The grit in the cogs of humanity since time immemorial
A "system auto-generated error" left hundreds of eager would-be Raspberry Pi owners fuming yesterday after they were told their prized micro-computer boards wouldn't arrive until the summer. The notification was made by Element 14, one of Raspberry Pi's two suppliers. Punters access the firm's website to check on the status …
The grit in the cogs of humanity since time immemorial
To be honest, checking that the device doesn't screw up other system seems perfectly sensible to me. Were RPi unaware of this before they started shipment?
No they weren't unaware - but dev boards are frequently not tested. The plan was to test the educational release (i.e. the cased release) later in the year.
The media got a bit excited, the foundation had to offload the distribution (and scale up production).
They're no longer dev boards by any reasonable metric.
RS and Farnell are (annoyingly, because I'm in the lucky 10k) doing the right thing.
It would appear if they'd ordered one, they might have realized the issues they were going to have to face. I presume the next step will be to slowly trickle out Raspberry Pi units, then announce they've found a design flaw, realize they're not selling at enough to cover RMA/repairs and and...
"The media got a bit excited, the foundation had to offload the distribution (and scale up production)."
But were they unaware that this would mean lengthy certification, causing months of delay to shipments?
Were RS and Farnell aware of that?
It seems to me that they were offering shipment dates which they ought to have known were never possible.
I managed to avoid said project, despite being very interested in it from exposure to its "predecessor in essence" the GP32 / GP2X. That's where the idea for that device came from, and at first it looked like a really good project.
3 years on, people are STILL waiting for those first-day orders and even getting bypassed by new orderers who want to pay ridiculous sums of money for what is basically a glorified GP2X /BeagleBoard. And *ALL* due to disgusting business practices that are excused by being "amateur" / "nonprofit" / etc.
Sorry, if you're selling a product, you're a business. You may be nonprofit but still a business. And you ALWAYS have certain obligations to your customer / the law no matter what happens or what definition you want to use.
Sadly, the RPi seems to be heading down the exact same lines, even with its "fanatics" posting on the forums about how people should shut up because "it's non-profit so what do you expect"? I expect the same as any other business gives me, and especially the same as any other business supplying a school would give that school. And if you're AIMING at schools, as stated on national TV broadcasts, amateur hour isn't the way to go.
I DIDN'T pre-order the OP precisely because of all that mismanagement, despite being there from before Day One (and still wouldn't touch them now, if I'm honest). I did pre-order the OP only because I was given the impression it was an ORDER that could be fulfilled almost immediately and then got stung in the whole "launch day" debacle.
I predict the next problems will involve distribution, queues, ordering, profit margins, cases, "upgraded" versions, component shortages, etc. I reckon that by the time the suppliers get to the point where they can churn through the customer lists for sending out orders, the hype will have died down and the number of people interested will be minimal.
I truly HATE seeing good ideas fail because nobody knows how to do business, or bothers to find someone who does.
Knowing how to do business was the first thing the West lost. Once that was gone, businesses were free to send all our jobs to China.
That really doesn't make any sense at all.
I am sure that the first XBOX 360 prototypes shipped to game studios were not CE marked. Developers need systems that are good enough for development, not perfect systems. There are problems with the current Raspberry Pi that will presumably be fixed in the model B+. The alignment of the magnetic jacks on the cicuit board is a bit crooked and perhaps the Boot ROM in the Broadcom microcontroller will be updated to support class 8 cards, but who cares? When these problems are fixed I will simply go out any buy a cased retail model of tbe model B. What I need is a development system now. I would like to suggest an experiment to Farnell, collect 10,000 part built Raspbery Pis from the factory in China where they are sitting awaiting delivery of magnetic jacks and put them in ebay. My prediction is that they would sell for £100 each.
They've already sold the first 10k - they can't sell them again.
For £100 you can get a Beagleboard, which is faster, has more features, is well tested and supported with a range of peripherals.
You'd have to be a bit mad to pay £100 for the Raspberry Pi. Not sure you'd find 10,000 of them.
> They've already sold the first 10k - they can't sell them again.
Actually they haven't :-) All they have is a few thousand pre-orders and a few hundred thousand
"expressions of interest".
Payment has not been collected, units have not been shipped, so no sale has been final.
I was not suggesting that they sell the first 10k. I was suggesting that they probably have a further 10k sitting in a factory in China awaiting the delivery of magnetic jacks that serious Raspberry Pi developers would be prepared to use.
I have an Atmel AVR development system sitting on my desk in front of me and I can assure you that it does not have a CE mark on it. It is ridiculous to suggest that the first few batches of Raspberry Pis will be shipped to anyone other than serious developers such as myself.
No it's not ridiculous at all, the Raspberry Pi launch was all over the popular media. It even made super hot deal at Hotukdeals, definitely not a serious developer site. "Regular" people were jumping over themselves to buy one.
In hindsight the first batch should never have been marketed widely as a cheap PC. There's a boatload of other issues too (like the SD card incompatibilities and the lack of any X11 acceleration) to sort out too before it becomes something a something the public will find enjoyable.
In their desire to beat Lady Gaga's popularity they just flew too close to the sun. Let's see if they can recover.
"For £100 you can get a Beagleboard, which is faster, has more features, is well tested and supported with a range of peripherals."
Actually, the Beagleboard ISN'T 'tested' or certified either. From the Beagleboard manual:
This evaluation board/kit does not fall within the scope of the European Union directives regarding electromagnetic compatibility, restricted substances (RoHS), recycling (WEEE), FCC, CE or UL, and therefore may not meet the technical requirements of these directives or other related directives.
Negative dear AC, straight from the 2nd page of the Beagleboard manual:
"NOTE: This equipment has been tested and found to comply with the limits for a Class B digital device, pursuant to Part 15 of the FCC Rules."
It has been FCC certified. That other phrase is just a general disclaimer.
It's interesting what you say about marketing, because apart from blog posts on the website and occasional press releases, very little actual direct marketing was done. It was almost all word of mouth organic stuff. Which is very difficult to turn off - do you a) Take a call from the BBC wanting to talk about it b) Not take a call from the BBC?
Prototypes are not 'placed on the market'.
Everything 'placed on the market' or in layman's terms sold is supposed to comply with *all* EU regulations relevant to whatever it is. A CE mark is an indication that the manufacturer or importer claims that something does comply with all relevant EU regulations and the manufacturer/importer may be required to prove that claim is valid.
There is no question that the Pi should be CE marked. The question is which EU regulations are relevant, if the Pi complies with those regulations, and how much time/work/expense the manufacturer/importer feels is needed to provide evidence supporting their claim of compliance.
So no they can't legally sell prototypes without a CE mark, they could give them away, they could mark them regardless to sell them and 'wing' it. The hype around the Pi and subsequent level of scrutiny means winging it is likely not a wise choice.
> For £100 you can get a Beagleboard, which is faster, has more features
...Making it a crap choice if you'r writing code to target the Pi...
Wow. A meaningless argument. Good one.
> A meaningless argument.
An argument is not meaningless just because you haven't understood the argument.
Let me spell it out for you in excruciating detail.
If you are writing code for a specific board - such as, in this case, a Pi - then writing it on a different platform is of limited use; this is how you find out too late that your code relies on the extra capability of the platform you have used.
How is CE testing taking more than about 10 minutes for someone at RS or farnell suitably armed with a scope?
If this was Russia, people would be blaming Putin for running interference.
Because it does takes a lot more than about 10 minutes for someone at RS or farnell suitably armed with a scope!
In fact I don't think a scope is one of the pieces of testing equipment used.
They could have probably argued that as this is not 'Finished Equipment' but a component, they could cover CE requirements with a Certificate of Incorporation, rather than conformance?? (is a graphics card conformant to I.T radiated emissions levels if it is run outside of a nice metal grounded case)?
They argued that with BIS and the answer is "no".
Which begs the questions why they didn't ask BIS earlier and what professional advice they had which led them to believe CE compliance was not needed?
It looks like the foundation was being driven on wishful thinking until they got their distributors on board and it was only then that issues which should have already been resolved were identified.
The Arduino is CE approved, so it's no surprise the Raspberry Pi would also have to be for even stronger reasons (being marketed as a personal computer)
The even scarier bit is if they need FCC compliance to sell in the US. This is something that will take at least 50 days. Not something they can ignore either, since again popular boards like the Arduino AND the Beagle are FCC certified.
If Apple can get away with all sorts of misleading claims for its products around the world, why can't Raspberry?
The CE mark shows a product is compliant with certain EU legislation. e.g. toys carry a CE mark to demonstrate they do not pose a risk of strangulation / asphyxiation, protruding sharp bits, toxic materials, etc.
For the Pi it would be for electromagnetic emissions. i.e. to demonstrate emissions from the device don't interfere with other devices and that emissions from other devices don't interfere with the Pi.
I expect it involves a day or two in a lab and filing some paperwork. More arduous is cracking open every sealed product and sticking a CE sticker on it, e.g. on the USB port.
But CE means China Exempt (or should do) so what's the problem? Sod the EU. Just get some stickers printed.
Or buy them on e-bay like everyone else.
This is what happens when enthusiastic, starry-eyed academics deal with the real world of manufacturing supply chains, off-shoring, linguistic misunderstandings, transport logistics, worker rights, contract management, exclusivity agreements, marketing, etc, and the inevitable technical cock-up along the way. An advance appreciation of such problems could have been pre-empted, and a bit of real-world sense would suggest that the arrival of the devices should only have been publicized when there was 100% certainty that all delivery stages were fully proven.
I am of course still enthusiatically waiting for my slice of PI and grateful for their endevours, notwithstanding.
Nah, this is Element14/RS's cock up. Since they are supposed ot be doing the contract manufacturing, its them who forgot to mention (way back when the contract was set up) that this would be needed.
I remember the RPi being compared to the old BBC Micro, in its expected impact in education and hobbyist circles.
The BBC had terrible problems too on its hasty first release with a suspiciously named “Version 0.1” operating system.
I guess the lesson is not to rush these things out the door before they’re ready.
it might be possible to run W8 on it,
The irony is appreciated :-)
On a related note, suggesting on the Pi forums that it might be suitable for running W8 on ARM, will first make one subjected to a deluge of vitriolic posts from "rabid, Linux beardy-wierdies", foaming at the mouth with rage that you would dare mention the word Windows, and secondly, some rather more reasoned responses regarding the fact that the particular ARM cpu being used on the Pi is not suitable for running Windows and that the 256Mb of RAM would make it less than an attractive proposition in terms of performance, even if it could be run.
Personally, I think the hacking community being what it is, and I use the word hacking here in its original sense of the word, will find some way of porting Windows 8 to the ARM 6 instruction set and getting it working, just because of the challenge. One has only to look at somewhere like the XDA Developers forums to see what is possible by the right minds turning their efforts towards a common goal.
One thing I do know, is that the little Pi, causes a nerdgasm in all of us that grew up coding in the 80's and I am very much looking forward to receiving mine in due course.
A nerdgasm in all of us that grew up coding in the 80's and have never seen an ARM-based dev board before... :-)
There's no Windows driver for the GPU on the Pi either and that would have to be reverse engineered first. Broadcom only supplies a closed source binary driver for Linux.
It's the same problem for any Android port.
Reverse engineering a GPU is years of work even with all the good intentions of the XDA forums.
The real annoyance at the question isn't that "window"s is mentioned, it's that the answer is in the FAQ, the sticky at the top of the forum, in 50 other messages, AND you didn't even indicate that you would be asking about windows (and wasting everyone's time) in the title.
If you ask this question on the Raspi forums you are an idiot and deserve all the vitrol you get for not taking the time to read anything.
They can probably pay a specialist to look at the device, file a self assessment and then affix a stickers to the things. Maybe they can even ship without waiting for approval or acknowledgement I don't know.
You'd have thought that someone in RaspberryPi would've had a definitive answer about requiring CE certification *before* they tried selling them.
Their whole launch has been a fiasco - only the Brits (of which I'm one) can come up with a good idea and then really screw it up!
Now I know where they got the idea for the name "Raspberry"...
They did think of it but as the plan was to sell a small quantity to a bunch of developers, it wasn't needed ahead of the Educational' launch, but then the massive reaction moved the goalposts, and RS/Farnell havn't helped by agreeing to distribute it without CE certification then backtracking when the penpushers got involved.
You can avoid the need for FCC certification on development boards if they are are only intended for test and evaluation use in a laboratory environment. I assume that the requirements for CE exception are fairly similar. Most development boards I've used can legitimately take advantage of this - even if they are mainly targeted at hobbyists. However, once you start pitching the product for use by the general public you really need to have all the paperwork in order...
CE marking does not absolutely require any agency outside of the manufacturer to test the items for compliance.
You can self certify, basically just writing a declaration of conformance.
It is if you are challenged that you need to produce evidence of compliance (in the form of test results etc.)
For example all the powerline networking kit that are blatantly NOT EMC compliant all have CE certification (from manufacturer). Look at how they have been made to stop selling taht kit after complaints..Oh wait.
It is a useless system that does nothing to protect consumers. It is just a theatrical show.
Not at all, the PLT stuff is clearly in the interest of the *majority* of users and I applaud Ofcom for taking a difficult, but pragmatic, view.
Most PLT stuff only interferes with services below 30Mhz and looking at spectrum in that region the only ones who might be inconvenienced are amateur radio enthusiasts. There's no critical services at risk.
On the upside PLT allows many more users to connect to the Internet, where Wifi wouldn't work and where it's not feasible to rewire. Thus allowing PLC is a fair and democratic decision.
Now with devices running in the high Mhz such as Pi there's a clear risk of interfering with services in higher bands such as mobile phones, TV, DECT phones, GPS... In this case, if left unchecked, a few Raspberry Pi users could inconvenience a great many more users of these services. So the obligation to certify is again a good thing for the majority of consumers.
(I'm a Foundation license holder - AND - PLT user btw)
The problem is not PLT, it's cheap, badly built and non-conforming PLT.
Ofcom shouldn't take a pragmatic view - it's in no one's favour to permit shoddy kit to come onto market (and be distributed by BT).
PLT is not some 'great enabler', getting the great unwashed onto t'internets. There's no excuse for Ofcom to ignore their statutory duties; part of which are protecting the primary users of the spectrum, whether that's the military, or radio amateurs. In fact, the risk from an uncertified device like the pi is, in comparison, tiny!
Even if the pi turned out to be a spark-gap transmitter in disguise, you could fix the issue by putting it in a shielded box.
/Mine's the one with the two-part RAE in the pocket
Just the electromagnetic part of this can take some time, as the devices must be operated in each of their configurations with emissions and susceptibility to electromagnetic effects measured.. It's possible that if, sold to consumers, they may fail of approval, even though if, sold as components to be assembled into larger equipments, they might not need tests or marking at all.
Personally, as an Amateur Radio operator of some 53 years' standing (and an electromagnetic compatibility engineer over 30 years) I don't like to bring home radio interference generators. No one who uses shortwave would, if he thought about it.
One question that I have is how can people have ported a Linux software package to the Raspberry Pi when no hardware is yet available? The answer is presumably that around 500 Raspberry Pis have already been shipped to serious software developers around the world.
The problem is that Open Source development is supposed to be a level playing field. What I think should happen is that people be allowed to request that the Raspberry Pis that they have on order be shipped to them without magnetic jacks now on the understanding that they are sold as seen. I can not see how a large amount of educational software will be available in the middle of the year when the model A ships unless incomplete model Bs are shipped to developers now.
The point isn't that a large amount of educational software should be available when the model A ships. The point is for people to tinker with them. Why should software developers be prioritised over hardware hackers?
Perhaps you might want to look into the development VMs that people have been using, given that you're a serious software developer intending to create stuff for this platform.(cough).
I believe the answer to "how can people have ported Linux distributions to the Pi when no hardware is yet available" is that the Linux distribution community understand what the target environment is and are able to emulate it under QEMU for testing distributions.