"it's worked very well for Linux."

Really? As someone struggling to get FC3 configured, googling for
information produces many more people with Linux problems than there are
solutions out there.

BTW anyone care to help with my problem getting Slimserver 5.4.0 to start up
correctly?

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of Phil Karn
Sent: 13 March 2005 08:26
To: Slim Devices Discussion
Subject: [slim] Open firmware for SB2?


Patrick Dixon wrote:
> IMHO, the two biggest threats to Slim Devices' competitive advantage are:
>
> * Product design - most 'normal' people think the Roku styling is better.


Maybe. Personally, I think basic functionality and reliability are far
more important. Then again, my Squeezeboxes are all black.

> * Simple software installation - most 'normal' people can't (or can't be
> bothered) to spend hours reconfiguring their computer to get an

application
> running - if it doesn't work reliably straight from the tin, they'll just
> send it back and move on.


Absolutely!

> The second produces a major dilemma - the opensource community is
> notoriously geeky and seems to just love wading though reams of poorly
> documented or undoccumented source code to re-configure it for some

strange
> combination of a Linux installation. But if the company concentrates on
> supporting and making the software work seamlessly with Windows and iUnix,
> it will probably alienate the geeks.


I don't think that's really a big dilemma. These sorts of "sponsored
open source" projects work best when the volunteers work on the features
that personally interest them, and the commercial sponsor acts as the
project "glue" -- merging patches, conducting regression testing, and
managing the release cycle. I can't see how any geek could oppose the
mere existence of a stable Windows version (though that's arguably a
contradiction in terms) so long as the code he's interested in remains
open and hackable.

What I *do* find discouraging is the distressing unreliability of even
the 5.4 version of the server software. I shouldn't have to install the
version du jour just to get a fix for a bug that keeps crashing my
server in routine usage.

There ought to be two code bases: a relatively stable, no-frills version
with an emphasis on robustness, and an experimental version with all the
latest gimmicks. As new features prove themselves and become stable,
they can be backported to the stable version. This is hardly a novel
concept; it's worked very well for Linux.

Most of the volunteers would probably prefer to play with the
experimental release, while the people at Slim Devices would maintain
the stable version. After all, their product is pretty much useless
without a server to drive it.

--Phil