[maemo-developers] Corporate ownership of open source projects [LWN]

[From: Igor Stoppa]

Hi,
On Sat, 2008-05-03 at 19:58 +0200, ext Hans J. Koch wrote:
> Am Sat, 03 May 2008 17:47:08 +0300
> schrieb Igor Stoppa <igor.stoppa at nokia.com>:
> 
> Hi Igor,
> 
> > 
> > In the end I think what would be realistically possible - and i'm
> > already completely sure that many won't be satisfied and will
> > complain - is that Nokia provides one person whose sole task is to
> > support the community by mantaining the closed code and providing new
> > binaries that link against recent libraries. 
> 
> You're right, I'm not satisfied. The acid test for a Linux device based
> on OpenSource software is that I can compile my own kernel. I want to
> be able to download 2.6.25 from kernel.org now, apply some hardware
> specific patches (GPL, of course) and compile that. With this kernel,
> all hardware of the device should be functional. Whenever a company
> designs such a device, hardware should be chosen that makes this
> possible.
> Note that I'm not talking about firmware. IMHO firmware belongs to the
> hardware, and there might be valid reasons to keep it closed, e.g. to
> fulfill legal requirements for the WLAN chip or for the battery charger
> you mentioned.
> 
> In userspace, closed source applications are OK as long as I have all
> information required to write my own replacement. 
> 
> No help from Nokia is required, except giving us information. I accept
> it if I have to fix the kernel patches for a new kernel version myself.
> And of course I accept that it's my own problem if something doesn't
> work.
> 
> > The community would
> > still be able to set the direction for the development and ask for
> > updates, so that these closed areas would not hold back work done in
> > the open part. Which is the majority.
> 
> I disagree. Any closed part in a Linux system forces developers to
> implement stupid wrappers and workarounds somewhere. If I'm forced to
> use a certain kernel or glibc version, it's not an open system.

Probably i was not clear: i'm just saying that if a new kernel comes out
and the only impediment in using it is that the proprietary module is
not compliant, a new module compiuled against the new kernel should be
made available.

> > And I think it would be only fair that, having Nokia enjoyed the
> > benefits of taking these shortcuts (mostly can be summarized in not
> > using better/more open HW), now it will take also the pain of
> > providing continued support for the closed components.
> 
> That is just additional cost caused by the closed components. Use open
> components and use the money for something more useful.

And I thought that I was writing in good english. Poor me. I'm talking
about the past and you want open components. Is Nokia supposed to go
around and replace all the sold devices with open hw?

> > 
> > This is my personal opinion - not to be taken as a promise or plan -
> > but as an advice in what the community could do/demand to keep the old
> > devices alive.
> 
> I'll continue demanding open hardware. I'm really fed up with so-called
> Linux devices where I can't make the hardware work with free software.
> I've a similar problem ATM with the Eee PC's unsupported WLAN and its
> buggy BIOS...

And that is good, but once certain things have happened, there is no way
back. An I think I have explained what I think are the obstacles in
opening specs of existing hw. Not that i wouldn't be happy to be proven
wrong.

> > Anyway note that in order to do proper low level kernel development,
> > one needs also measuring tool and special boards that allow for
> > precise measuring of what the sw is doing.
> 
> Oh, please, leave that to the people who want to do kernel development.

Yeah, I happen to be one of them, sorry for the personal point of view.
OTOH the fact that you don't care doesn't mean others are not
interested. If you talk about an open device, it's the whole stack that
should be targeted, no?

> > Nobody in the community
> > has such setup,
> 
> How do you know that? I compile kernels for ARM devices almost every
> day and also build the bootloaders for them.

But not our own boards, i guess. If you do it would be interesting to
know how you have obtained one. Compiling for your custom or eval board
means very little. To give you an example, every pad must be checked so
that it has the right idle configuration depending on what is connected
to it, or high currents can be seen. Hard to do without schematics.

> > so some help from Nokia for validation would still be
> > required.
> 
> What do you want to "validate"? I agree if my warranty is void if I use
> my own kernel, just let me do it.

Maybe i'm too idealistic, but my idea of open source is a working model
where if one has an interesting piece of code, the code can be committed
back to the community.

Code "compile tested" or "boot tested" should not be good enough for
being included in repositories that are meant to be used for rolling
kernels that will be provided to less experienced users.

I don't care if you void your warranty and blow up your device, but i
get interested if you start distributing such kernels to common users.

What i was talking about would be a way to have Nokia approval for code
that affects sensitive areas that community developers have no way to
verify.

>From that point of view, TI has improved its feedback to the ml and many
TI folks provide such verifications done on their reference boards.

I'd like to see Nokia doing the same.

-- 
Cheers, Igor

---

Igor Stoppa
Next Generation Software
Nokia Devices R&D - Helsinki