-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client? I expect they are stuffed when they can't even mention eoma-68!
It's just, I've been trying to work on my Operation: Marketing (Op:Mark) but with this sad news I've not sure which direction to go. I had been plaining to write marketing material for a device.
I have a huge amount to evolve and currently Op:Mark is my mess also my progress is dreadfully slow. I wanted to make more progress on it before I mentioned it to you but arr heck here it goes: http://rhombus-tech.net/op:marketing
On Mon, May 19, 2014 at 4:40 PM, Alexander Stephen Thomas Ross maillist_arm-netbook@aross.me wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client? I expect they are stuffed when they can't even mention eoma-68!
well the problem is that they didn't keep us informed of numbers, as they came in. we therefore lost face with our large client (the factory in china) because we had to give them excuses instead of an order that was promised.
four months in, after asking many many times we find that they had 10% of expected orders, which they could have informed us of within the first 4 weeks not after 4 months.
after apologising deeply to our embarrassed factory, whom we had gone to extraordinary lengths to arrange an order of only 2500 units (they normally deal with minimum 10 to 50k runs, their machines have a capacity of 20 million units a week so this is a huge favour they were doing us), we found an alternative factory, on very short notice, who could do 500 units of the PCBs.
... three weeks later, complete silence..
my business partner has therefore permanently crossed them off of our client list - he was very pissed. i don't have control over what decisions he makes.
It's just, I've been trying to work on my Operation: Marketing (Op:Mark) but with this sad news I've not sure which direction to go.
if we receive orders, we can place the order with the factory. everything is set up for that.
thanks alexander - i'll take a look.
l.
On Monday, May 19, 2014 17:21:13 Luke Kenneth Casson Leighton wrote:
On Mon, May 19, 2014 at 4:40 PM, Alexander Stephen Thomas Ross
maillist_arm-netbook@aross.me wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client? I expect they are stuffed when they can't even mention eoma-68!
well the problem is that they didn't keep us informed of numbers, as
This is not accurate, but I'm not going to get into details of what was a series of private conversations, most of which Luke was not even involved with, in public. I like to think I have slightly better judgment than that. :/
On Tue, May 20, 2014 at 5:09 PM, Aaron J. Seigo aseigo@kde.org wrote:
On Monday, May 19, 2014 17:21:13 Luke Kenneth Casson Leighton wrote:
On Mon, May 19, 2014 at 4:40 PM, Alexander Stephen Thomas Ross
maillist_arm-netbook@aross.me wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client? I expect they are stuffed when they can't even mention eoma-68!
well the problem is that they didn't keep us informed of numbers, as
This is not accurate,
ok. let me be more specific. you did not keep us informed of numbers within a reasonable time-frame. delays of several weeks before responding with a status update is, i believe, reasonable to summarise as "not keeping us informed of numbers".
if you had advised us immediately that you only had 200 to 250 units after one month when we requested after one month that you inform us of numbers, we would have had time to do something.
but you did not respond to my associate's request.
he had to ask several times. each time you did not respond.
eventually after several months we received a response. by then it was too late.
but I'm not going to get into details of what was a series of private conversations,
which my associate kept me appraised of, as you may not be aware that i speak with him nearly every day. aaron: if you genuinely had kept my associate informed, providing him the status updates he requested when he requested them, we would not be having this conversation. we would have been able to do something.
l.
On Tuesday, May 20, 2014 18:10:18 Luke Kenneth Casson Leighton wrote:
ok. let me be more specific.
*sigh* Again, not accurate in the least, but I'm not having this conversation here.
Please, Luke, respect that. I'm only responding at all because I'm not OK with such misinformation about my person in public. I am not going to drag anyone else (e.g. you) through the mud (deserving or not) or get into a useless argument over the details of something that can not at this point be undone. I'd appreciate a similar amount of respect shown from you in the name of civility and, well, professionalism.
On Tue, May 20, 2014 at 7:01 PM, Aaron J. Seigo aseigo@kde.org wrote:
On Tuesday, May 20, 2014 18:10:18 Luke Kenneth Casson Leighton wrote:
ok. let me be more specific.
*sigh* Again, not accurate in the least,
i think the only thing i can say at this point is this: that you can repeat that belief helps to explain why my associate made the decision to blacklist you.
l.
Please, Luke, respect
<jaw drops>
http://liliputing.com/2014/05/novena-open-laptop-raises-700-thousand-crowdfu...
Its even got a deluxe rod end bearing to prop its door open :(
The stereo speakers are mounted side by side (at last) :( :(
</jaw drops>
Acquired four 3D printers in the last few weeks. One of them is the Formlabs laser curing device with 25 micron resolution and the other 3 are fdm printers. Driving it with openscad making boxes, helical gears, tapered screws, jaw grips, etc - all the things mechanical engineers take years to to learn to design and manufacture cut short to a few days, and built from scratch, thanks to 3D printers and open scad scripts.
Also notice freecad 0.13 is nothing like its predecessors. It can do operations along prescribed paths (something which openscad lacks at the moment), and there is a (hacked) way to do assemblies. Wings3D is also fantastic, and meshlab takes cares of checking the STL to get objects into printable state.
So who needs open source parametric case designed in openscad? and 3D samples of it printed?
It should be possible to print 3 cases per day small production runs and evolve the case design to something respectable in weeks.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 21/05/14 09:21, joem wrote:
Please, Luke, respect
<jaw drops>
http://liliputing.com/2014/05/novena-open-laptop-raises-700-thousand-crowdfu...
Its even got a deluxe rod end bearing to prop its door open :(
The stereo speakers are mounted side by side (at last) :( :(
</jaw drops>
Marketing marketing marketing marketing and promotion
bunny has: *a good flowing on his blog. hes famous for xbox reverse engineering *updates are posted in (one) place thats easy to see and read, theres a comfortable writing style and flowing text. *there is no "politics" like there is with eoma-* and software freedom. techys hate anything with/about politics, My conclusion is techys hate it because they do not like emotional difficult/painful things and avoid them. if it makes them making difficult decisions or puts what they do as negative or be a source of conflict. *bunny has only mentioned acceptable amounts of software freedom related stuff and when he has, it's been techy acceptable format. In the context of being able to change every thing, non-free bits would be limiting and so frustrating. *the product is for enthusiasts by enthusiasts, hmm - peer group. * note it's also a board, not some fancy upgradable thing that has a chance of only doing one or 2 products and dyeing, = risk in there minds and unnecessary compromises or something. the normal way is tried and tested. I feel more confitable with the average norm.
*techeys see eoma-68 as limiting. what no ffpga! what 1ess usb!, not many gpi's! they don't get it and nor do they like the thought of "things getting in the way". things like cubieboards, olimex boards fit there mind set. they don't get the big picture. they don't get that most of the features they want can be had. I expect theres more they don't get.
It's a big sign that there is a problem with what info there is easily available about eoma-*. I've also mad mistakes in my few discussions. sorry. I know your all busy busy busy. you need others who get it to help. which this newbie trying to do but my progress is very slow. I could do with help from others on the side line too :)
Disclaimer: Please note this post is my opinion, a fat amount of what I have written is my conclusions based on chats I've had with a few techys! just covering my self.
It should be possible to print 3 cases per day small production runs and evolve the case design to something respectable in weeks.
sounds great!
finally one more email out of my drafts :)
Let's also note he aimed for a higher priced specialty market. His prices were so high he could do it in small run and didn't need to attract the same volume. If people are willing to pay $500-$2000 the whole manufacturing equation changes. This way you can do small runs and pay a lot of NRE and still be viable.
I wouldn't have thought you could find even the hundreds needed at those kinds of prices. I guess I was wrong. I'm sure as you mention his following and reputation were a big part of this.
So kudos to him, I don't think a lot of us could have had the same success in getting people to buy at those prices.
Sent from my android device.
-----Original Message----- From: Alexander Stephen Thomas Ross maillist_arm-netbook@aross.me To: Linux on small ARM machines arm-netbook@lists.phcomp.co.uk Sent: Fri, 23 May 2014 7:49 AM Subject: Re: [Arm-netbook] Improv And Operation:Marketing for EMOA-*
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 21/05/14 09:21, joem wrote:
Please, Luke, respect
<jaw drops>
http://liliputing.com/2014/05/novena-open-laptop-raises-700-thousand-crowdfu...
Its even got a deluxe rod end bearing to prop its door open :(
The stereo speakers are mounted side by side (at last) :( :(
</jaw drops>
Marketing marketing marketing marketing and promotion
bunny has: *a good flowing on his blog. hes famous for xbox reverse engineering *updates are posted in (one) place thats easy to see and read, theres a comfortable writing style and flowing text. *there is no "politics" like there is with eoma-* and software freedom. techys hate anything with/about politics, My conclusion is techys hate it because they do not like emotional difficult/painful things and avoid them. if it makes them making difficult decisions or puts what they do as negative or be a source of conflict. *bunny has only mentioned acceptable amounts of software freedom related stuff and when he has, it's been techy acceptable format. In the context of being able to change every thing, non-free bits would be limiting and so frustrating. *the product is for enthusiasts by enthusiasts, hmm - peer group. * note it's also a board, not some fancy upgradable thing that has a chance of only doing one or 2 products and dyeing, = risk in there minds and unnecessary compromises or something. the normal way is tried and tested. I feel more confitable with the average norm.
*techeys see eoma-68 as limiting. what no ffpga! what 1ess usb!, not many gpi's! they don't get it and nor do they like the thought of "things getting in the way". things like cubieboards, olimex boards fit there mind set. they don't get the big picture. they don't get that most of the features they want can be had. I expect theres more they don't get.
It's a big sign that there is a problem with what info there is easily available about eoma-*. I've also mad mistakes in my few discussions. sorry. I know your all busy busy busy. you need others who get it to help. which this newbie trying to do but my progress is very slow. I could do with help from others on the side line too :)
Disclaimer: Please note this post is my opinion, a fat amount of what I have written is my conclusions based on chats I've had with a few techys! just covering my self.
It should be possible to print 3 cases per day small production runs and evolve the case design to something respectable in weeks.
sounds great!
finally one more email out of my drafts :)
_______________________________________________ arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 23/05/14 16:02, neal@pengpod.com wrote:
Let's also note he aimed for a higher priced specialty market. His prices were so high he could do it in small run and didn't need to attract the same volume. If people are willing to pay $500-$2000 the whole manufacturing equation changes. This way you can do small runs and pay a lot of NRE and still be viable.
I wouldn't have thought you could find even the hundreds needed at those kinds of prices. I guess I was wrong. I'm sure as you mention his following and reputation were a big part of this.
no he woundn't have got hundreds of orders, but didn't he get a few hundred not 230ish? it helped, that was all I was saying. I'm noooo expert but a bloody newbie. I could do with a less, err, umm, "asserting" writing style, again please help me.
So kudos to him, I don't think a lot of us could have had the same success in getting people to buy at those prices.
oh yes well done to him, hes made some good hardware but like I said a completely differ market.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
I guess it's a tragic case of them (mpl) feeling down and lack motivation/stressed, etc, sad & sad that it's cost us the nice factory but oh well. moving on....
Operation Savage what does this mean for the tablet? do you have access to the designs, if so all or some? Could it be savaged?
So to confirm. What would this product be?
Is this the dead simple to make board the exposes the interfaces? Could we have a improved interfaces expose board?
What about calling it the "Expose" or BodgeIt or ...?
* More compact like the Improv? I'd love the eoma-68 card to go underneath the Expose - -* ...and even better for the Expose to be the the same dimensions as the emoa-68 card. * I'd appreciate a built-in tiny 4 port usb hub * Flexable power circuity with monitoring. So one can use battery’s and the system will know there status. * built in or external addon board for battery charging li-* or other battery types? (I just want easy way of adding a battery and for my system to know when it's low. it might be a 12v battery powering other stuff (ie on my camera rig, 12v bat powering my external little 12vlcd for my camera) or heck ni-cd or Li-*. There's microUPS, but thats on the big side and is via a serial interface https://www.indiegogo.com/projects/microups-for-raspberry-beaglebone-cubiebo... )
* Audio out and/or Headphone Out * bonus: audio inputs like mic and line in but if there going to be nosiey things then they don't matter
but for audio that means usb audio or the multi-job ardunio-like chip, so what about including the multi-job ardunio-like chip, and have it available to provide some extra options? I don't know. or do you just include a 5 or 6 port ultra compact usb hub instead? hmm
On Tue, May 20, 2014 at 5:54 PM, Alexander Stephen Thomas Ross maillist_arm-netbook@aross.me wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
I guess it's a tragic case of them (mpl) feeling down and lack motivation/stressed, etc, sad & sad that it's cost us the nice factory but oh well. moving on....
well... i'm not sure you understand the implications. without that factory's resources there really is no project.
Operation Savage what does this mean for the tablet? do you have access to the designs, if so all or some? Could it be savaged?
we need money, it's quite simple. i can then send the design off to a prototyping house, get one made up and tested, and we can carry on.
so this is why i am in holland earning money. until i have money, the project is shelved. find the money, the project is not shelved.
l.
On Tuesday, May 20, 2014 18:12:59 Luke Kenneth Casson Leighton wrote:
so this is why i am in holland earning money. until i have money, the project is shelved. find the money, the project is not shelved.
money and time. the first run at the tablet was projected to take a number of months from where everything was at the time and ended up taking a year and still not having all hardware design complete. unless something has changed in the meantime, i'd expect to face similar time constraints with another go- around ... even if the official answer given is "oh, no time at all!" (been there..)
Apologies if this turns out to be a stupid suggestion.
Calculations for prototypes have so far been based on runs of O(100) fully assembled boards. Would there be any possibility of paying for a handful of PCBs, a couple of sets of components (farnell, RS, digikey) and finding a volunteer with some equipment to assemble by hand (paste, oven, ...)?
I'm no expert (it probably shows), but the board doesn't seem hyper dense in terms of components. Of course, there would be no econmoies of scale at all, but it might provide an affordable way of inching past the all-important debugging and validation stages.
Best wishes
On 20/05/14 18:12, Luke Kenneth Casson Leighton wrote:
we need money, it's quite simple. i can then send the design off to a prototyping house, get one made up and tested, and we can carry on.
On Tue, May 20, 2014 at 6:59 PM, Boris Barbour barbour@biologie.ens.fr wrote:
Apologies if this turns out to be a stupid suggestion.
Calculations for prototypes have so far been based on runs of O(100) fully assembled boards. Would there be any possibility of paying for a handful of PCBs, a couple of sets of components (farnell, RS, digikey)
yes of course (except some of them are only available from china, but that's ok)
and finding a volunteer with some equipment to assemble by hand (paste, oven, ...)?
if they're prepared to do it, yes. there are however quite a lot of components, they're quite small, but it is doable.
I'm no expert (it probably shows), but the board doesn't seem hyper dense in terms of components.
it is, but there's nothing smaller than 0402. the smallest drill sizes are i think 7mil, but there are not many of those (they're for the Micro-HDMI)
Of course, there would be no econmoies of scale at all, but it might provide an affordable way of inching past the all-important debugging and validation stages.
we achieved that 100% debugging and validation phase back in i think it was october of last year.
so we are talking about fully-tested, fully working hardware that is ready to go into larger production runs for clients.
l.
On 20/05/14 21:16, Luke Kenneth Casson Leighton wrote:
we achieved that 100% debugging and validation phase back in i think it was october of last year.
so we are talking about fully-tested, fully working hardware that is ready to go into larger production runs for clients.
Sorry - I didn't realise it had progressed quite that far.
It's a crying shame how difficult it is to find arm laptops (also tablets and of course phones) to run debian* without jumping through hoops. The chromebooks show that it is entirely possible, but there are always driver problems, you can't install natively, etc. I really like the eoma idea.
Anyway, I'd be a customer of the first products.
*distribution of your choice
On Tue, May 20, 2014 at 9:34 PM, Boris Barbour barbour@biologie.ens.fr wrote:
On 20/05/14 21:16, Luke Kenneth Casson Leighton wrote:
we achieved that 100% debugging and validation phase back in i think it was october of last year.
so we are talking about fully-tested, fully working hardware that is ready to go into larger production runs for clients.
Sorry - I didn't realise it had progressed quite that far.
yes! that's why we entered into an agreement with aaron's company that they would order 2500 units.
It's a crying shame how difficult it is to find arm laptops (also tablets and of course phones) to run debian* without jumping through hoops.
... and with the eoma idea, it's not just solved once and can be inserted into multiple devices and *still* you only had the porting to do just the once, but also the kernel infrastructure needed to support multiple devices actually makes it easier to port to the next CPU card as well.
The chromebooks show that it is entirely possible, but there are always driver problems, you can't install natively, etc. I really like the eoma idea.
Anyway, I'd be a customer of the first products.
... please put your name down on the list at the preorders page, so that when it happens - once i have enough money to get the prototypes done without relying on other people - it's possible to let you know.
l.
On 20/05/14 21:56, Luke Kenneth Casson Leighton wrote:
... please put your name down on the list at the preorders page, so that when it happens - once i have enough money to get the prototypes done without relying on other people - it's possible to let you know.
Done, although for some reason it appears as a separate page rhombus-tech.net/allwinner_a10/orders/boris rather than on the main list. Basically, I'm up for anything that can make a working product without too much construction or hacking. Amounts are advisory - I'd be excited if it really could work.
I guess a mini-fanless-desktop would be least bother to design (in fact, is there anything to do?).
I wonder if it would be possible to repurpose older laptop cases.
On Tue, May 20, 2014 at 10:21 PM, Boris Barbour barbour@biologie.ens.fr wrote:
On 20/05/14 21:56, Luke Kenneth Casson Leighton wrote:
... please put your name down on the list at the preorders page, so that when it happens - once i have enough money to get the prototypes done without relying on other people - it's possible to let you know.
Done, although for some reason it appears as a separate page
... if you go one level down you will see the complete list.
rhombus-tech.net/allwinner_a10/orders/boris rather than on the main list. Basically, I'm up for anything that can make a working product without too much construction or hacking. Amounts are advisory - I'd be excited if it really could work.
I guess a mini-fanless-desktop would be least bother to design (in fact, is there anything to do?).
not a lot - i have the design already done as a 1st prototype. http://rhombus-tech.net/community_ideas/mini_desktop/news/
I wonder if it would be possible to repurpose older laptop cases.
that's one of the ideas that's been considered, although it is a hell of a lot of work. apart from anything, as the bloom laptop team found out it takes two professors and an entire team of students *over two hours* to work out how to disassemble standard laptops into their 140+ constituent parts.
l.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Opps just see aarons 2 replys before I did my post. apologises about what I err umm wrote,assumed,etc in my prev post as I didn't see them before I sent my post.
On Monday, May 19, 2014 16:40:17 Alexander Stephen Thomas Ross wrote:
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client?
I would happily support this; as it is, orders are simply coming in far too slow, and are virtually non-existent from the eoma fan-club.
I expect they are stuffed when they can't even mention eoma-68!
If there was promotional value in doing so or if there was an ecosystem around eoma-68, we would certainly do so. Unfortunately there isn't, and there has been very little demonstrated interest in addressing that.
+++ Aaron J. Seigo [2014-05-20 17:06 +0200]:
On Monday, May 19, 2014 16:40:17 Alexander Stephen Thomas Ross wrote:
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client?
I would happily support this; as it is, orders are simply coming in far too slow, and are virtually non-existent from the eoma fan-club.
Hmm. I was expecting a plasma tablet announcment at some point, based on eoma-68, but apparently something has gone wrong and that's now not happenning? Is that right? I feel like I've missed some important information (although I did notice this list go very quiet).
Is there going ot be a plasma tabet at some point? I have a use for one...
In the meantime my playslive card is handy, but I'm somewhat stalled on mailine sunxi support, which is coming, slowly.
I know these things are difficult - I was quite impressed that we got this far. On the other hand, having got this far it seemed likely that we would get further too...
Wookey
On Wed, May 21, 2014 at 1:22 AM, Wookey wookey@wookware.org wrote:
+++ Aaron J. Seigo [2014-05-20 17:06 +0200]:
On Monday, May 19, 2014 16:40:17 Alexander Stephen Thomas Ross wrote:
If there was late help from extra individuals marketing and promoting eoma-* standard devices like Improv and there were magically the number of pre-orders needed would it be possible to resume makeplaylive being a client?
I would happily support this; as it is, orders are simply coming in far too slow, and are virtually non-existent from the eoma fan-club.
Hmm. I was expecting a plasma tablet announcment at some point, based on eoma-68, but apparently something has gone wrong and that's now not happenning?
aaron kindly funded the prototype development, based on the specifications that i gave him. he agreed to those specifications but it became clear later that he may not have thought through the implications at the time, as time progressed and the bar was raised by "competing" monolithic products.
so by prototype 2 when we finally had something to do software development on of the firmware, and i repeated the specification to him, to cut a long story short in effect he changed his mind and wanted a better specification.
checking the news page, i have had the first revision prototype of that new specification ready to be sent to a factory since 10 nov 2013. i mentioned that it needs around $1500 for components and PCB printing at the time.
Is that right? I feel like I've missed some important information (although I did notice this list go very quiet).
Is there going ot be a plasma tabet at some point? I have a use for one...
In the meantime my playslive card is handy, but I'm somewhat stalled on mailine sunxi support, which is coming, slowly.
I know these things are difficult - I was quite impressed that we got this far. On the other hand, having got this far it seemed likely that we would get further too...
it basically needs money, that's all. that's what it will take for the projects to continue. i've mentioned this a number of times, and as i have not yet received any responses i am earning the money myself so that i can continue.
i tried it once: i cannot do full-time project management, software development, component sourcing and hardware development *and* a full-time job without serious health consequences so it is one or the other.
so it is very simple. i will get this done; the question is whether anyone else would like to help out financially to get it done. if they do, it will get done sooner.
l.
it basically needs money, that's all. that's what it will take for the projects to continue. i've mentioned this a number of times, and as i have not yet received any > responses i am earning the money myself so that i can continue.
i tried it once: i cannot do full-time project management, software development, component sourcing and hardware development *and* a full-time job without serious health > consequences so it is one or the other.
so it is very simple. i will get this done; the question is whether anyone else would like to help out financially to get it done. if they do, it will get done sooner.
l.
How much money is needed for the EOMA68 CPU Card?
Are you looking for product prepayment to fund final development, an investment, or a donation?
Craig Ballew Madisonville, TN
Time is the most valuable thing a man can spend. ~ Theophrastus
This message and any attachments are intended only for the use of the addressee and may contain information that is privileged and confidential. If the reader of the message is not the intended recipient or an authorized representative of the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited which may be enforced through criminal and/or civil proceedings. If you have received this communication in error, please notify me immediately by e-mail and delete the message and any attachments from your system.
On Wed, May 21, 2014 at 2:00 AM, Craig Ballew arm-netbook2013@mymemopad.com wrote:
it basically needs money, that's all. that's what it will take for the projects to continue. i've mentioned this a number of times, and as i have not yet received any > responses i am earning the money myself so that i can continue.
i tried it once: i cannot do full-time project management, software development, component sourcing and hardware development *and* a full-time job without serious health > consequences so it is one or the other.
so it is very simple. i will get this done; the question is whether anyone else would like to help out financially to get it done. if they do, it will get done sooner.
l.
How much money is needed for the EOMA68 CPU Card?
a batch of 2,500 is about $USD 40 each, that's including case but no FCC/CE, and a batch of only 500 i got an estimate from another supplier, PCB only, if i recall correctly... was about another $5. the reason for PCB only is because they are very busy, and for only 500 units the time it takes to organise a quote including contacting the casework supplier it would have been unfair to have them do that. anyway expect about another $5 for casework and assembly, so around the $50 mark for 500 units.
Are you looking for product prepayment to fund final development, an investment, or a donation?
i'm open to all three.
btw, joe... you'll like this i am sure, but it is a bit of a risk.
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
... pause so that joe can reply "hooray".
the other i think should be a EINT-capable GPIO. also i am thinking to take over one of the 5V power lines as a "power on" button.
now the thing here is that if done carefully these could all be "compatible" (i.e. not do any damage) to existing EOMA68 CPU Cards.
but the time to do that is _now_: it cannot be done later. it does mean a minor CPU Card board revision, but if this project takes off with EOMA68 as it is then we are stuck with a very limited set of SoCs.
so, it is a bit of a risk/delay but i think it will be worth it.
l.
On 23/05/14 22:57, Luke Kenneth Casson Leighton wrote:
a batch of 2,500
That's great for once the whole thing has taken off. I know the Eoma has been very carefully designed to cheaply mass-produced (and for the design to be recycled). But I think it's unrealistic to expect those kinds of numbers until a fairly polished product is shipping. We would be headed for crowd-sourcing failure. I think we need some way to progress and build up momentum that involves much smaller runs and is still affordable.
Several years back I designed an analog circuit for work (4 layers, 20 odd SMT ICs of non-negligible cost, connector). I needed 20 pieces and did not feel like building them myself (too many). In the end I sent the whole caboodle to a company in England* who took the design, made the PCB, sourced the components and assembled the boards. All for 2500 GBP. That gives me hope that the expense for much smaller runs would not be so great. I think we'd have a lot more chance of finding 20 people who'd stump up 100 each to get through one or two revisions.
Joe mentioned 3D printing for any case work. That's a great idea for these kinds of smallish prototype runs. It's ideal for iterating a design and it is perfectly feasible to order single units or small numbers.
* For what it is worth, these were the people: http://www.cogent-technology.co.uk/ . They were great for me, but I didn't shop around that much or try to haggle, so it may be possible to find even cheaper. But they were perfectly willing to take on such a small job.
On Fri, May 23, 2014 at 10:44 PM, Boris Barbour barbour@biologie.ens.fr wrote:
the components and assembled the boards. All for 2500 GBP. That gives me hope that the expense for much smaller runs would not be so great. I think we'd have a lot more chance of finding 20 people who'd stump up 100 each to get through one or two revisions.
boris, i appreciate you're still catching up, here: we went through the 20 boards process already, last year. the first run was a cock-up (junior designer even got the board size wrong! i didn't make it clear enough, apart from anything). the second was 5 units (success), the third was 30 (also success - hand-assembled the connectors, it showed in some cases, but the recipients were able to sort out those few units that had solder bridges).
so with access to an extremely large factory with fully-automated assembly lines that are capable of doing 20 million units *A WEEK* we managed to negotiate them down to a *very* small run of 2500 units. it was only through personal contacts that we managed to do that: their normal sample batch MOQ is 50k to 100k because the lines are so fast that it simply isn't worth their time to roll out the tapes and reels.
so, if we make the minor revisions, i would want to get a batch of 2 to 5 made up just to confirm that they're good. i have a company in the UK that is prepared to do that. then because we know that the rest of the PCB works fine (NAND, HDMI, SD/MMC etc.) i think we could go straight to 100 to 500 units on a crowdfunded site.
l.
On 24/05/14 00:08, Luke Kenneth Casson Leighton wrote:
i appreciate you're still catching up, here
So now I'm caught up.
so, if we make the minor revisions, i would want to get a batch of 2 to 5 made up just to confirm that they're good.
(I've said I'll contribute.)
i have a company in the UK that is prepared to do that. then because we know that the rest of the PCB works fine (NAND, HDMI, SD/MMC etc.) i think we could go straight to 100 to 500 units on a crowdfunded site.
So where can I order my eoma-concept-demonstration mini-pc?
Boris
On Fri, May 23, 2014 at 11:25 PM, Boris Barbour barbour@biologie.ens.fr wrote:
On 24/05/14 00:08, Luke Kenneth Casson Leighton wrote:
i appreciate you're still catching up, here
So now I'm caught up.
:)
so, if we make the minor revisions, i would want to get a batch of 2 to 5 made up just to confirm that they're good.
(I've said I'll contribute.)
awesome.
i have a company in the UK that is prepared to do that. then because we know that the rest of the PCB works fine (NAND, HDMI, SD/MMC etc.) i think we could go straight to 100 to 500 units on a crowdfunded site.
So where can I order my eoma-concept-demonstration mini-pc?
let me draw up a schedule, first. i have to sleep - i'll start the micro-pc PCB tomorrow.
l.
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
I think it's a good idea.
EOMA-68 should be compatible with most SoCs. If SATA prevents EOMA is compatible with most SoCs, I think it is better to delete SATA.
I guess the idea is to change SATA for (another) USB 2.0. Thus, EOMA-68 would have a USB 3.0 and a USB 2.0. I think this is a good idea, since all devices (tablets, phones, portable game devices...) use several USBs, but only very few devices use SATA.
On Sat, May 24, 2014 at 10:36 AM, Miguel Garcia gacuest@gmail.com wrote:
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
I think it's a good idea.
EOMA-68 should be compatible with most SoCs. If SATA prevents EOMA is compatible with most SoCs, I think it is better to delete SATA.
I guess the idea is to change SATA for (another) USB 2.0. Thus, EOMA-68 would have a USB 3.0 and a USB 2.0. I think this is a good idea, since all devices (tablets, phones, portable game devices...) use several USBs, but only very few devices use SATA.
exactly. *sigh* now i have to carefully reroute portions of the CPU Card...
On Sat, 2014-05-24 at 11:36 +0200, Miguel Garcia wrote:
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
I think it's a good idea.
EOMA-68 should be compatible with most SoCs. If SATA prevents EOMA is compatible with most SoCs, I think it is better to delete SATA.
If all you want is to make a product that works for tablets, go ahead. But SATA is a disk standard, while USB is a peripheral bus. As I understand it (and I could be wrong, please educate me), the USB bus generally has to go through CPU, while SATA can often be driven with DMA. I believe the RasPi currently has this problem with its USB busses.
So, if I can echo some of what Luke has said previously, this is for tablet AND tv AND laptop AND mini-computer. Please please please do NOT cut out the SATA connector.
I remain interested in buying a card, to go with my MEB with the upside-down SATA.
Derek
2014-05-25 1:49 GMT+02:00 Derek LaHousse dlahouss@mtu.edu:
So, if I can echo some of what Luke has said previously, this is for tablet AND tv AND laptop AND mini-computer. Please please please do NOT cut out the SATA connector.
I remain interested in buying a card, to go with my MEB with the upside-down SATA.
Yes, EOMA-68 is designed for all types of devices. Therefore, Luke should think best for most devices. If SATA prevents the use of many SoCs and SATA only benefits some (few) devices; thinking of the best for most devices, SATA must be removed.
Luke can always add a USB-to-SATA IC in the MEB, and the MEB will have SATA.
On Sun, May 25, 2014 at 8:41 AM, Miguel Garcia gacuest@gmail.com wrote:
Luke can always add a USB-to-SATA IC in the MEB, and the MEB will have SATA.
... or any 3rd party who creates a correctly-compatible base board.
l.
Luke can always add a USB-to-SATA IC in the MEB, and the MEB will have SATA.
FWIW, for my own use (mostly low-power servers, running 24/7), SATA-via-USB is a non-starter because my experience has always been that these are not nearly as reliable as true SATA. Second problem is the usual poor support for things like S.M.A.R.T or control of spin-down-when-idle. For my use case the performance difference has never been a show-stopper (after all, I lived with a wl-700ge as one of those servers until fairly recently), but of course in many cases it can be yet another deal-breaker.
Stefan
sön 2014-05-25 klockan 15:26 -0400 skrev Stefan Monnier:
these are not nearly as reliable as true SATA. Second problem is the usual poor support for things like S.M.A.R.T or control of spin-down-when-idle.
Depends a lot on the USB->SATA bridge used. Both smartctl and hdparm works fine with at least one of mine to control both SMART settings and access to security feature functions. Haven't tried spindown settings, but if security feature group works then I would assume the spindown parameter settings would work as well.
Regards Henrik
On Sun, May 25, 2014 at 8:26 PM, Stefan Monnier monnier@iro.umontreal.ca wrote:
Luke can always add a USB-to-SATA IC in the MEB, and the MEB will have SATA.
FWIW, for my own use (mostly low-power servers, running 24/7), SATA-via-USB is a non-starter because my experience has always been that these are not nearly as reliable as true SATA.
for low-power servers when SoCs with USB3 arrive it will always be possible to get a higher quality USB3-to-SATA IC.... on the baseboard... for those situations that need it.
.... and the cost of the CPU Card for other purposes that make it affordable and desirable will not be overburdened by the cost of the SATA interface.
look at the cost of the iMX6 ($36) compared to the RK3188 ($12) - and you start to understand what's driving this.
l.
On Sun, May 25, 2014 at 12:49 AM, Derek LaHousse dlahouss@mtu.edu wrote:
On Sat, 2014-05-24 at 11:36 +0200, Miguel Garcia wrote:
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
I think it's a good idea.
EOMA-68 should be compatible with most SoCs. If SATA prevents EOMA is compatible with most SoCs, I think it is better to delete SATA.
If all you want is to make a product that works for tablets, go ahead. But SATA is a disk standard, while USB is a peripheral bus. As I understand it (and I could be wrong, please educate me), the USB bus generally has to go through CPU,
the standard mentor usb hard macros have associated source code which is DMA driven.
allwinner licensed mentor's usb hard macros, and cheerfully completely ignored mentor's source code on which the linux kernel driver for many years has been based, writing their own usb driver which places some considerable load on the CPU.
it has yet to be seen whether they have the good sense to get with the program when they start doing USB3.
So, if I can echo some of what Luke has said previously, this is for tablet AND tv AND laptop AND mini-computer. Please please please do NOT cut out the SATA connector.
we have a choice derek: don't have a successful standard because the number of available SoCs are too far apart, or have a successful standard that has many SoCs per year to choose from over the next decade.
that isn't much of a choice, so, logically, we go with the success option.
l.
Hello,
On Sat, 24 May 2014 19:49:17 -0400 Derek LaHousse dlahouss@mtu.edu wrote:
[]
As I understand it (and I could be wrong, please educate me), the USB bus generally has to go through CPU, while SATA can often be driven with DMA.
This is ridiculous, any bus can be driven by DMA, and how well that works depends on particular controller (and driver) implementation, not on a bus.
I believe the RasPi currently has this problem with its USB busses.
That's probably because Raspberry Pi is made of botched hosed hardware and software, and good people warned on staying away from it, because from it, more bad things can be learned, than good (http://whitequark.org/blog/2012/09/25/why-raspberry-pi-is-unsuitable-for-edu... , etc.)
[]
On Sun, 2014-05-25 at 11:24 +0300, Paul Sokolovsky wrote:
On Sat, 24 May 2014 19:49:17 -0400 Derek LaHousse dlahouss@mtu.edu wrote:
As I understand it (and I could be wrong, please educate me), the USB bus generally has to go through CPU, while SATA can often be driven with DMA.
You are wrong.
Well, I'm certainly happy to be wrong.
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net: <snip>
btw, joe... you'll like this i am sure, but it is a bit of a risk.
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
Since we going for interoperability USB makes more sense indeed. USB1,2,3 is somewhat backward compatible, interoperable and is seems the better choice.
Can we get away with a USB2 only on a USB3 connector? Otherwise the SoC is forced to have USB3
No native SATA would reduce the use-case for a NAS.
If only thunderbolt would be widespread interface on SoC's.
<snip>
so, it is a bit of a risk/delay but i think it will be worth it.
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Mon, May 26, 2014 at 2:38 PM, mike.valk@gmail.com mike.valk@gmail.com wrote:
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
<snip>
btw, joe... you'll like this i am sure, but it is a bit of a risk.
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
Since we going for interoperability USB makes more sense indeed. USB1,2,3 is somewhat backward compatible, interoperable and is seems the better choice.
yes. read the EOMA68 spec. the section on USB is based on exactly this premise... and explicitly bans SoCs such as some of the TI ones which implement 480mbit/sec high-speed *only* on the USB2. if such SoCs were to be used they would need to be firewalled behind a USB2 Hub IC that could do the down-level (to USB1.1 and 1.0) protocol conversion.
Can we get away with a USB2 only on a USB3 connector?
of course.
2014-05-26 15:51 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
On Mon, May 26, 2014 at 2:38 PM, mike.valk@gmail.com mike.valk@gmail.com wrote:
2014-05-23 22:57 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
<snip>
btw, joe... you'll like this i am sure, but it is a bit of a risk.
the inclusion of SATA on EOMA68, it is cutting us off from a ton of CPUs that are coming out for the tablet,tablet,tablet,tablet market - the $12 rk3188 for example, which would otherwise be perfect.
so i am inclined, especially because i anticipate USB3 SoCs coming along over the next 8-9 years, to replace SATA with USB2 and 2 other lines. i think, joe, that one of them should be the "TTL high Power Line" for the voltage levels on GPIO (and UART).
Since we going for interoperability USB makes more sense indeed.
USB1,2,3 is
somewhat backward compatible, interoperable and is seems the better
choice.
yes. read the EOMA68 spec. the section on USB is based on exactly this premise... and explicitly bans SoCs such as some of the TI ones which implement 480mbit/sec high-speed *only* on the USB2.
Yeah that was a lame direction. Especially since USB2 certification demands a USB1.0/1.1 compatiblity.
if such SoCs were to be used they would need to be firewalled behind a USB2 Hub IC that could do the down-level (to USB1.1 and 1.0) protocol conversion.
Can we get away with a USB2 only on a USB3 connector?
of course.
That means that every USB3, slave, device must accept a USB1/2 link, from a master, EOM68 card, even if a USB3 connector is present on the device holding the EOMA68 card. USB3 requires 5 extra pins, IIRC.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Mon, May 26, 2014 at 3:04 PM, mike.valk@gmail.com mike.valk@gmail.com wrote:
That means that every USB3, slave, device must accept a USB1/2 link, from a master,
yyyup.
USB3 requires 5 extra pins
4 coming off the CPU Card. what's the 5th one?
2014-05-26 16:10 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
On Mon, May 26, 2014 at 3:04 PM, mike.valk@gmail.com mike.valk@gmail.com wrote:
That means that every USB3, slave, device must accept a USB1/2 link, from a master,
yyyup.
USB3 requires 5 extra pins
4 coming off the CPU Card. what's the 5th one?
Apparently: GND_DRAIN: Ground for signal return
http://en.wikipedia.org/wiki/USB_3.0#Pinouts
Probably it used to "drain" signal leakage from the SuperSpeed cable pairs. But how that is to be terminated on the devices properly, I have no idea. I'm not sure if it's necessary in the EOMA connector either.
Couldn't find anything useful on the Internet either. http://electronics.stackexchange.com/questions/107669/usb3-with-fewer-wires
Just a notice in this one: http://www.onsemi.com/pub_link/Collateral/AND9114-D.PDF
Page 7: Quote on PCB design: "Place grounds between high speed pairs and keep as much distance between pairs as possible to reduce crosstalk"
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Tue, May 27, 2014 at 9:40 AM, mike.valk@gmail.com mike.valk@gmail.com wrote:
2014-05-26 16:10 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
On Mon, May 26, 2014 at 3:04 PM, mike.valk@gmail.com mike.valk@gmail.com wrote:
That means that every USB3, slave, device must accept a USB1/2 link, from a master,
yyyup.
USB3 requires 5 extra pins
4 coming off the CPU Card. what's the 5th one?
Apparently: GND_DRAIN: Ground for signal return
ok great. thanks mike.
http://en.wikipedia.org/wiki/USB_3.0#Pinouts
Probably it used to "drain" signal leakage from the SuperSpeed cable pairs. But how that is to be terminated on the devices properly, I have no idea. I'm not sure if it's necessary in the EOMA connector either.
no, that's my understanding.
l.
On Wednesday, May 21, 2014 01:38:25 Luke Kenneth Casson Leighton wrote:
so by prototype 2 when we finally had something to do software development on of the firmware, and i repeated the specification to him, to cut a long story short in effect he changed his mind and wanted a better specification.
Wow. Just .. wow. Luke, I'm tired of you blaming me for failures that had nothing to do with me.
The reality is that we ended up waiting for the case work design, which I paid several thousand for, and which never happened.
We did get PCB prototypes done up as well, and they were botched. Remember the package sent to Philadelphia? The collection of screen, PCB and case which was not assembled as expected and which did not work even once it was? The one we forwarded on to you to work out the hardware issues on the PCB, and which ended up as a huge patchwork of wires you had to add to get things working? (That was some good work on your part, btw, to get that board near working ..) ... and then a new PCB was drafted and *you and your partner* started talking about new SoCs ...
Rather than pumping *up* the specification as you asserted, I was the one who asked you to *drop* the screen resolution in early 2013 to save production costs. I did not ask for more CPU, more RAM, more disk, etc. over that time. Things were kept extremely steady despite the changing competitor landscape, something I repeatedly noted as a challenge to marketing the device as the timeline continued to stretch out.
Your continued misrepresentation on this list is repulsive. Learn to take some responsibility, or at the very least not blame others using lies to save your own face. :/
2014-05-23 12:22 GMT+02:00 Aaron J. Seigo aseigo@kde.org:
On Wednesday, May 21, 2014 01:38:25 Luke Kenneth Casson Leighton wrote:
so by prototype 2 when we finally had something to do software development on of the firmware, and i repeated the specification to him, to cut a long story short in effect he changed his mind and wanted a better specification.
Wow. Just .. wow. Luke, I'm tired of you blaming me for failures that had nothing to do with me.
<snip>
Flamewar. Somebody call the Firbirgade!
The way I see it. Everybody already got scorched. 1. Aaron, your idea for a opensource KDE tablet failed twice now. 2. Luke, he put a lot of energy, into this project. 3. Luke's associate, whom ever that may be, for taking the challenge 4. The factory, QiMod?, for not getting the order they thought they get 5. The community here, "arm-netbooks", due to weeks/moths of silence and then being presented with failure and a blame-game/flamewar
I think that's enough burn marks for everyone.
Thus far I only see mis-communications, or the failure to recognize what the other party receives out of the communication.
I often think that my story is absolute and complete. But how true that may be to me. It doesn't mean it is complete or clear enough that the recieving party understands my story, even if they think they do.
Luke, I get this endeavor is frustrating and painfully slow and full of pitfalls, You try get your idea, dream, to reality. But when it goes to slow you get from pulling into a pushing mode. I relate to that, I'm dutch, but in the end it always bites me. One can control only one person, yourself, and even that control is limited, the rest need persuasion, and that requires patience and observation.
Aaron, you took over this project and got it underground or relocated. For the sake of speed or funding, whatever. But it went completely silent here. Although you may believe you haven't been hiding anything, including yourself, I, personally, don't think you have been open and forthcoming enough. Thats all.
I wonder were we are right now.
EOMA68 Standard. Is it complete enough to call it version 1.0? EOMA68-Allwinner A20 Card. Is it production ready? 1. PCB? 2. Parts? 3. Casework?
EOMA68-Freescale IMX6 Card. Is it production ready? 1. PCB? 2. Parts? 3. Casework?
Interface Board, MEB. Is it production/hacker ready? 1. PCB? 2. Parts? 3. Casework?
The iMX6 development is becomming increasingly active in the OSS community, CuBox-I/C1/HummingBoard, Novena, Utillite.
Even the GPU has a OSS Driver on the right track.
Lets be friends and start pulling the cart together.
I know funding is always a problem. Crowdfunding is only viable when having a working prototype. Let find a way to circumvent that.
I really hope that is becomes a realty. But maybe it gets superceeded by PhoneBloks/Project Ara. Who knows. I really wonder how they are going to fit a SoC + Memory on such a small form. Or are they going to serialize memory access as well? We'll see.
<snip>
-- Aaron J. Seigo _______________________________________________ arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Friday, May 23, 2014 16:58:02 mike.valk@gmail.com wrote:
But it went completely silent here.
Honestly, I didn't consider "here" being a place I needed to communicate regularly; most of the work we did around EOMA68 was done without communication here, and I never saw Improv as being the torch bearer for EOMA68.
Although you may believe you haven't been hiding anything, including yourself, I, personally, don't think you have been open and forthcoming enough.
I'm sorry you feel that way.
I wonder were we are right now.
I can't speak to the EOMA68 bits as I really have nothing to do with that other than as a potential customer.
Interface Board, MEB. Is it production/hacker ready?
- PCB?
- Parts?
- Casework?
No casework done (other than one quick prototype we did up at a hackerspace), but I have one of the completed Improv devices (MEB + A20 EOMA68) sitting on my desk which works flawlessly; I've run it constant-on for days at a time running server and desktop applications. It's ready. We could bang off any number of these tomorrow with the right funding. It's been that way since Improv was announced.
On Fri, May 23, 2014 at 5:24 PM, Aaron J. Seigo aseigo@kde.org wrote:
I can't speak to the EOMA68 bits as I really have nothing to do with that other than as a potential customer.
aaron, i'm not sure if you fully understand. you've been blacklisted by my associate. that is an irreversable decision. around 2 months ago he gave you notice to remove *all* mention of EOMA from *all* web sites under your control. i don't have control over what my associate does, nor do i have any influence over him in this regard, because my reputation with him has been destroyed by what has transpired.
so i am very sorry to have to spell it out, but you will *never* be a customer of *any* EOMA or QiMod products, *ever*, and you will *never* be granted a license to make EOMA-compatible products. and that's not my decision, but we both have to live with that.
messing up someone's relationship with a multi-billion dollar factory has consequences, ok?
l.
Responding belatedly as someone pointed me to this and an article written online that covered this issue .. I'd like there to be at least public record somewhere of the reality of things. Personally I had moved on, but as it apparently continues to circulate and escalate on its own, I feel the need to respond publicly.
People on the list here can feel free to skip over this email if you were not involved in the project Luke is referencing, though it does hold a cautionary tale for anyone who would look to EOMA68 and those behind it as a viable solution.
On Friday, May 23, 2014 22:59:41 Luke Kenneth Casson Leighton wrote:
On Fri, May 23, 2014 at 5:24 PM, Aaron J. Seigo aseigo@kde.org wrote:
I can't speak to the EOMA68 bits as I really have nothing to do with that other than as a potential customer.
aaron, i'm not sure if you fully understand. you've been blacklisted by my associate. that is an irreversable decision. around 2 months ago he gave you notice to remove *all* mention of EOMA from *all* web sites under your control.
I never received any such notice.
messing up someone's relationship with a multi-billion dollar factory has consequences, ok?
As does taking 1000s of $ from a me as your client and never producing the promised product.
A realistic appraisal would be to look at QiMOD's string of failures over the entirety of 2013 that I had absolutely nothing to do with. QiMOD failed to meet deadline after deadline that you had set for yourselves. Those failures stretched out over a long enough period of time that both my own resources for the projects were exhausted and external sources of funding that were at my disposal were withdrawn due to the (rightfully) perceived risk in continuing with QiMOD as hardware partner.
Blaming a possible customer representing a maximum order of 10k pieces (going back to early 2013, better times indeed) for ruining *your* relationship with a multi-billion dollar factory is, frankly, astounding. When everything is riding on such a small order, it becomes obvious just how precarious things were on the EOMA / QiMOD side.
Given that there are zero other customers banging down your door and that we *never* signed a production PO, I dare say you are trying to deny responsibility. Always easier to blame another and let their reputation bear the consequences of your mistakes?
I hope others thinking of doing business with you take that as the warning it ought to be.
you will *never* be a customer of *any* EOMA or QiMod products
This is the definition of a paper tiger.
aaron: it does not matter what you think happened. what matters is that you failed to keep a supplier informed, you embarrassed them in front of the factory (a multi-billion dollar state-sponsored company), massively disrupting the entire project as a result.
i do not have time to deal with this. please can you unsubscribe yourself from this list so that i do not have to do it for you.
l.
2014-06-03 15:05 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
aaron: it does not matter what you think happened. what matters is that you failed to keep a supplier informed, you embarrassed them in front of the factory (a multi-billion dollar state-sponsored company), massively disrupting the entire project as a result.
i do not have time to deal with this.
Then don't. Don't respond.
please can you unsubscribe yourself from this list so that i do not have to do it for you.
I feel that that will counter productive. Wat is done is done. Having Aaron on or off the ML won't change anything. Forcing it will only imply censorship.
Trying to silence someone, is a bad choice. The noise will only grow bigger.
You and your associate have chosen to not to do business deals with Aaron or his team. Fine, understandable.
Don't forget that Aaron intentions were probably good and probably still are. The means to result have failed.
Aaron still represents a part in the KDE/ActivePlasma movement. You'll bumb into each other again in the future when either of the goals is met.
You both lost something. Searching for blame will not change it.
You have said your part, Aaron has said his part. We/I have said mine/ours.
Let people decide for their own
Let's move one. All of us.
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Fri, May 23, 2014 at 11:22 AM, Aaron J. Seigo aseigo@kde.org wrote:
On Wednesday, May 21, 2014 01:38:25 Luke Kenneth Casson Leighton wrote:
so by prototype 2 when we finally had something to do software development on of the firmware, and i repeated the specification to him, to cut a long story short in effect he changed his mind and wanted a better specification.
Wow. Just .. wow. Luke, I'm tired of you blaming me for failures that had nothing to do with me.
aaron: i have kept silent on a number of things, when i should not have done. yes i made mistakes: they're part of a learning curve and product development, and we are doing this all at a much lower cost than would normally be attempted. normally, mistakes would be hidden from the client, because the amount of money paid would be *vastly* higher than just the operational costs.
i have spoken to other companies asking them if they would like to help: many of them misunderstood my request and sent me back quotes for full product development: those quotes were often somewhere around $USD 250,000. in total, on both the CPU Card and the tablet PCB, we've gotten away with around $20k so far. which is extremely good going.
when you threatened me (that if we made a competitor product to Improv you would ensure that we never sold a product to a software libre community ever again. i wasn't intending to: i did reassure you of that, so the threat that you made was completely unnecessary) but i decided to let that pass, and, at a critical time in the project, allowed you take control of the project, in effect.
i figured what the heck: you have the experience, you have the confidence, and the contacts, let's see what happens. so i gave you the database of all the people whom we had collected, and we skipped the phase of doing a run of up to 150 units, where the people on the database included those people who would have been happy to pay in some cases up to $USD 150 for a board, as a loyal way to get the project to the next phase.
... what happened?
well, we ended up with our relationship with *two* factories being destroyed, and we have a small dedicated client list who have paid you money for a product that they are never going to get, until you return their money to them. even when you do that, some of those loyal people may be so disillusioned that they give up.
but there is more.
the design team that you selected ignored my advice: i let you go through that learning curve as well.
the team that you selected started trying to take total control of the EOMA68 standard, making promises on public forums that they had absolutely no authority to do, placing themselves at risk of being sued for misrepresentation, as well as bringing the standard and QiMod Ltd into disrepute as a result, through that misrepresentation. it was extremely embarrassing to have to correct chris publicly on the makeplaylive forum, but it had to be done.
they also, by ignoring my advice to use lower-cost china-based components from suppliers that i had already found, delayed the entire operation by something like 8 to 10 weeks as they had to redesign the board after selecting USA-based suppliers (of china-manufactured components) and finding that the cost of components was often 400% higher.
overall the team that you selected, who have a lot of experience in PCB design, took *OVER THREE MONTHS* to design a board with under 20 components.
and you're complaining about _me_ making errors on a board with over 200. do you think it is reasonable to continue to complain about me making mistakes on 2 revisions of a quite high-density board on my first major PCB design project when experienced people that you selected cannot deliver a board with 1/10th the number of components?
so, once we effectively handed over complete control to you:
* you destroyed our relationship with two factories * you adversely affected our relationship with the software libre community who has been following the project loyally for a few years now * you didn't deliver on the promise that you made * you didn't keep us informed in a timely fashion as to whether you were likely to be able to keep to that promise * you selected an incompetent team that comprised one person that was known to be *OPPOSED* to the EOMA68 concept and another that misrepresented the standard on public forums * you jeapordised not one but TWO factory relationships by not responding to our communications * you delayed the project by several additional and unnecessary months through the selection of an incompetent team who took three months to design a board with under 20 components.
now, i don't think it appropriate to make judgements, because that is inappropriate. what *would* be good is if you were able to accept these things as lessons and do better next time. by that... ach this is really quite subtle, let me try to explain.
if you have someone who has access to a huge factory, and they are presenting you with the opportunity to gain access to that resource, and if you want access to that resource, then you should damn well make sure that they are kept REALLY happy!!
so if they say "you didn't answer our questions in a timely fashion" what in god's name are you doing *denying* that they even asked you those questions... not once, but TWICE!
if you *really* were truly interested in making this a success, instead of trying to continue to shout me down in a frustrated fashion, you would be apologising profusely and going "eek! this is really bad, how can i make it right, how can i make this a success?"
this is harder than it looks, especially on a budget. retrospectively, we all learned a lot, we're not hugely happy, but we pick ourselves up and carry on. unfortunately i don't have control over my associate's decisions: he has completely lost faith in the software libre community *entirely* (including in me) so he does not listen to my advice any more if it has anything to do with software libre or anyone else *in* the software libre community.
which reminds me: have you removed all mention of EOMA68 and EOMA products from all web sites yet, as my associate requested? that would be good. i *really* do not want to have to tell him that there are still pictures of EOMA68 CPU Cards on makeplaylive or any other forums, or any public discussions going on on forums that you control, a full two months after he sent you a notice to remove them. trust me: he's not someone you want to mess about with.
l.
On Wednesday, May 21, 2014 00:22:23 Wookey wrote:
Hmm. I was expecting a plasma tablet announcment at some point, based on eoma-68, but apparently something has gone wrong and that's now not happenning? Is that right? I feel like I've missed some important information (although I did notice this list go very quiet).
This entire project is on hold indefinitely. That has been noted on the MPL forums as well; we haven't been keeping that secret. The tablet project stretched out much longer than it should have and exhausted my resources in the process, which is what was driving tha tproject.
Is there going ot be a plasma tabet at some point? I have a use for one...
That depends on the ability to find a reliable partner and funding in future.
On Fri, 23 May 2014, Aaron J. Seigo wrote:
This entire project is on hold indefinitely. That has been noted on the MPL forums as well; we haven't been keeping that secret. The tablet project stretched out much longer than it should have and exhausted my resources in the process, which is what was driving tha tproject.
You mean the forum where people have noticed more than 2 weeks ago from external source (this mailing list) that there is a huge problem with the future of Improv and you have not taken 10 minutes to inform them, neither before nor after they discover this, while you continue publishing articles and comments on Google+ with your usual verbosity on other matters ; matters on which you do not claim to have spent $200.000 but you judge more important to communicate ?
Stéphane.
Guys,
Although it's hard to avoid the temptation of a good blamestorming session, it is a shame to see libre people fighting for little gain. Personally, I don't think it's a good idea to use the public mailing list for this.
Getting libre-friendly hardware off the ground is really difficult, and it's not something that anybody has a lot of previous experience with (except Arduino + Raspberry to an extent). I think it is only to be expected that mistakes will be happen, unrealistic assumptions will be made and overoptimistic expectations will not be fulfilled, leaving people in more than awkward situations. What's important is to learn from the mistakes and move forward.
The fund-raising success of the Novena stuff suggests one path forward. The EOMA project will find it a lot easier to raise funding and gain traction if it can demonstrate a prototype product integrating a prototype eoma board. There are some working prototypes of the board already, if I understand correctly. What would be the easiest (=cheapest) product to build? I'm guessing a mini- pc, since there is no screen to integrate and no real space constraint, although a latptop or a tablet would be a bit more exciting. What would be required?
Best wishes,
Boris
On Fri, May 23, 2014 at 2:18 PM, Boris Barbour boris.barbour@ens.fr wrote:
Guys,
Although it's hard to avoid the temptation of a good blamestorming session, it is a shame to see libre people fighting for little gain.
boris: i have to inform the people who have been counting on us to deliver as to what is going on, and why.
Personally, I don't think it's a good idea to use the public mailing list for this.
i totally disagree. in fact i much prefer that *all* details be covered on public mailing lists. i have always disliked private conversations for software libre projects, and now you know why.
The fund-raising success of the Novena stuff suggests one path forward.
bunny went to MIT where he learned hardware design.
bunny has personal funds which he has received from clients and has been able to redirect his own personal money into getting the project off the ground.
he was able to use those personal funds to create prototypes, as well as use his considerable experience to not have to pay anyone else to do the component sourcing and PCB design.
once he reached the point where the PCB was demonstrably working, *then* he was able to approach case designers, and again use his personal funds to create a prototype.
once he had a case design, *then* he was able to get onto the fundraiser sites and raise funds.
now here is the difference:
1) when i started this project, i had zero hardware design experience, and zero funds.
2) i have been operating this project for four years with a $USD 40,000 personal debt continuously kept at bay, putting all and any personal money received into the project *instead* of paying off that debt, but have only received an average of $USD 15,000 income *per year* over the past 4 years, with outgoings averaging around $20,000. various family members and in one case complete strangers have at various times given or lent me personal funds in order to keep going.
3) i did not have the experience to do PCB design, so i asked other people if they would like to help. not one of those people who offered has delivered.
4) i was therefore forced to either pay up to $10,000 of personal funds to get PCBs designed or
5) i was forced to learn PCB design in order to get the project done.
so it is a completely different story, completely different scenario. and yes i have already approached bunny to ask him if he would like to help: he declined.
so, having had so many libre people let me down i am going to do this myself, controlled entirely 100%. they've been offered the opportunity to help out and to share in the profits: my conscience which tells me to prioritise libre people and give them the opportunity to participate is therefore in the clear if i now operate this 100% entirely controlled and receive 100% of the profits.
once it is financially self-sustaining i can revisit that decision.
The EOMA project will find it a lot easier to raise funding and gain traction if it can demonstrate a prototype product integrating a prototype eoma board.
that was done already *months* ago. august or september 2013.
There are some working prototypes of the board already, if I understand correctly. What would be the easiest (=cheapest) product to build? I'm guessing a mini- pc, since there is no screen to integrate and no real space constraint,
correct. you are looking at the mini desktop pc basically. i completed the first revision in december 2013.
although a latptop or a tablet would be a bit more exciting. What would be required?
money, boris. very simple. about $1500 or so per PCB revision including component population and buying. it can be done cheaper but only by someone who is skilled at PCB population.
however that is $1500 per *iteration* and it will likely take at least 2 iterations. there is a technique for reducing the cost and risk, by splitting the PCB into modules and then integrating each part. however in this case as space is quite tight (100mm x 100 or so) it is difficult to do that. and time-consuming.
l.
Boris Barbour wrote:
if I understand correctly. What would be the easiest (=cheapest) product to build? I'm guessing a mini- pc, since there is no screen to integrate and no real space constraint,
The problem is you have to answer the question of "what does this proposed new product offer that products that are already on the market do not".
I know you will say "the ability to replace the CPU card and avoid throwing the rest of the product away". However there are two problems with that.
1: we are talking about a new unproven product line. What confidence does a potential customer have that there will ever be a second generation of CPU card? 2: it relies on their being significant value in the product OUTSIDE the cpu card. If the rest of the product is little more than a breakout board for the CPU card that value isn't there.
So since the modularity advantage is rather theoretical at this point you have to offer something compelling outside of the modularity, whether that is some unique feature, better performance at a given price point than has previously been available or something else. Improv simply didn't offer that, it had a substantially higher price point than the cubie2 which had the same hardware and was already on the market.
On Friday 23 May 2014 16:56:08 peter green wrote:
The problem is you have to answer the question of "what does this proposed new product offer that products that are already on the market do not".
If you are simply interested in performance now, this isn't the project for you. The value proposition for me is maximally free arm hardware. For instance allowing me to run debian natively. I'm completely fed-up of "images on sd cards", "download the driver from google here" etc. I want debian on a disk and upgradeable as normal. Eoma goes a long way in that direction and offers the longer-term avantages of low cost, reuse, modularity and even more freedom. Obviously the economies of scale will not be there initally, but I would definitely pay above the odds for a working product to help the project get started.
On Fri, May 23, 2014 at 4:15 PM, Boris Barbour boris.barbour@ens.fr wrote:
On Friday 23 May 2014 16:56:08 peter green wrote:
The problem is you have to answer the question of "what does this proposed new product offer that products that are already on the market do not".
If you are simply interested in performance now, this isn't the project for you. The value proposition for me is maximally free arm hardware. For instance allowing me to run debian natively. I'm completely fed-up of "images on sd cards",
have you seen the recent discussions on debian-arm about the port of debian-installer to sunxi?
l.
On 24/05/14 00:40, Luke Kenneth Casson Leighton wrote:
have you seen the recent discussions on debian-arm about the port of debian-installer to sunxi?
No. Although it sounds good, I'm not actually sure what that means - that I'll soon be able to run the installer on lots of hardware?
All I know is that I should be able to run debian without hassle on arm like I run it on intel stuff, and benefit from low power, long battery life (for mobile stuff) and low cost (eventually). But nearly all available equipment is tediously constrained in some way or other. I'll be happy to start with a small, silent fanless box.
let me draw up a schedule, first. i have to sleep - i'll start the micro-pc PCB tomorrow.
Waiting...
On Fri, May 23, 2014 at 11:55 PM, Boris Barbour barbour@biologie.ens.fr wrote:
On 24/05/14 00:40, Luke Kenneth Casson Leighton wrote:
have you seen the recent discussions on debian-arm about the port of debian-installer to sunxi?
No. Although it sounds good, I'm not actually sure what that means - that I'll soon be able to run the installer on lots of hardware?
it means no more f*****g about with 4 gigabyte downloads just to get *someone else's* decisions on what software should be installed on *your* hardware, having to undo all of their arbitrary decisions, no more having to work out how to resize SD card images, and you can install directly onto SATA as one of the disks... or do NFS root partitions.... everything.
All I know is that I should be able to run debian without hassle on arm like I run it on intel stuff, and benefit from low power, long battery life (for mobile stuff) and low cost (eventually). But nearly all available equipment is tediously constrained in some way or other. I'll be happy to start with a small, silent fanless box.
USB, VGA, SD/MMC, HDMI, 10/100 Ethernet, 1Gb RAM, 4Gb NAND, Dual-Core A20, a few pins GPIO and an I2C interface. would that do for a start?
l.
On 24/05/14 09:06, Luke Kenneth Casson Leighton wrote:
USB, VGA, SD/MMC, HDMI, 10/100 Ethernet, 1Gb RAM, 4Gb NAND, Dual-Core A20, a few pins GPIO and an I2C interface. would that do for a start?
Mirco-pc indeed :-) Is there a(n optional) disk in this system? You mentioned SATA (vs USB 3) in an earlier email. I suspect that more memory would be useful. But, whatever: the project will have to walk before it can run.
On Sat, May 24, 2014 at 9:33 AM, Boris Barbour barbour@biologie.ens.fr wrote:
On 24/05/14 09:06, Luke Kenneth Casson Leighton wrote:
USB, VGA, SD/MMC, HDMI, 10/100 Ethernet, 1Gb RAM, 4Gb NAND, Dual-Core A20, a few pins GPIO and an I2C interface. would that do for a start?
Mirco-pc indeed :-) Is there a(n optional) disk in this system? You mentioned SATA (vs USB 3) in an earlier email.
yes. i want to keep this very basic and achievable, and also realistically stand a chance of selecting alternative SoCs in the future. the future's going SATA-less (tablet s,tablets,tablets,tablets,tablets) and with USB3 being 5gbit/sec i don't see that as a problem.
putting down a USB-to-SATA IC would approximately double the components, and increase the risk. much better all round if people just add their own USB-to-SATA dongle.
I suspect that more memory would be useful. But, whatever: the project will have to walk before it can run.
... exactly. i already asked: it's $8,000 for a board-level redesign to add 2Gbyte RAM and 10/100/1000 Ethernet.
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On 24/05/14 16:07, Luke Kenneth Casson Leighton wrote:
On Sat, May 24, 2014 at 9:33 AM, Boris Barbour barbour@biologie.ens.fr wrote:
On 24/05/14 09:06, Luke Kenneth Casson Leighton wrote:
much better all round if people just add their own USB-to-SATA dongle.
OK.
I suspect that more memory would be useful. But, whatever: the project will have to walk before it can run.
... exactly. i already asked: it's $8,000 for a board-level redesign to add 2Gbyte RAM and 10/100/1000 Ethernet.
So that's certainly off the menu for now.
On Fri, May 23, 2014 at 3:56 PM, peter green plugwash@p10link.net wrote:
Boris Barbour wrote:
if I understand correctly. What would be the easiest (=cheapest) product to build? I'm guessing a mini- pc, since there is no screen to integrate and no real space constraint,
The problem is you have to answer the question of "what does this proposed new product offer that products that are already on the market do not".
I know you will say "the ability to replace the CPU card and avoid throwing the rest of the product away". However there are two problems with that.
1: we are talking about a new unproven product line. What confidence does a potential customer have that there will ever be a second generation of CPU card?
i cannot answer this question as it stands, because trust is something that *you* have to give. the decision to trust is *yours*, not mine. so the honest answer to that is whether *you* trust *my* committment to the project's success.
now, i can point you in a direction which will allow you to make that decision, bearing in mind that it is down to you. i am a complex and gifted person with a lot to learn about communication and people, often keeping silent when i should not, and speaking out truthfully when people do not wish to hear. but... hey, i am usually well ahead of the curve when it comes to computer innovations.
one day something that i do ahead of everyone else will be highly successful and for once i might even be the one that benefits financially enough from it to be able to keep it completely under my control and on the course that i know will bring the full benefits that i originally envisaged.
2: it relies on their being significant value in the product OUTSIDE the cpu card. If the rest of the product
products PLURAL peter. please take a look at the community ideas page on the rhombus-tech.net wiki: http://rhombus-tech.net/community_ideas/ and when doing so, please bear in mind that this project has a decade-long timespan. unlike other product family design concepts, this one is not on a desperate design-it-as-fast-as-you-can-sell-it-as-fast-as-you-can-before-the-CPU-is-overtaken-by-competitor-products lifecycle.
So since the modularity advantage is rather theoretical at this point you have to offer something compelling outside of the modularity, whether that is some unique feature, better performance at a given price point than has previously been available or something else. Improv simply didn't offer that, it had a substantially higher price point than the cubie2 which had the same hardware and was already on the market.
which means that you have misunderstood the concept, and the full value of the concept.
the first thing you need to understand is that a comparison of a short-term development board to a long-term PRODUCT FAMILY concept is at best completely misguided, and misses the whole purpose of this inititative.
allow me to illustrate through some simple statements:
if you want only cubie (or other developer boards at the lowest possible price), you are at the wrong web site.
if you want a tablet, and a pc, and a laptop, and a games console, and you want to operate all those machines simultaneously, and you are happy to throw them in landfill and replace them each when even the slightest component breaks, you are at the wrong web site.
if however you want to save around 30% on the cost of two (or more) equivalent monolithic products, because you are happy to SHARE THE CPU CARD BETWEEN PRODUCTS, then you are at the **RIGHT** web site.
if you want to upgrade those products in the future, saving over the lifetime of SEVERAL products on eco-waste in the process by being able to recycle the older CPU Card, you are at the **RIGHT** web site.
does that give you a clear enough distinction between the EOMA concept and the monolithic limited product concept?
the committment of customers to EOMA product concept is a *long-term* one, it is *not* a short-term "buy the latest and greatest fashion product and discard it in landfill when bored with it" one.
that's why our relationship with that large billion-dollar factory in china is so important, because we have the possibility to influence them in their product ranges, and due to their size their own bureaucracy is making it very challenging to keep up with the fast-moving pace of the ARM world, now. our product concept allows them to create the same "shell" on a *DIFFERENT* much longer timescale than the CPU cards, which, in bulk, they can make plenty of and then switch very quickly to a new CPU **WITHOUT** having to totally redesign an entire [monolithic] product range.
so it is a good match. they get it. but they *cannot* commit to an R&D budget. at all. they don't have one. at all. they only do ready-made designs (made by other people). and then make 100,000 to 10,000,000 of them :)
l.
arm-netbook@lists.phcomp.co.uk
-
Aaron J. Seigo -
Alexander Stephen Thomas Ross -
Boris Barbour -
Boris Barbour -
Craig Ballew -
Derek LaHousse -
Henrik Nordström -
joem -
Luke Kenneth Casson Leighton -
Miguel Garcia -
mike.valk＠gmail.com -
neal＠pengpod.com -
Paul Sokolovsky -
peter green -
Stefan Monnier -
Stéphane Goujet -
Wookey