On 3 July 2012 19:33, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
The thing to
remember is that it is not finished yet. This is not the
final product.
When I looked (after it was hyped i nthe press, and after I had to
navigate past various Rpi pages to get the bits I wanted from Farnell), I
didn't see anything to suggest that this was not the final product.
Neither in the press nor on the sites I (briefly) looked at.
The motto of Open Source is "release early,
release often". They
shoved the hardware out the door as soon as they could. The OS(es)
is(are) not finished, there is no case, there is no ecosystem or
infrastructure around it.
I can think of 3 things wrong with that :
1) It's not open. You yourself said that. Given a piece of oepn-sourve
softwre, it is possible (if you are sufficiently technical) to figure out
a lot from the source. Given a piece of closed-source software, running
on a machine which at the time had no released scheamtics (they are
released now, but as you say, things change), you have an alomst
impossible job figuring it out without docuemtnation
2) It's not software
3) And therefore it's not free. A peice of software can be easily copied,
and thus it cna be made esentially free. A piece of hardware cannot.
DOn;t get me wrong, a piece of hardware is a physical thing and it;s
obvious you have to pay for it. But even thought the Rpi is cheap, it's
still expensive enough that I (for one) would bot buy one unless I knew
it coule be useful. Buying things you haev no use for is simply wasting
money.
AS a result, releasing in a 'beta' version and not stating it was not
finished seems like a rather bad idea.
No, the hardware is not FOSS. It's?not FOS and it's not S(/W).
But it is designed and intended to run FOSS OSes. Exclusively, in
fact. It is intended to be a cheap Linux machine. The fact that other,
non-FOSS OSes are being ported to it - well, one so far, RISC OS - is
incidental.
As for the unfinishedness and the beta state of this first version - I
agree. It's barely been mentioned. It comes up in interviews and
things occasionally. I think that they really ought to have emphasized
it, not hidden it.
But then, whereas I admire the spirit and intent of the project, I
don't actually like the execution, myself. I don't like the
proprietary GPU, I don't like the closed-source graphics driver, I
don't like the crappy old-spec low-powered underperforming ARM core. I
don't like the rather underspecced amount of RAM, which isn't really
adequate to run Linux with a GUI. I don't like the fact that the
Ethernet port is a USB device. I don't like the fact that there's no
fast local storage.
Now, some of these compromises were doubtless necessary to make it so
very cheap, but then again, here is a commercial rival:
http://liliputing.com/2012/05/via-apc-a-49-android-computer-with-an-arm11-c…
http://apc.io/
Item Description
Model APC 8750
Software Android 2.3 (PC System)
Chip VIA 800MHz Processor
Memory DDR3 512MB Memory
2GB NAND Flash
Graphics Built-in 2D/3D Graphic
Resolution up to 720p
Input and Output HDMI
VGA
USB 2.0 (x4)
Audio out / Mic in
microSD Slot
Network 10/100 Ethernet
Size 170 x 85mm (W x H)
Neo-ITX Standard*
US$49, or ?31.25.
Better CPU, more RAM, more local storage, but inferior graphics and no GPIO.
The AllWinner A10 SoC is a 1-1.25GHz ARM SoC with a complete system
with GPU onboard for $7 in bulk. It's going to shake up this market
and I think soon there will be multiple RasPi rivals at around the
same price-point.
If this means that an ecosystem of ARM Linux distros for low-end
desktops develops, and lots of ARM Linux apps for low-spec machines,
then all might be well - if they can all run each other's software,
fine. If the kernel and the underlying graphics drivers are different
but at the OS API level they're intercompatible, fine.
If they're all different iterations of the ARM CPU core and thus are
/not/ software compatible, problem.
If the latter happens, I think RasPi may lose out, because rivals will
be able to offer much more powerful machines for competitive prices.
If some of those rivals also offer functionality equivalent to GPIO,
and bundle some nice features like optional dual cores (which Linux
can exploit well) or on-board SATA, then I think RasPi may end up
getting slaughtered.
OTOH, if that happens, then perhaps they can just release a RasPi SE
or Mark II or something, with an AllWinner CPU and the same
form-factor or something.
There is a chance, I reckon, that the RasPi may end up like the IBM
PC. Quite respectably successful in its own right, but more important
for the whole market sector it spawned of enhanced-but-compatible
machines.
At the end of the day, if it means that the endless march of boring
beige x86 boxes is replaced or supplanted or supplemented with little
silent, cool-running, all-solid-state no-moving-parts legacy-free ARM
machines running Linux, getting kids in the developing world online at
prices so cheap that a couple of billion poor people can afford them,
then I will be delighted and I don't care who does it - RasPi
Foundation, Via, nVidia or whoever.
Hell, if the Chinese come up with a rival system based on the
Longson/Godson family of MIPS-compatible chips, I'm cool with that as
well. Variety is good.
Maybe it's time for the RISC resurgence.
--
Liam Proven ? Profile:
http://lproven.livejournal.com/profile
Email: lproven at cix.co.uk ? GMail/G+/Twitter/Flickr/Facebook: lproven
MSN: lproven at
hotmail.com ? Skype/AIM/Yahoo/LinkedIn: liamproven
Tel: +44 20-8685-0498 ? Cell: +44 7939-087884