Funding Global Domination Mk II "The Console"
Stroller
stroller at stellar.eclipse.co.uk
Tue Dec 9 16:13:26 CET 2008
On 8 Dec 2008, at 12:23, Arigead wrote:
> ...
> So you'd have a PC Motherboard, TV out don't know what standard. Do
> all
> modern TV's take the same digital connection all over the world? Then
> you'll have a TV in signal for recording TV. Again I've no idea if
> that
> would mean different hardware all over the world. Ethernet, USB,
> Remote
> control. Nice plastics.
>
> A few companies have toyed with the idea of producing Open Console but
> to my knowledge nobody has ever really done it well ...
Hi there,
I, too, have thought about what open hardware I'd like to see on the
market.
Regarding video players, I think a terrabyte of storage - or any hard-
drive - is excessive for a front-end box, and what I'd really like to
see is a small-front end unit which streams data from the hard-drives
on a back-end server or NAS. I also think that - if you want "just
basically a PC with a disk-drive" - one could probably just buy one
from the host of all the options already out there (Apple MacMini /
iTV, Asus Eee Box, various Mini-ITX form-factor machines).
I also fear there's little "unique selling point" to OpenMoko getting
involved with something that's "just basically a PC" - what's to stop
people from using the same software on a repurposed PC? What's
compelling about buying the hardware?
I have, however, been using MediaTomb a little recently, which is
software for the backend server to stream your music & videos to a set-
top device. Many player devices are supported, but mostly these are
all completely closed, hardware boxes one buys off the shelf at PC
World or BestBuy. I use the PS3 in my livingroom to playback from a
directory of videos on the MediaTomb server, but the PS3 isn't a
terribly good player, as its very fussy about codecs. Like so many
other devices its playback software is closed source.
The problem with silent video front ends is power vs noise. And the
best front ends for MediaTomb have hardware decoding on board. But
they're closed source: http://www.popcornhour.com/onlinestore/
So what I think would be idea would be for OpenMoko - or someone like
them - to produce a low-power media box with hardware decoding, and
open-source drivers for the video-accelerator chip. The Popcorn
devices have such an accelerator, but did I mention they're closed-
source? (VIA used to do CPUs c 1ghz with extra MPEG decoders, but
they have been very late to the game with their open-source "chrome"
drivers).
I should add that complaints about the Popcorn are that its interface
is slow & clumsy - this is easily the sort of thing that users can fix
if they have access to the sources, whilst someone-like-Openmoko can
stay concentrated on low-level drivers - Openmoko's core competency -
for the video hardware. I should add that a built-in TV tuner is
undesirable - should it be for cable, satellite or terrestrial digital
TV - but USB ports would allow the user to tuners as they wish. There
are open-source drivers already available for a number of devices, so
there's no need to waste resources on that.
I have a Linux-based device here which also suffers from slowness &
general quirkiness. It is an IP-KVM for remote access. Mine is branded
Addison-Hughes, but Peppercorn & many other manufacturers sold very
similar devices apparently based on the same software stack - the web-
based GUI is very distinctive. I think the OEM of that was Taiwanese,
but good luck getting any information about it, as they clearly only
sold to people buying their video-capturing chipsets. My IP-KVM is a
godsend, allowing me to work on multiple computers - reinstalling
Windows for customers, for example - without needing a monitor for
each one (and the space that would consume) and to do so without
leaving the comfort of the chair at my main workstation. But it is, as
I said, slow & quirky: there is SO much that the open-source community
could do to improve this device, but I don't even bother asking for
sources because bitter experience has shown that most mainstream
suppliers - basically anyone using Linux in their embedded devices
unless they go out of their way to make it a marketing focus - will
only provide the minimum sources they're legally required to.
EDIT: I was about to add that there used to be an open-source project
that had started working on this & then died; I was sure that was the
scenario 6 months ago, but I now find them successful! Well done! I
shall be reading more after I hit send.
http://okvm.sourceforge.net/kvmoverip.html
http://www.opengear.com/
This bitter experience stemmed from trying to unlock the Wanadoo
Livebox ADSL router to use it with another ISP. I managed to find a
firmware file, run `strings` on it and discovered it uses Linux, but
it transpired that the closed source `adsld` will reject any PPPoA
logon entered in the GUI if it doesn't end in @wanadoo.com. This has
been hacked, but freedom has been severely compromised. The Linksys
WRT routers have shown what cool stuff the open-source community can
do, but AFAIK there are still no routers available with open ADSL
drivers. There are HUNDREDS of devices on the OpenWRT supported
hardware list, but they're all somewhat lacking if you're dependent
upon DSL - Openmoko would be unique if they were to bring an Open ADSL
router to market.
Since I'm writing such a long email, I would like to mention that I've
recently come to better appreciate Openmoko's full openness. THANK YOU
OPENMOKO! In the last week or two I have been looking at a hardware
device - I won't mention its name because I don't want to antagonise
the guy - which users can build and which claims to be "open source",
but isn't, depending upon your definition. I've always thought myself
that if I ever wrote $significant_application I would release it under
a "not for commercial use" license (and offer commercial licenses
alongside), and I thought to do so for exactly the reasons this guy
gives. He says he "doesn't want other people making a profit off [his]
hard work", but when I consider working on his project as a 3rd-party
I just find myself turned-off and almost nauseated by his licensing
mess. The developer has implemented a really neat idea and shown its
potential - he has managed to get the community engaged, but I feel
that's hampered by the number of devices he can turn out of his shed
(and many people don't want to have to build electronics PCBs for
themselves). And when you look at the device, it's nothing really
novel or unique - only a "twist" on existing ideas. Manufacturers have
been turning out functionally very similar devices for decades and
when I look at the chips he has used and posts about related stuff on
electronics forums I can see that he's implemented his device in the
obvious way that any hobbyist or electronics engineer would do, were
they faced with the same set of requirements. The only difference
between this device and something you'd buy in the local high street
is that he has posted a blog describing some of his techniques, but
the market for his "twist" is perhaps too small & niche for the mass-
market. When I want to produce something similar I just feel a big
"fuck you" over his licensing - I want to write the software from
scratch, GPL everything and steal his market, just because he's closed
me out from his existing work. I come to his device with quite a
different perspective, and initially I had little overlap of interest.
Stroller.
More information about the community
mailing list