On Sat, 2012-04-07 at 11:28 +0100, R. Diez wrote: 
> 
> > +For general information about Or1ksim, post to both OpenRISC mailing
> > lists:
> > +
> > +  [email protected]
> > +  [email protected]
> > +
> > +(yes we do mean crosspost - they have disjoint sets of readers).
> >
> >The following process applies to both the SVN HEAD and release
> >branches.
> >+
> >+* Submit the patch to both OpenRISC mailing lists 
> 
> 
> I tend to posts patches from time to time, and I feel that you are 
> abusing your position of power in or1ksim in order to pursue a
> political
> agenda, namely that you would like to move all OpenRISC development
> to 
> OpenCores.org . You are trying to get all OpenRISC source code under
> OpenCores' Subversion and move all mailing list traffic there too.

Hi Ruben,

That is a most unfair accusation, and I am not sure why you are making
it.

Openrisc.net was set up by Jonas Bonn, one of the major recent
professional contributors to the OpenRISC project in April 2011. He set
it up to provide functionality, particularly mailing lists, a Wiki and
later git support, which he felt was missing from the main opencores.org
website. He also disliked the requirement at that time to register even
for read access to the source code. Those were issues that many of us
across the community agreed with.

The owners of opencores.org, ORSoC AB, responded after a couple of
months by providing mailing lists, which is why there are OpenCores
mailing lists, an opencores Wiki and git support. And it has to be clear
that the pressure from Jonas was important in making those changes
happen.

We had a meeting of many of the major OpenRISC contributors in Stockholm
last July. Johan Rilegard, the CEO of ORSoC AB and Marcus Erlandsson,
their CTO took part, as did I, Jonas Bonn, Julius Baxter, Olof Kindgren,
Stefan Kristiansson and a number of others you will recognize from IRC
and the mailing lists.

It is perhaps worth noting that many of the major contributors to this
meeting are not just hobbyists, but are or have been paid professional
contributors. My company, Embecosm, was paid to develop the GCC 4.5 tool
chain to production standard, Olof works for ORSoC AB, Julius used to
work for ORSoC AB, Jonas was paid to develop Linux for OpenRISC.

We discussed many things, including what we felt was wrong with
OpenCores and how it was managed. We discussed the possibility of
setting up OpenCores as an independent foundation, but ORSoC AB made it
clear this was not a route they were ready for. I for one was willing to
accept this, having lived through the period in 2007 when opencores.org
ran without commercial backing and was almost unusable.

Jonas was asked if he would merge his openrisc.net OpenRISC mailing list
into the opencores.org OpenRISC mailing list. He indicated he was
prepared to consider this, but so far has not taken up the offer.

In the meantime *all* those present agreed to cross-post to both mailing
lists, to ensure no one was left out. That is why for example the
openrisc.net wiki and mailing lists are promoted and linked to from the
opencores.org website.

Up until the last few weeks everyone has been happy to do that. I am not
sure why you feel it is so important to break that concensus.

> In this particular case, the OpenCores list has little traffic that's
> not already in this list, and forcing the cross-posting on us is in my
> view a move to undermine this list in the hope that it will fade out
> of existence.

Not true I am afraid. When you register at OpenCores for the OpenRISC
project you are by default also registered for the opencores.org
OpenRISC mailing list. So the OpenCores mailing list has a much wider
readership, although most active contributors are registered on both
lists.

> In a truly open environment, the maintainers would be happy to take
> any patches they can get, with whatever means, whether the patch's
> body is inlined or not. Should they have a concern about the patch's
> contents, they would then discuss it themselves with other maintainers
> or interested people wherever they like, without forcing anybody to
> post to any list in particular.

Personally I don't care whether patches are inlined or not, since my
mail client will inline text attachments for me anyway. However it is a
widely accepted convention in the open source software community, and as
you will note from Stefan Kristiansson's email this morning, others like
it. In particular git can accept inlined mail messages directly (using
git am).

All projects have agreed mailing lists to use for patch submission. Just
try posting a GCC patch to the GDB mailing list if you don't believe me.
In the case of OpenCores the community has been happy for the past year
to cross-post to two mailing lists, recognizing that we have an
unresolved issue.

Now one of the great strengths of open source is that you can fork
projects. However with forking comes a risk - you divide your forces. It
has happened once already before. Beyond Semiconductor
(http://www.beyondsemi.com/page/) split off as a commercial operation to
develop OpenRISC based chips, taking with it Damjan Lampret, Matjaz
Breskvar and other OpenCores founders, and their input to the project
has now been lost.

Open source hardware is still small. If we are to change the commercial
proprietary world, we must stay unified and with a single clear presence
in the world. For the past 13 years, OpenCores has been that single
clear presence, and I have not seen a compelling argument from anyone to
change that.

I hope this explains the background, and you will realize I was
documenting the agreed community position. Something I have been doing
at OpenCores for the past decade.

Best wishes,


Jeremy

-- 
Tel:      +44 (1590) 610184
Cell:     +44 (7970) 676050
SkypeID: jeremybennett
Email:   [email protected]
Web:     www.embecosm.com

_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to