Seneca breaks silence on Fedora Remix

Chris Tyler posted an update today.

Estimated release date is 20th, but lots of work to do.

http://lists.fedoraproject.org/pipermail/arm/2012-July/003631.html

New estimated release date for Fedora Remix from Chris Tyler:

"aim to ship next week", with 3.2.27 kernel.

http://meetbot.fedoraproject.org/fedora-meeting-1/2012-09-26/fedora-meeting-1.2012-09-26-20.00.log.html

Also, the RPF is reportedly working with Fedora on firmware licensing:

>> I heard that there won't be an official Fedora build of the Raspberry Pi remix because

>> the firmware license is not compatible with the Fedora firmware guidelines.

>

> Correct at this point in time. I meant an official remix stable release.

> Fedora legal and the board are working with the foundation to try and resolve the firmware issue

http://lists.fedoraproject.org/pipermail/arm/2012-September/004085.html

Seneca blew it at Pi launch and instead of communicating with end users they just bailed. Then at the end of August the new release candidates quietly appeared from nowhere, but still communication (and support) remains woeful. Are we to expect a sudden (and miraculous) transformation in their customer service skills, or will those who "invest" in Fedora find themselves high and dry at some point in the future?

One of the reasons that FOSS growth remains less than stellar is that too many people are willing to keep garbage on life support rather than hit it on the head with a shovel and quietly bury it. A small extinction event would create space for smart folks who are willing to learn from the mistakes of the past and are able to create applications in line with the current environment.

I can't believe that FOSS advocates are Creationist in their beliefs!

Sorry, I went a bit ranty there for a moment! Still, good ideas that are poorly implemented are just as much of a waste of time as beautifully marketed bad ideas. The end result is what matters...

> Seneca blew it at Pi launch and instead of communicating with end users they just bailed.

My impression of what happened is that Seneca and the RPF were in a "partnership" that

neither really wanted to be in, and neither was willing to make the necessary investments

to make the project a success. 

RPF clearly wanted Ubuntu as the preferred OS, but that didn't work out.  They turned to

Fedora instead, but as far as I can tell, made absolutely no investment other than a single

alpha board.   I haven't heard the RPF citing any advantages of Fedora over Debian,

other than noting that the Fedora folks had put a lot of effort into it.  When the first release

had problems, they dropped it like a hot potato.

Similarly, Fedora considers ARM a secondary architecture, and considers the RPi Remix to be

a second-class member of the ARM family.  They realize that the RPi would require a lot of work

to do X11 acceleration, and to do a hard-float version, and to integrate the RPi kernel changes

into the current kernel, especially with device tree support and the troublesome USB driver. 

They realize that the RPi isn't suitable for self-hosted compilation in their build farm, due to lack

of ram and lack of SATA.  They are focused mostly on ARMv7 and later architectures, and don't

want to be supporting ARMv6 if and when ARM gets promoted to a primary architecture.

They also aren't set up to support hundreds of thousands of RPi users.

> Seneca blew it at Pi launch and instead of communicating with end users they just bailed.

My impression of what happened is that Seneca and the RPF were in a "partnership" that

neither really wanted to be in, and neither was willing to make the necessary investments

to make the project a success. 

RPF clearly wanted Ubuntu as the preferred OS, but that didn't work out.  They turned to

Fedora instead, but as far as I can tell, made absolutely no investment other than a single

alpha board.   I haven't heard the RPF citing any advantages of Fedora over Debian,

other than noting that the Fedora folks had put a lot of effort into it.  When the first release

had problems, they dropped it like a hot potato.

Similarly, Fedora considers ARM a secondary architecture, and considers the RPi Remix to be

a second-class member of the ARM family.  They realize that the RPi would require a lot of work

to do X11 acceleration, and to do a hard-float version, and to integrate the RPi kernel changes

into the current kernel, especially with device tree support and the troublesome USB driver. 

They realize that the RPi isn't suitable for self-hosted compilation in their build farm, due to lack

of ram and lack of SATA.  They are focused mostly on ARMv7 and later architectures, and don't

want to be supporting ARMv6 if and when ARM gets promoted to a primary architecture.

They also aren't set up to support hundreds of thousands of RPi users.

Thanks for the reply coder27 - so, the plot thickens...

I think that Eben mentioned in an interview that they originally had an eye on Ubuntu, but Canonical dropped support for v6 a very long ime ago (after 9.04?), so I can't really see it as a credible reason for rolling out a piece of hardware with a broken O.S. nearly three years later. Also, why did they throw their weight behind Fedora, rather than say, Arch, or Debian, I wonder? I suspect your impression regarding an unwillingness to make investments is right on the money. Perhaps Seneca put in a very cheap bid without realising what they were taking on. So why persist? As you suggest Fedora aren't set up for consumer-level support - I feel I need a degree in "Geek" to extract any information from their utterly confusing sprawl of an internet presence! This is more of a reflection on my lack of hacker-fu than a criticism of their set up (which has probably served Fedora enthusiasts reasonably well over the years), but crikey, it's hard work in there for a mere human...

I suspect that with the benefit of hindsight there wouldn't be a lot of Linux development houses willing to build a boutique O.S. for a superceded architecture without first taking a fat cheque and an assurance that the Foundation would keep their end of the bargain regarding firmware / driver development and any other reasonable support. It doesn't explain the Fedora debacle though (or the impression that no-one at Fedora Towers is willing to concede that this is a debacle).

Perhaps Fedora are still under contract. Or perhaps they are battling on out of pride. *shrug*

> So why persist?

Good question.

I believe Seneca's funding comes mostly from government and other grants,

and it probably looks good on such grant applications to say you're working on

a popular educational device.

Their primary motivation doesn't seem to be to capture market share on the RPi,

or to regain "recommended OS" status.  They seem to be weighing the concern

of appearing to be taking too long to generate a release against the concern

of issuing a release that will not be well received, like the Fedora 14 release.

Yeah I agree that they completely botched the release then subsequent comunications(or the lack of).

I even sent an email to one of the boss men there and offered help....nothing ...nada

Any way it will be interesting to see what if anything of use they will actually produce, they have certainly lost traction and with several Disros specifically for the ' Pi now available then "WHY ??? " is perfectly valid question, why would you want to use a distro which is maintained by folk who are totally incomunicative... pitty really!

BTW if they have Sold or have orders for .5 million units are they still a minority platform on a minority Arch ?

RPF clearly wanted Ubuntu as the preferred OS, but that didn't work out. 

IMHO, that shows a clear disconnect from reality. Ubuntu have a track record of only doing what interests themselves. It was never going to work out unless Mark Shuttleworth decided he wanted to be involved.

Similarly, Fedora considers ARM a secondary architecture, and considers the RPi Remix to be

a second-class member of the ARM family.  They realize that the RPi would require a lot of work

to do X11 acceleration, and to do a hard-float version,

Raspbian is in much the same situation, but it came together quite quickly based on a single individuals interest in making it happen.

and to integrate the RPi kernel changes

into the current kernel, especially with device tree support and the troublesome USB driver. 

The kernel changes necessary are beginning to be pushed into mainline, that effort may take some time though and there's an ongoing effort to redo the usb driver in mainline - not due to the RPi but due to lots of other devices using the same synopsis device.

It all suffers from the same problem though, it takes sustained community involvement to kickstart some of these efforts. None of the RPF, Fedora, Ubuntu etc seem interested in funding things like the accelerated X driver, so these things may simply never happen.

> BTW if they have Sold or have orders for .5 million units are they still a minority platform on a minority Arch ?

Don't confuse "secondary" with "minority".

A primary architecture means, among other things, having strict release criteria.

A serious regression that affects a primary architecture will delay a release, but

the same problem affecting only a secondary architecture will not.  A secondary

architecture may not get released at the same time as the primary architectures.

The x86 fedora builds are unified with a single kernel image, not so currently for

ARM, although that's the goal.

Fedora 17 was released on 29 May 2012 for primary architecures.

Fedora 17 for ARM was released on 18 June 2012 for the officially supported ARM

machines, not including the RPi Remix, which is not officially part of Fedora.

http://fedoraproject.org/wiki/Architectures/ARM/Fedora_17_GA

"Raspberry Pi support is not included in Fedora 17 GA because it does not meet the

Fedora release guidelines. Specifically, the Raspberry Pi relies upon a custom kernel

version, not upstream, as well as special GPU binary blobs which are not acceptable

under the current firmware exception."