has an NDA cooperation with Marvell for the wireless chips that they want to use for the One Laptop per Child-project. The idea of this is that both parties think Marvell will be more open in the future, but this is absolutely not the path they should walk, according to OpenBSD's Theo de Raadt. "I am getting really tired of 'open source' people who work against the open source community. Our little group can probably take credit for having 'opened up' more wireless devices than the rest of the community, and therefore we feel we have a better grasp of the damage OLPC has done here."
Thread beginning with comment 168888
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE: Right!
by twenex on Thu 5th Oct 2006 22:38 UTC in reply to "Right!"
twenex
Member since:
2006-04-21

I'm confused. How can you write an open-source driver with an NDA? The source should clearly show details of the implementation the NDA was meant to keep private, non?

Reply Parent Score: 3

RE[2]: Right!
by diegocg on Thu 5th Oct 2006 22:48 in reply to "RE: Right!"
diegocg Member since:
2005-07-08

Yes, the source code does indeed reveal many of the things that the NDA spec tells you. But certainly it's _not_ a spec, it doesn't reveals everything.

In my book:

- closed drivers: bad
- open drivers, no spec: good
- open drivers and/or specs: great

Just don't ask me why Theo spend his time arguing that "open drivers, no spec" are somehow a openness crime. It's not the best situation possible but certainly they're great news for the open source community. Except for Theo, that is. I hope this is not related to the fact that the open driver redhat is writing is for linux and not for openbsd...

Edited 2006-10-05 22:56

Reply Parent Score: 3

RE[3]: Right!
by Marcellus on Fri 6th Oct 2006 06:41 in reply to "RE[2]: Right!"
Marcellus Member since:
2005-08-26

- open drivers, no spec: good

It's not "good", only "less bad".

With closed drivers you have to assume that the people who wrote the driver knows what to do.
With open driver, but no spec, you have to assume that the people who wrote the open driver actually knows what they are doing with their hidden specs.

I'd rank it as:
-closed drivers: bad
-open drivers, closed spec: less bad
-open drivers, reverse engineered: decent
-open drivers, open spec: good

Reply Parent Score: 3

RE[3]: Right!
by Tor85 on Fri 6th Oct 2006 07:39 in reply to "RE[2]: Right!"
Tor85 Member since:
2006-07-04

I wouldn`t be so sure, that it can be ported evrywhere. If Red-hat write driver it will be probably under GPL license. That`h mean, that main *BSD systems won`t port it into theire kernel.
Well I`m pro-GPL person, but I`d like to see evrything work in *BSD systems either.
Ofcourse I can be wrong and maybe Red-hat will do driver in BSD-like license. No idea.

Reply Parent Score: 2

RE[3]: Right!
by ddpbsd on Fri 6th Oct 2006 14:07 in reply to "RE[2]: Right!"
ddpbsd Member since:
2006-04-29

Just don't ask me why Theo spend his time arguing that "open drivers, no spec" are somehow a openness crime. It's not the best situation possible but certainly they're great news for the open source community. Except for Theo, that is. I hope this is not related to the fact that the open driver redhat is writing is for linux and not for openbsd...

And when Intel stops supporting their older products where do we go? To developers without documentation.

What about when a driver doesn't work on sparc64 (or has major issues)? Where do we go? Developers without documentation.

That doesn't sound "good" to me.

Reply Parent Score: 3

v RE[3]: Right!
by Mitarai on Thu 5th Oct 2006 23:19 in reply to "RE: Right!"
RE[4]: Right!
by twenex on Fri 6th Oct 2006 00:46 in reply to "RE[3]: Right!"
twenex Member since:
2006-04-21

Nothing I've said is any indication that I'm not in favour of open source drivers. It isn't any more complicated than proprietary software, actually.

Reply Parent Score: 3

RE[2]: Right!
by Soulbender on Fri 6th Oct 2006 03:26 in reply to "RE: Right!"
Soulbender Member since:
2005-08-18

"The source should clearly show details of the implementation the NDA was meant to keep private, non?"

It's perfectly possible to obfuscate code to the point that it's virtually impossible to understand without having the specs.

Reply Parent Score: 5

RE[3]: Right!
by twenex on Fri 6th Oct 2006 14:28 in reply to "RE[2]: Right!"
twenex Member since:
2006-04-21

True enough.

Reply Parent Score: 1

RE[2]: Right!
by Morty on Fri 6th Oct 2006 08:43 in reply to "RE: Right!"
Morty Member since:
2005-07-06

The source should clearly show details of the implementation the NDA was meant to keep private, non?

No, not neccesary, lots of implementation details will not have any relevance on the driver sourcode. It's more a convinience thing. It's hard and expensive work to go through an internal specification and design documentation, to remove the information desired or requred to keep private. And you still have to include enouch information to make the specifiaction usefull.

In many cases it's plain simpler and cheaper to make anyone needing the info sign an NDA, and give access to the internal specification and design documentation.

Reply Parent Score: 2