On Sat, Apr 7, 2012 at 8:21 PM, R. Diez <[email protected]> wrote:
>
>> [...]
>> You're quite happy to use Linux and GCC, yet almost all the development
>> on those fully open source tools has been by fully paid professional
>> engineers.
>> [...]
>
> That model works well because no single private company controls the Linux 
> kernel development. They contribute to it, but do not control it. The Linux 
> kernel people exchange their views in a mailing list that's not controlled by 
> any single company either.
>
> Imagine for a moment that the Linux kernel and its e-mail lists were hosted 
> by IBM. All other competitors would fork immediately.
>
>
>> [...]
>> It's no different from SourceForge, GitHub or any of the other similar 
>> websites.
>> Like all those websites, the owners invest in the equipment and software to 
>> run the
>> site and hope to make a return, primarily through advertising.
>> [...]
>
> It is different to SourceForge or GitHub in that they do have a registration 
> wall in place for most of the source code repositories. The only exception I 
> know of is the recent opening of the OpenRISC repository, which was not "long 
> ago" as you suggest. But some components typically used by OpenRISC projects 
> are still behind the registration wall.
>
> Another important difference is that SourceForge and Co. do not have a direct 
> financial involvement with the projects they host.
>
>
>> I've taken the numbers off the Wiki, since they are almost certainly out
>> of date - I don't know when they were put there (I could find out by
>> trawling through the Wiki history, but I have better things to do).
>
> I can do that for you. It was Marcus Erlandsson, from ORSoC.se, the diff is 
> here:
>
>   
> http://opencores.org/wiki1/index.php?title=OR1K:Community_portal&diff=397&oldid=383
>
> That was the only change, a blatant marketing/advertising makeover. 8-)
>
>
>> Well of course if you cross-post to two mailing lists, then it will be
>> very clear if your email is being filtered. This accusation gets hurled
>> around for time-to-time, but no one has ever provided evidence. If you
>> have the evidence post it.
>>
>> You want to be a little careful in who you accuse. A lot of the readers
>> of these mailing lists have put a great deal of very professional work
>> into developing this technology for many years. Presumably that is why
>> the likes of NASA, NXP, and Samsung use it in products.
>
> I have nothing against private company contributions or professional 
> contributors. Advertising doesn't bother me either. I have not accused any of 
> you of censuring anything either, you are just playing the emotional card in 
> order to support your views and your paying customer.
>
> I'm simply saying that it's too risky to centralise all development and 
> communications on a private company's web site, as they could succumb to that 
> kind of temptation once all communications are under their control. We have 
> no need to risk it and give any single private company that kind of power.

Hi Ruben

These are concerns we voiced rather vigorously at a meeting held with
most of the project's proponents about a year ago at ORSoC's offices
in Stockholm. ORSoC, it appears, get the message and have been
listening to the community and progressively changing their
controversial policies such as registration to access the source code.

I'm not sure I share the level of concern you have, in this case,
about the project being essentially hosted and controlled by a single
commercially-interested entity, but I do admit there's the potential
for funny behaviour should the controlling party take a dislike to a
particular set of developers or the way the project is heading.

Lately it has been clear that the project is more than just ORSoC's
involvement via OpenCores - the community is growing and so too the
number of commercial interests who are getting involved. I would think
there's enough momentum from outside to sidestep OpenCores and
continue development elsewhere in the event of an upset, and they are
probably aware of this.

But it's my opinion that this would be unfortunate because OpenCores,
whether it's a model open source development site or not, is the
gateway to open source digital design projects on the internet and I
think should be the entry point for the OpenRISC project for the time
being.

>
> The Linux model could work quite well here. The OpenCores site could provide 
> a certified/supported/better experience for paying customers or 
> enhanced/patched/whatever versions of the software/cores/etc, that's fine. 
> Everybody gets their kernel from Red Hat/Ubuntu/whatever, but the kernel has 
> its own independent Web site. We should not give up all control of the whole 
> development and communications to OpenCores. Just imagine the company behind 
> them gets bought up by Sun in the future, there could be a sudden, nasty 
> change of direction.
>
> It also worries me that seemlingly most key decision makers and committers 
> are financially tied to a single private company behind the project. A 
> healthy open-source community does include such members, but certainly looks 
> more varied. My take is that the current model and ways put many people off.

I would say the most active maintainers lately, and even in the last
year, have been the least financially involved with OpenCores. If
you're talking specifically about decisions about how OpenCores itself
is run, then yes, that appears to be entirely down to Marcus
Erlandsson at ORSoC, and we must convince him before we get any
changes through.

But regarding the technical content of the project, I would say we've
seen more from the community than from those doing commercial work.
For instance, although variously commercially involved with the
project previously, Jeremy, Jonas and myself are all still very much
involved and contributing in our spare time (what little we have
outside of our day jobs). Stefan and many others are also doing great
work and long hours on this great project purely for the enjoyment and
challenge. Although Olof works for ORSoC, I understand he spends a
great deal of his spare time, too, helping with the project despite
being occupied with un-related tasks during the day. In fact, it is my
opinion we don't see enough of ORSoC in the community, considering
they employ many more engineers who are doing this stuff day-to-day!
But there's many people who I don't think are directly commercially
interested but who should have an equal say with those that are, on
issues relating to technical or management aspects of the project.

Something we came up with at the meeting in Stockholm last year, but
was never fully revealed and has not eventuated, was a charter for an
OpenRISC project committee, which would deal with such things as
architectural evolution and the management of the source code base,
amongst other things. Roughly, we had proposed a simple committee
setup, where people were voted onto it by the community so long as
they had the best interests of the project in mind, and they
ultimately had the final say in any decisions that were not straight
forward.

We haven't really needed something like this yet, although I suspect
one might be useful to solve the issue (in the 'git vs. svn' thread)
of how projects are maintained in which VCS etc. But I suspect it'd
also be the sort of thing that could put your mind at ease, knowing
that the commercial entity who owned the webservers and employed a few
of the maintainers wasn't running the show entirely. Although, in
summary, I'm not convinced it is given the size of the community and
the amount of contribution we're getting from people who actually have
almost no commercial interest.

Cheers

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

Reply via email to