yeah we thought about that, but ExpressCard means PCI-e. so we investigated how many SoC "embedded" CPUs have PCI-e and it's like about... 5. several from marvell, but they're the high-end "server" style ones, with power budgets wayyyy over the reasonable limit.
then we looked instead at PCI-e "PHY" chips. there's _one_: it's a PCI-e to USB converter, which says it all, really.
the other option is an FPGA: they're just... horrendously expensive. something like the Zynq-7030 would be able to do it in hardware using the on-board FPGA, but that's not out yet.
the only other option would be bit-banging! updating multiple GPIO pins to emulate a 33mhz ISA bus *shudder*.... yeah overall we just went "ok, forget it - let's just take over PCMCIA at the connector/header/socket level".
ephhh, you're right - i'm trying to figure out the best way to express it! there exist patches for forms (using YAML) in ikiwiki but the version that's on there is "stable", running multiple sites, so... i kinda settled for a manual form for now. if this went ballistic yeah i think i'd be in trouble, and would probably convert to a python wiki, which at least i can safely/happily program in.
i know, raenex - i've discussed this at some length, and have good justifiable reasons for sticking with what i'm doing. when writing "full articles" for publication, i do exactly that: use formal language. thanks for understanding.
thanks lennier1. yes we have someone in the same line of business, he's looking at placing an order for 1k units, precisely because the cost of development of what is effectively the major bit of the work - the CPU card - is so much lower. then, they can do a 2 to 4 layer board for the remaining bit, covering all the peripherals. that way, they've just got themselves a decent profit-margin back, even on low-volume production runs of their product, because the main CPU card is a "consumer-grade" off-the-shelf part.
I'm not a programmer or a hardware hacker. I don't know anything about soldering circuit boards. I'm just a guy who likes to surf the net, write stories, play games, hang out on-line, and so on. What is the availability of this $15 device going to mean for me?
I mean, at least (as far as) I know the Raspberry Pi is going to be producing fully-realized devices that I can buy, plug in a keyboard and monitor and Ethernet cable, and I'm done. It sounds like this project is just about building a circuit board. And while it's nice it will be 40% cheaper and three times as fast, I'd like to know what I could do with it if someone came up to me on the street and handed me one.
if they _literally_ handed you one on the street, you'd be able to plug in a USB-OTG-powered hub, then you could put in a keyboard and a USB ethernet, and also an HDMI monitor, and some headphones.
if they also included the "micro-header" that is also a planned product, you'd also be able to plug in an ethernet cable (without the USB internet dongle), and you'd not have to plug in that USB-OTG hub, you'd be able to put a standard hub on instead, and also power it from a 5V PSU, and you'd also be able to connect a standard (externally-powered) eSATA drive.
oh, and there's a MicroSD slot, so if the man-on-the-street had taken out its OS card before giving it to you, you'd be able to download ready-built standard GNU/Linux OS distros, shove them in and go.
so hell no, it's most definitely *not* about just "building a circuit board" - that's just the first step.
I blame poor tolerances in the computer's PCMCIA slot - too much clearance let the card shift enough that the pins didn't quite line up. This seemed to be a common problem back in PC cards were popular - we always ended up with a computer or two at the office with pins mangled so badly that we had to replace the PCMCIA slot module.
Apparently I'm not the only one to have this problem, since the howto guide on fixing the pins mentions the problem:
hey guess what, dude? if you manage to mangle it irretrievably, then thanks to the modular design you've only got one part to replace, not the entire device, eh? is that good apples or what?:)
jock, thanks for the clarification: i didn't explicitly mention it before, but it's worthwhile now, having made the mistake of not mentioning it earlier. the book that i read which describes the differences - patiently and in-depth - is Professor Yunus's book "Creating a World without Poverty". Professor Yunus is an Economics Professor, formed the Grameen Bank, pioneered "Micro Loans", and is the joint winner of the 2006 Nobel Peace Prize. his book is just awe-inspiring, but crucially it describes why Ltd Company, not-for-profit and charity is not as appropriate a vehicle as a CIC. ok i leave it at that:)
I imagine the reason that they reused the PCMCIA design is for reuse of the tooling for the case and also the durability of the connectors. The PCMCIA connectors have durability ratings of 10K insertions. Many card edge connectors have only a durability of 100-200 insertions. The simply made fascia plate keeps these new cards from being inserted into legacy sockets. The cpu card might be swapped from a laptop to a desktop, set-top-box, car PC, cluster rack, etc etc. You could make devices with a simple cover plate to keep the cpu module from being easily ejected if you wish. Devices such as laptops, set-top-boxes, etc might be easily upgraded to a newer or more powerful cpu or more RAM by simply swapping the cpu module.
*snort*. i wish i'd written what you'd written, it's spot-on:) can i borrow what you wrote, put it (attributed) somewhere on the wiki? seriously:)
I work in a company where some of our products are basically Full-HD TFT displays with integrated ARM-based computers (glorified nettop components) running a company-internal Linux distro.
Having one of these to replace/upgrade their computer like you'd switch the optical drive in a business laptop would certainly cut down costs.
oo - lennier1, i want to talk to you:) i'd love to know how much these cost (the product your company has) because esp if it's gnu/linux based and has a decent amount of RAM i'm sure it would be desirable by many Software (Libre) Developers. please could you email me, to talk more? thanks!
michael, hang on dude! you're asking me to try to run before we walk, ok?
if you're not familiar with the way this stuff works... ok: the only thing we can do - *right now* - is set an "upper bound" based on known costs.
so *right now*, and *at this stage*, which is "alpha stage", we can say that, based on the fixed NREs of $USD 2,000, the more people that place preorder pledges, the more that we can subdivide those costs across the total number of people.
*right now* we have approximately 30 pre-order committments which, if they were all "alpha" committments, that $2000 would mean $75 per person.
what we _do_ know is the "upper bound" as well - based on the $15 figure from the SoC manufacturer, which is $15.
in between those two values the costs are a *direct* relation to the size of the order.
i've explained this - clearly - on the preorders page. i hope. i think. if it's not clear, please do say so.
but if you absolutely want a fixed price i'm sorry, that's just simply not possible. ok, it is, but you'll have to place cash on the table to get it.
do you have cash up-front that you can put on the table? because if so, then that's absolutely fine: i can then go to the Factory and say "we have a cash order for NNNNN units, please can you quote an *exact* figure for us?" they will then ask for a deposit - which you will have to pay - before they proceed with the work.
do you see how that is a completely different kind of deal from the one that we're doing? what we're doing is to *collaborate* with the Factory. we've done a deal which summarises as "we won't charge you for software engineer time if you don't charge us for hardware engineer time", and ensuring that the EOMA-PCMCIA CPU card can fulfil both *their* customer requirements as well as our mass-volume sales opportunities *and* fulfil the requirements of Software (Libre) Developers...... you see how radically different that is from the "standard" business deal of "pay a factory in china to get it done (y'all), add that to the product retail cost and multiply up by some margin equals profit"?
it's... _complicated_, michael:) i'm doing my best to keep it simple, but also following the lessons learned from other related products like the PixelQI screen (they borrowed factory time at christmas of a taiwanese LCD manufacturer) and Goldelico's GTA04 product, designed by Dr Schaller. (Dr Schaller deliberately picked components that are available right down to Qty 1).
ok i'll leave it at that, hope that helps clarify.
p.s. it turns out that we do have someone who is willing to place an order for 1,000 units. i've asked - and will keep pressing - the factory for a quote based on those quantities. when it's available, i'll update the pages accordingly.
Sadly the corps have figured out "designed for the dump" gives them their biggest profits hence why everything is so flimsy and easily broken now. Personally I wish the FOSS guys all the luck in the world, i'd love a cell phone or laptop where parts were as easy to get and interchange as your average desktop but I doubt the corps would ever let that happen, it'd cost them too much profit.
we've spoken to several of them, already. the profit margins anticipated by the large companies are, exactly as you surmise, too low. thus, we have no competition. question for you: tell me where you can get a GPL-compliant 1ghz+ ARM Cortex laptop with 1gb or 2gb of RAM, anywhere in the shops with a 1280x800 LCD or better for $150 and i will quit working on this and go buy it.
ok, that's just one point covered - i've said quite enough on this discussion already, i'll leave it for a while, ok?:)
allo mr hairyfeet - good question. much of the reasoning _is_ "for the hell of it", but it is definitely more than that. what i haven't mentioned is that my associates have contacts with very very large PRC manufacturers. we really really are in a position to go, stage by stage, from a prototyped system all the way to massive-scale production that would dwarf even Dell (because companies such as IBM and HP actually use one of the PRC manufacturers that we are in communication with). now this is *not* "bragging" - i'm mentioning it because you asked, even though there is the risk of many people reading this to go "we don't believe you, who the hell do you think you are??".
if we were planning this to be a "tiny niche" product, we would have picked SO-DIMM form-factor, not re-use of legacy PCMCIA. SO-DIMM form-factor is perfect for R&D purposes: look at the pricing and functionality of modules from colibri, or directinsight and so on. they're expensive, and they don't scale into mass-production (sadly).
but, the other thing that you're missing is that when you match up low-cost china-based factories with the kinds of retail pricing expected by Software (Libre) Developers and so on, using pricing levels set as a precedent by things precisely like the BeagleBoard, Arduino or better the Leafpad Maple, you actually make quite good profit margins.
i'm not sure if you're aware of this but the profit margins on mass-volume products, especially at the consumer-grade level, are pretty damn small. like... a few dollars small.
ironically, therefore, the profit margins of medium-volume sales would be about the same order as the mass-volume sales! if we get this right, we can transition from one to the other, *without* having "the usual" massive NREs associated with having to do *complete* product redesign each and every time (thanks to the modular design).
so it's a complex strategy, based on bootstrapping up from zero NREs, which, itself, is far from being silly or impossible, is again _part_ of the strategy of keeping the overall costs down.
been doing that for 2 years, dude. consider yourself lucky that i'm a software (libre) developer, not a marketing droid. mwahahahah all your base are belong to us, we are Ltd Company pathological liars who will do annnnything to get your moneyyyy mwahahah. http://www.thecorporation.com/
yes. now, what's really cool about the idea of doing an open hardware CPU is that when it succeeds, and when the rhombus-tech initiative succeeds, there will be a ton of compatible motherboards sold as mass-volume products that end-users, who don't want to go to the trouble of reinventing the laptop wheel or the tablet wheel can just go "yes! i'll have one of those, and now i have a laptop with an open hardware CPU" and even "yes! i have a smartphone with an open hardware CPU".
try that with anything other than a modular architecture and see how far it gets you.
ah, but then can you take the CPU out of the phone and put it into a low-cost beowulf supercomputer cluster? yes, seriously: one of the options that's possible with these little CPU cards, because they have SATA-II interfaces (proper ones) and also use such little power, is to plug them into a massive rack, 1gb RAM, 1ghz CPU speed, NEON instruction set per CPU, hell you'd have an ultra-low-power supercomputer in no time! if only bloody ARM would release information about how to use the GPU on the MALI 400 MP for scientific purposes we'd be laughing.
if you tried to force it in, you would mechanically damage the laptop and/or the card, and once you'd done that, the chances are that you'd blow up the card and/or the laptop as well.
Or you couyld just use the laptop. Am i missing something here?
you're missing something:) the design concept is that the EOMA-PCMCIA CPU card *is* the laptop... but only when the modular CPU card is plugged into an EOMA-PCMCIA-compliant laptop Motherboard that's *designed* to take these CPU cards. see example motherboards here: http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA#Example_Motherboards
to have an x86 CPU in a laptop already (cost of $300+) and to then put in an extra low-cost CPU card that does pretty much the same job as far as 98% of computer users are concerned, well... that would just be silly. why not just have a modular mass-volume laptop plus CPU card that can retail for about $95, eh?:)
... there's a website that I can order one from at that price, which will deliver with 7 days.
pete: done. we have a deal:)
please feel free to fill in a preorder which says exactly this. you want "Stage: stable". if you think it would be better to have a stage "7 day delivery at the stated price" then please feel free to say so, but bear in mind that it may be better for you to wait until the product's in Hypermarket Retail Stores and you can buy them off-the-shelf (literally). of course, you miss out on all the fun that way...:)
good man! feel free to fill in the preorder form http://rhombus-tech.net/allwinner_a10/orders/ i'm a bit reluctant to do it on your behalf [aitch tee tee pee slashdot dot org slash tilde hatta]
please do bear in mind that in the early stage we're *not* going to sell completely untested cards in mass-volume right away, that would be foolish. we're going to follow the process that Dr Schaller has been doing on the development of the GTA04 - http://wiki.openmoko.org/wiki/GTA04_revisions as have various other projects, OpenPandora included.
so, early alpha boards go out to people prepared to take a risk, but who have the money spare (under $100, gosh, wow, break the bank why not) to consider "what the heck, this is cool, let's support this initiative" but at the same time have some expertise in embedded GNU/Linux development, and they might actually get something that works perfectly first time, and they're the ones that got it, before anyone else.
beta boards go out to people who want something that, hardware-wise, is pretty much guaranteed to work 100%, but maybe the software's not all there, and they might have to (gosh) get involved and help write it.
stable boards go out to people who really would "just like something that works, thank you, where's the debian distro image, where's the instructions for putting everything onto an sdcard, heck, where can i buy a pre-loaded MicroSD card so i don't have to do that, even".
so it's a known trade-off: the principles of Software (Libre) Development as applied to hardware: release early, release often. exactly the sort of thing that you never normally see in the development of hardware products, and i think it's pretty damn cool to be able to witness and be part of something that *isn't* GPL-violating. at bloody last.
I think they're just using the physical PCMCIA connector, not the PCMCIA pin-out. I think it's so that the entire thing can be plugged in to a variety of hardware devices: small form-factor computers, TVs, tablets, whatever. There are benefits to that approach. Wouldn't it have been nice, for example, if you could have upgraded your original iPad by simply ejecting the motherboard and inserting a new one? There might not have been any reason to replace the screen and battery. This sort of modular approach resolves that. And it opens up opportunities for hardware manufacturers if they know they can get a whole computer in a known form-factor. It would relieve them of an otherwise huge part of the product design.
exactly! now, why did you say this as an anonymous coward?:) the problem with the above is that what end-users would love is exactly what consumerism hates! planned obsolescence is what it's called, i believe. we don't like that sort of thing round here, y'all:) hence the initiative is being done under the umbrella of a Community Interest Company, because it removes the absolute requirement to maximise profits over-and-above-all-else [CICs just have to not make a loss, and there are *no* Shareholders, and no dividends to pay out]. thus, there is no driving force, for example, to justify planned obsolescence, nor is there a justification to remove the dual motherboard design, because to do so would result in larger profits through the argument that "when it breaks people will spend more money with us, buying a whole new one, profit is higher".
the other reason for the modular approach is that there are now restrictions on air-shipment of explosives (lithium batteries being an explosion hazard). so by having a modular design where there is a battery compartment that can take AA or AAA batteries, and you can upgrade later to a lithium pack, Mass-Volume Hypermarket Retail Stores can do "Just in Time" ordering of the main parts of the device(s), get them shipped in by Airfreight, as well as stock up on Sea-freighted modular battery packs in large volumes which, if they truly run out or don't arrive in time, doesn't matter because the sales staff can direct people to buy the bit of plastic that holds standard AA/AAA batteries.
we've thought this through from lots of different angles. really:)
yaa, who said anything about competing?:) feel free to buy a 700mhz ARM11 unit for $25 when it's available. we're going in incremental stages. if you've seen what happened to projects like the OpenPandora, the OpenMoko and so on, you'll appreciate why. http://rhombus-tech.net//faq/#index2h2
"mass-volume" is code usually for 100k+ pricing. the pricing quote for the raspberry pi is equally based on mass-volume (100k) pricing. the pricing quote for the raspberry pi equally excludes profit, shipping, tax, packaging, delivery, handling, tax, customs duty, tax on customs duty, agent shipping handling fees, tax on agent shipping handling fees, customs duty on tax on agent shipping handling fees and so on. whilst that sounds like a joke it's not: each and every one of those costs _does_ actually exist.
now, in the case of the raspberry pi, because they are a not-for-profit foundation, they are *not allowed* to make a profit (definition of "not for profit"), thus there is no room for expansion or for investment. as the front page explains, and i think i put it on the FAQ as well, sales of products for this initiative is being done via a "Community Interest Company", thus, when we say "it's possible that all profits can be fed directly back into R&D for further products to the direct benefit of Software (Libre) Developers" or "a decision can be made to spend an entire years profits on buying modules and giving them away to charity or to Software (Libre) Developers for strategic purposes", you know that we really MEAN that.
if Rhombus Tech was a "Ltd Company" and tried that kind of "stunt", the shareholders would be absolutely screaming blue murder and would want the Directors' heads on a platter. CICs are pretty misunderstood and under-rated, but they're a much better vehicle for what is being planned.
http://rhombus-tech.net//faq/#index4h2 - re-use of *existing* connectors, housings and assemblies keeps the price right down. yes you're absolutely right: expecting a complete new design of connector to be reasonably affordable is impossible.
the whole initiative is based around leap-frogging over the normal barriers to entry for products. use Software (Libre) Developers for the software engineering. use off-the-shelf parts as much as possible. do a deal with the factory ["we won't charge you for software engineer time if you won't charge us for hardware engineer time"]. use pre-existing casework designs from China-based Industrial Flea Markets (don't get the wrong idea, here - these Markets are the size of football pitches and 7 stories high!) and so on.
That's because the law can comprehend the difference between faiths and fanboyism and making a joke on a mandatory census. This is where the law is sensible.
ahh, and i thought that it was because blair managed to get himself possessed by rakshasas because his wife was arseing about with ouija boards. that it was sith all along explains everything.
Slashdot Mirror
hi pedrop,
yeah we thought about that, but ExpressCard means PCI-e. so we investigated how many SoC "embedded" CPUs have PCI-e and it's like about... 5. several from marvell, but they're the high-end "server" style ones, with power budgets wayyyy over the reasonable limit.
then we looked instead at PCI-e "PHY" chips. there's _one_: it's a PCI-e to USB converter, which says it all, really.
the other option is an FPGA: they're just... horrendously expensive. something like the Zynq-7030 would be able to do it in hardware using the on-board FPGA, but that's not out yet.
the only other option would be bit-banging! updating multiple GPIO pins to emulate a 33mhz ISA bus *shudder*. ... yeah overall we just went "ok, forget it - let's just take over PCMCIA at the connector/header/socket level".
ephhh, you're right - i'm trying to figure out the best way to express it! there exist patches for forms (using YAML) in ikiwiki but the version that's on there is "stable", running multiple sites, so... i kinda settled for a manual form for now. if this went ballistic yeah i think i'd be in trouble, and would probably convert to a python wiki, which at least i can safely/happily program in.
i know, raenex - i've discussed this at some length, and have good justifiable reasons for sticking with what i'm doing. when writing "full articles" for publication, i do exactly that: use formal language. thanks for understanding.
thanks lennier1. yes we have someone in the same line of business, he's looking at placing an order for 1k units, precisely because the cost of development of what is effectively the major bit of the work - the CPU card - is so much lower. then, they can do a 2 to 4 layer board for the remaining bit, covering all the peripherals. that way, they've just got themselves a decent profit-margin back, even on low-volume production runs of their product, because the main CPU card is a "consumer-grade" off-the-shelf part.
I'm not a programmer or a hardware hacker. I don't know anything about soldering circuit boards. I'm just a guy who likes to surf the net, write stories, play games, hang out on-line, and so on. What is the availability of this $15 device going to mean for me?
I mean, at least (as far as) I know the Raspberry Pi is going to be producing fully-realized devices that I can buy, plug in a keyboard and monitor and Ethernet cable, and I'm done. It sounds like this project is just about building a circuit board. And while it's nice it will be 40% cheaper and three times as fast, I'd like to know what I could do with it if someone came up to me on the street and handed me one.
if they _literally_ handed you one on the street, you'd be able to plug in a USB-OTG-powered hub, then you could put in a keyboard and a USB ethernet, and also an HDMI monitor, and some headphones.
if they also included the "micro-header" that is also a planned product, you'd also be able to plug in an ethernet cable (without the USB internet dongle), and you'd not have to plug in that USB-OTG hub, you'd be able to put a standard hub on instead, and also power it from a 5V PSU, and you'd also be able to connect a standard (externally-powered) eSATA drive.
oh, and there's a MicroSD slot, so if the man-on-the-street had taken out its OS card before giving it to you, you'd be able to download ready-built standard GNU/Linux OS distros, shove them in and go.
so hell no, it's most definitely *not* about just "building a circuit board" - that's just the first step.
I blame poor tolerances in the computer's PCMCIA slot - too much clearance let the card shift enough that the pins didn't quite line up. This seemed to be a common problem back in PC cards were popular - we always ended up with a computer or two at the office with pins mangled so badly that we had to replace the PCMCIA slot module.
Apparently I'm not the only one to have this problem, since the howto guide on fixing the pins mentions the problem:
http://www.ehow.com/how_8690286_fix-bent-cardbus-pins.html
hey guess what, dude? if you manage to mangle it irretrievably, then thanks to the modular design you've only got one part to replace, not the entire device, eh? is that good apples or what? :)
jock, thanks for the clarification: i didn't explicitly mention it before, but it's worthwhile now, having made the mistake of not mentioning it earlier. the book that i read which describes the differences - patiently and in-depth - is Professor Yunus's book "Creating a World without Poverty". Professor Yunus is an Economics Professor, formed the Grameen Bank, pioneered "Micro Loans", and is the joint winner of the 2006 Nobel Peace Prize. his book is just awe-inspiring, but crucially it describes why Ltd Company, not-for-profit and charity is not as appropriate a vehicle as a CIC. ok i leave it at that :)
I imagine the reason that they reused the PCMCIA design is for reuse of the tooling for the case and also the durability of the connectors. The PCMCIA connectors have durability ratings of 10K insertions. Many card edge connectors have only a durability of 100-200 insertions. The simply made fascia plate keeps these new cards from being inserted into legacy sockets. The cpu card might be swapped from a laptop to a desktop, set-top-box, car PC, cluster rack, etc etc. You could make devices with a simple cover plate to keep the cpu module from being easily ejected if you wish. Devices such as laptops, set-top-boxes, etc might be easily upgraded to a newer or more powerful cpu or more RAM by simply swapping the cpu module.
*snort*. i wish i'd written what you'd written, it's spot-on :) can i borrow what you wrote, put it (attributed) somewhere on the wiki? seriously :)
I work in a company where some of our products are basically Full-HD TFT displays with integrated ARM-based computers (glorified nettop components) running a company-internal Linux distro.
Having one of these to replace/upgrade their computer like you'd switch the optical drive in a business laptop would certainly cut down costs.
oo - lennier1, i want to talk to you :) i'd love to know how much these cost (the product your company has) because esp if it's gnu/linux based and has a decent amount of RAM i'm sure it would be desirable by many Software (Libre) Developers. please could you email me, to talk more? thanks!
ahh, you've dealt with SoC manufacturers rushing things out the door before, i see? :)
michael, hang on dude! you're asking me to try to run before we walk, ok?
if you're not familiar with the way this stuff works... ok: the only thing we can do - *right now* - is set an "upper bound" based on known costs.
so *right now*, and *at this stage*, which is "alpha stage", we can say that, based on the fixed NREs of $USD 2,000, the more people that place preorder pledges, the more that we can subdivide those costs across the total number of people.
*right now* we have approximately 30 pre-order committments which, if they were all "alpha" committments, that $2000 would mean $75 per person.
what we _do_ know is the "upper bound" as well - based on the $15 figure from the SoC manufacturer, which is $15.
in between those two values the costs are a *direct* relation to the size of the order.
i've explained this - clearly - on the preorders page. i hope. i think. if it's not clear, please do say so.
but if you absolutely want a fixed price i'm sorry, that's just simply not possible. ok, it is, but you'll have to place cash on the table to get it.
do you have cash up-front that you can put on the table? because if so, then that's absolutely fine: i can then go to the Factory and say "we have a cash order for NNNNN units, please can you quote an *exact* figure for us?" they will then ask for a deposit - which you will have to pay - before they proceed with the work.
do you see how that is a completely different kind of deal from the one that we're doing? what we're doing is to *collaborate* with the Factory. we've done a deal which summarises as "we won't charge you for software engineer time if you don't charge us for hardware engineer time", and ensuring that the EOMA-PCMCIA CPU card can fulfil both *their* customer requirements as well as our mass-volume sales opportunities *and* fulfil the requirements of Software (Libre) Developers... ... you see how radically different that is from the "standard" business deal of "pay a factory in china to get it done (y'all), add that to the product retail cost and multiply up by some margin equals profit"?
it's... _complicated_, michael :) i'm doing my best to keep it simple, but also following the lessons learned from other related products like the PixelQI screen (they borrowed factory time at christmas of a taiwanese LCD manufacturer) and Goldelico's GTA04 product, designed by Dr Schaller. (Dr Schaller deliberately picked components that are available right down to Qty 1).
ok i'll leave it at that, hope that helps clarify.
p.s. it turns out that we do have someone who is willing to place an order for 1,000 units. i've asked - and will keep pressing - the factory for a quote based on those quantities. when it's available, i'll update the pages accordingly.
Sadly the corps have figured out "designed for the dump" gives them their biggest profits hence why everything is so flimsy and easily broken now. Personally I wish the FOSS guys all the luck in the world, i'd love a cell phone or laptop where parts were as easy to get and interchange as your average desktop but I doubt the corps would ever let that happen, it'd cost them too much profit.
we've spoken to several of them, already. the profit margins anticipated by the large companies are, exactly as you surmise, too low. thus, we have no competition. question for you: tell me where you can get a GPL-compliant 1ghz+ ARM Cortex laptop with 1gb or 2gb of RAM, anywhere in the shops with a 1280x800 LCD or better for $150 and i will quit working on this and go buy it.
ok, that's just one point covered - i've said quite enough on this discussion already, i'll leave it for a while, ok? :)
allo mr hairyfeet - good question. much of the reasoning _is_ "for the hell of it", but it is definitely more than that. what i haven't mentioned is that my associates have contacts with very very large PRC manufacturers. we really really are in a position to go, stage by stage, from a prototyped system all the way to massive-scale production that would dwarf even Dell (because companies such as IBM and HP actually use one of the PRC manufacturers that we are in communication with). now this is *not* "bragging" - i'm mentioning it because you asked, even though there is the risk of many people reading this to go "we don't believe you, who the hell do you think you are??".
if we were planning this to be a "tiny niche" product, we would have picked SO-DIMM form-factor, not re-use of legacy PCMCIA. SO-DIMM form-factor is perfect for R&D purposes: look at the pricing and functionality of modules from colibri, or directinsight and so on. they're expensive, and they don't scale into mass-production (sadly).
but, the other thing that you're missing is that when you match up low-cost china-based factories with the kinds of retail pricing expected by Software (Libre) Developers and so on, using pricing levels set as a precedent by things precisely like the BeagleBoard, Arduino or better the Leafpad Maple, you actually make quite good profit margins.
i'm not sure if you're aware of this but the profit margins on mass-volume products, especially at the consumer-grade level, are pretty damn small. like... a few dollars small.
ironically, therefore, the profit margins of medium-volume sales would be about the same order as the mass-volume sales! if we get this right, we can transition from one to the other, *without* having "the usual" massive NREs associated with having to do *complete* product redesign each and every time (thanks to the modular design).
so it's a complex strategy, based on bootstrapping up from zero NREs, which, itself, is far from being silly or impossible, is again _part_ of the strategy of keeping the overall costs down.
been doing that for 2 years, dude. consider yourself lucky that i'm a software (libre) developer, not a marketing droid. mwahahahah all your base are belong to us, we are Ltd Company pathological liars who will do annnnything to get your moneyyyy mwahahah. http://www.thecorporation.com/
yes. now, what's really cool about the idea of doing an open hardware CPU is that when it succeeds, and when the rhombus-tech initiative succeeds, there will be a ton of compatible motherboards sold as mass-volume products that end-users, who don't want to go to the trouble of reinventing the laptop wheel or the tablet wheel can just go "yes! i'll have one of those, and now i have a laptop with an open hardware CPU" and even "yes! i have a smartphone with an open hardware CPU".
try that with anything other than a modular architecture and see how far it gets you.
ah, but then can you take the CPU out of the phone and put it into a low-cost beowulf supercomputer cluster? yes, seriously: one of the options that's possible with these little CPU cards, because they have SATA-II interfaces (proper ones) and also use such little power, is to plug them into a massive rack, 1gb RAM, 1ghz CPU speed, NEON instruction set per CPU, hell you'd have an ultra-low-power supercomputer in no time! if only bloody ARM would release information about how to use the GPU on the MALI 400 MP for scientific purposes we'd be laughing.
Great, so this low cost computer can be plugged into the PCMCIA slot of a laptop.
ah NO! :) the mechanical design prevents insertion of EOMA-PCMCIA CPU cards into legacy PCMCIA slots:
http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA#Deliberate_Mechanical_Non-interoperability
if you tried to force it in, you would mechanically damage the laptop and/or the card, and once you'd done that, the chances are that you'd blow up the card and/or the laptop as well.
Or you couyld just use the laptop. Am i missing something here?
you're missing something :) the design concept is that the EOMA-PCMCIA CPU card *is* the laptop... but only when the modular CPU card is plugged into an EOMA-PCMCIA-compliant laptop Motherboard that's *designed* to take these CPU cards. see example motherboards here: http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA#Example_Motherboards
to have an x86 CPU in a laptop already (cost of $300+) and to then put in an extra low-cost CPU card that does pretty much the same job as far as 98% of computer users are concerned, well... that would just be silly. why not just have a modular mass-volume laptop plus CPU card that can retail for about $95, eh? :)
pete: done. we have a deal :)
please feel free to fill in a preorder which says exactly this. you want "Stage: stable". if you think it would be better to have a stage "7 day delivery at the stated price" then please feel free to say so, but bear in mind that it may be better for you to wait until the product's in Hypermarket Retail Stores and you can buy them off-the-shelf (literally). of course, you miss out on all the fun that way... :)
good man! feel free to fill in the preorder form http://rhombus-tech.net/allwinner_a10/orders/ i'm a bit reluctant to do it on your behalf [aitch tee tee pee slashdot dot org slash tilde hatta]
please do bear in mind that in the early stage we're *not* going to sell completely untested cards in mass-volume right away, that would be foolish. we're going to follow the process that Dr Schaller has been doing on the development of the GTA04 - http://wiki.openmoko.org/wiki/GTA04_revisions as have various other projects, OpenPandora included.
so, early alpha boards go out to people prepared to take a risk, but who have the money spare (under $100, gosh, wow, break the bank why not) to consider "what the heck, this is cool, let's support this initiative" but at the same time have some expertise in embedded GNU/Linux development, and they might actually get something that works perfectly first time, and they're the ones that got it, before anyone else.
beta boards go out to people who want something that, hardware-wise, is pretty much guaranteed to work 100%, but maybe the software's not all there, and they might have to (gosh) get involved and help write it.
stable boards go out to people who really would "just like something that works, thank you, where's the debian distro image, where's the instructions for putting everything onto an sdcard, heck, where can i buy a pre-loaded MicroSD card so i don't have to do that, even".
so it's a known trade-off: the principles of Software (Libre) Development as applied to hardware: release early, release often. exactly the sort of thing that you never normally see in the development of hardware products, and i think it's pretty damn cool to be able to witness and be part of something that *isn't* GPL-violating. at bloody last.
I think they're just using the physical PCMCIA connector, not the PCMCIA pin-out. I think it's so that the entire thing can be plugged in to a variety of hardware devices: small form-factor computers, TVs, tablets, whatever. There are benefits to that approach. Wouldn't it have been nice, for example, if you could have upgraded your original iPad by simply ejecting the motherboard and inserting a new one? There might not have been any reason to replace the screen and battery. This sort of modular approach resolves that. And it opens up opportunities for hardware manufacturers if they know they can get a whole computer in a known form-factor. It would relieve them of an otherwise huge part of the product design.
exactly! now, why did you say this as an anonymous coward? :) the problem with the above is that what end-users would love is exactly what consumerism hates! planned obsolescence is what it's called, i believe. we don't like that sort of thing round here, y'all :) hence the initiative is being done under the umbrella of a Community Interest Company, because it removes the absolute requirement to maximise profits over-and-above-all-else [CICs just have to not make a loss, and there are *no* Shareholders, and no dividends to pay out]. thus, there is no driving force, for example, to justify planned obsolescence, nor is there a justification to remove the dual motherboard design, because to do so would result in larger profits through the argument that "when it breaks people will spend more money with us, buying a whole new one, profit is higher".
the other reason for the modular approach is that there are now restrictions on air-shipment of explosives (lithium batteries being an explosion hazard). so by having a modular design where there is a battery compartment that can take AA or AAA batteries, and you can upgrade later to a lithium pack, Mass-Volume Hypermarket Retail Stores can do "Just in Time" ordering of the main parts of the device(s), get them shipped in by Airfreight, as well as stock up on Sea-freighted modular battery packs in large volumes which, if they truly run out or don't arrive in time, doesn't matter because the sales staff can direct people to buy the bit of plastic that holds standard AA/AAA batteries.
we've thought this through from lots of different angles. really :)
yaa, who said anything about competing? :) feel free to buy a 700mhz ARM11 unit for $25 when it's available. we're going in incremental stages. if you've seen what happened to projects like the OpenPandora, the OpenMoko and so on, you'll appreciate why. http://rhombus-tech.net//faq/#index2h2
"mass-volume" is code usually for 100k+ pricing. the pricing quote for the raspberry pi is equally based on mass-volume (100k) pricing. the pricing quote for the raspberry pi equally excludes profit, shipping, tax, packaging, delivery, handling, tax, customs duty, tax on customs duty, agent shipping handling fees, tax on agent shipping handling fees, customs duty on tax on agent shipping handling fees and so on. whilst that sounds like a joke it's not: each and every one of those costs _does_ actually exist.
now, in the case of the raspberry pi, because they are a not-for-profit foundation, they are *not allowed* to make a profit (definition of "not for profit"), thus there is no room for expansion or for investment. as the front page explains, and i think i put it on the FAQ as well, sales of products for this initiative is being done via a "Community Interest Company", thus, when we say "it's possible that all profits can be fed directly back into R&D for further products to the direct benefit of Software (Libre) Developers" or "a decision can be made to spend an entire years profits on buying modules and giving them away to charity or to Software (Libre) Developers for strategic purposes", you know that we really MEAN that.
if Rhombus Tech was a "Ltd Company" and tried that kind of "stunt", the shareholders would be absolutely screaming blue murder and would want the Directors' heads on a platter. CICs are pretty misunderstood and under-rated, but they're a much better vehicle for what is being planned.
http://rhombus-tech.net//faq/#index4h2 - re-use of *existing* connectors, housings and assemblies keeps the price right down. yes you're absolutely right: expecting a complete new design of connector to be reasonably affordable is impossible.
the whole initiative is based around leap-frogging over the normal barriers to entry for products. use Software (Libre) Developers for the software engineering. use off-the-shelf parts as much as possible. do a deal with the factory ["we won't charge you for software engineer time if you won't charge us for hardware engineer time"]. use pre-existing casework designs from China-based Industrial Flea Markets (don't get the wrong idea, here - these Markets are the size of football pitches and 7 stories high!) and so on.
That's because the law can comprehend the difference between faiths and fanboyism and making a joke on a mandatory census. This is where the law is sensible.
ah yes, that would explain why the australian government implied that it would break privacy laws in order to prosecute people. http://en.wikipedia.org/wiki/Jedi_census_phenomenon#Australia
ahh, and i thought that it was because blair managed to get himself possessed by rakshasas because his wife was arseing about with ouija boards. that it was sith all along explains everything.