Re: Crowdfunding an Ubuntu smartphone (right now)
On Wed 25 September 2013 21:32:13 Dr. H. Nikolaus Schaller wrote: > Am 25.09.2013 um 20:45 schrieb Stefan Monnier: > Production problems show almost immediately, even if there is only one > person. And they show after making let's say 20 units. I.e. it does not > need to produce let's say 1000 units to find real production problems. And > if you produce 1000 and find that 5 are bad, you don't worry as much as if > you have 2 bad in 20. Yield of 995/1000? Amazing!! A Yield of 950/1000 is already considered good. That's what PV runs (and later on production QA) are for, to determine and (if necessary) improve, and sustain your production yield. No users needed for that. Actually it would be a very poor idea to ship PV devices without tests to users and hope for them to find the "lemons". > > Also, if you can upgrade the screen and the CPU separately, you might > > attract a few other users, who aren't so interested in Freedom but do > > like the idea of customizing their phones. > > That is a dream that is not realistic. Every display has a different > connector (there is no standardization!). And every CPU has different > signals and power supply needs. I.e. you can swap an OMAP3505 for an > OMAP3530 > or an DM3730 but nor for an OMAP4 or OMAP5 or Snapdragon or i.MX6. > Because they are not designed for this way of use. Layman's idea of modularization, which never will fly in embedded. All the real progress made in embedded been based on new interfaces that were smarter and faster and smaller than the previously used ones. So it's like saying "if Industry PC Standard Architecture had used a standardized interface (like ISA) then Pcs had evolved faster and were cheaper than they are today". Actually it needed PCI since ISA wasn't appropriate for the next generation of hardware technology. Same with embedded chips, just there you have like 5 duzen interfaces, and most chips have more than one interface, SoCs have like 20 of them. /j - () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments (alas the above page got scrapped due to resignation(!!), so here some supplementary links:) http://www.georgedillon.com/web/html_email_is_evil.shtml http://www.nonhtmlmail.org/campaign.html http://www.georgedillon.com/web/html_email_is_evil_still.shtml http://www.gerstbach.at/2004/ascii/ (German) signature.asc Description: This is a digitally signed message part. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: Crowdfunding an Ubuntu smartphone (right now)
Am 25.09.2013 um 20:45 schrieb Stefan Monnier: >> I don't see reliability as a problem because it depends on what type >> of reliability you are thinking of: component, software, hardware, >> production, or availability. > > Small production runs means very few people have a chance of > discovering, let alone, fixing the various problems that can show up. Production problems show almost immediately, even if there is only one person. And they show after making let's say 20 units. I.e. it does not need to produce let's say 1000 units to find real production problems. And if you produce 1000 and find that 5 are bad, you don't worry as much as if you have 2 bad in 20. > >> In essence it goes to a modular approach - but "modular" typically drives >> cost up (at least for the version having the highest production numbers) >> and is in strong contradiction with miniaturization of handheld devices. > > In my part of the world, phones have been getting bigger rather than > smaller. Only in dimensions - but they became much slimmer in the same step. I.e. the volume has been constant. > And while modularity has a cost, it can be offset by economies > of scale (both in terms of production as in terms of > developping/debugging the kernel support) if that module can be reused > in more places. Free Software strives on standards and modularity. Of course it can. It is a matter of calculation. > Also, if you can upgrade the screen and the CPU separately, you might > attract a few other users, who aren't so interested in Freedom but do > like the idea of customizing their phones. That is a dream that is not realistic. Every display has a different connector (there is no standardization!). And every CPU has different signals and power supply needs. I.e. you can swap an OMAP3505 for an OMAP3530 or an DM3730 but nor for an OMAP4 or OMAP5 or Snapdragon or i.MX6. Because they are not designed for this way of use. > I'd be very happy to have a Free Phablet (and I actually wouldn't > necessarily need it to have cell-phone connectivity, as long as VoIP > works well), even if that's not my favorite form factor: at this stage, > I'm willing to settle for anything smallish. > >> It would be sufficient to bundle buying power (by summing up # of >> units for different projects), so that we get existing modules >> cheaper. I.e. if all projects would use let's say an DM3730+Memory, >> they still can be soldered into different devices. Or WLAN/BT and >> UMTS are already coming as SoC/MCP "modules". > > Right. That is a lower-leve of modularity than EOMA but it provides > similar benefits (not only direct cost, but also development&debugging). > >> So the trick is to use a bigger shopping bag and make a different meal >> out if it every day. > > Exactly. The various "Free Hardware" communities need to pool > their resources. Yes but I have no idea how this could happen. BR, Nikolaus ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: Crowdfunding an Ubuntu smartphone (right now)
> I don't see reliability as a problem because it depends on what type > of reliability you are thinking of: component, software, hardware, > production, or availability. Small production runs means very few people have a chance of discovering, let alone, fixing the various problems that can show up. > In essence it goes to a modular approach - but "modular" typically drives > cost up (at least for the version having the highest production numbers) > and is in strong contradiction with miniaturization of handheld devices. In my part of the world, phones have been getting bigger rather than smaller. And while modularity has a cost, it can be offset by economies of scale (both in terms of production as in terms of developping/debugging the kernel support) if that module can be reused in more places. Free Software strives on standards and modularity. Also, if you can upgrade the screen and the CPU separately, you might attract a few other users, who aren't so interested in Freedom but do like the idea of customizing their phones. I'd be very happy to have a Free Phablet (and I actually wouldn't necessarily need it to have cell-phone connectivity, as long as VoIP works well), even if that's not my favorite form factor: at this stage, I'm willing to settle for anything smallish. > It would be sufficient to bundle buying power (by summing up # of > units for different projects), so that we get existing modules > cheaper. I.e. if all projects would use let's say an DM3730+Memory, > they still can be soldered into different devices. Or WLAN/BT and > UMTS are already coming as SoC/MCP "modules". Right. That is a lower-leve of modularity than EOMA but it provides similar benefits (not only direct cost, but also development&debugging). > So the trick is to use a bigger shopping bag and make a different meal > out if it every day. Exactly. The various "Free Hardware" communities need to pool their resources. Stefan ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: Crowdfunding an Ubuntu smartphone (right now)
Hi Stefan, Am 25.09.2013 um 05:01 schrieb Stefan Monnier: > The main problem I see with such Free and/or Open phone is the "small > production" syndrome. Small productions mean high prices and low > reliability, whereas we need reasonable prices and reliability. Yes, small production is the key problem. I don't see reliability as a problem because it depends on what type of reliability you are thinking of: component, software, hardware, production, or availability. Each one can also be achieved with small production batches and improves with experience of the team running such a project. > > So we need to focus on making larger production. For that, we need to > widen the target "market". Yes, this is the core idea behind the Neo900 project - find a new "market" for an almost existing design and modifying it a little while sharing as many components as possible. In essence the Letux 3704 and Letux 7004 projects are no different from that. > I'm not sure how best to do that, but > I think the key is in making products that can be used in more > situations. > E.g. the EOMA project comes to mind: a single "SoC card" can potentially > be used in various devices (tablet, router, NAS, ...). > > If the core part of the hardware could be shared between communities > such as Openmoko (free phone), Raspberry, etc.. then it'd be easier to > get that core part produced at reasonable cost, and to have a reasonably > reliable kernel running on it. In essence it goes to a modular approach - but "modular" typically drives cost up (at least for the version having the highest production numbers) and is in strong contradiction with miniaturization of handheld devices. And another factor is that producing anything in higher quantities in advance means that someone is willing to put more money on the table in advance, not knowing if the product finally sells. I.e. the financial risk goes up. We all know that there are 10 times as many who would like to have something different than what is available... And sometimes nobody wants what has been made available. Some other thought: it must not be a module that *we* produce (like EOMA) to get production numbers up. It would be sufficient to bundle buying power (by summing up # of units for different projects), so that we get existing modules cheaper. I.e. if all projects would use let's say an DM3730+Memory, they still can be soldered into different devices. Or WLAN/BT and UMTS are already coming as SoC/MCP "modules". So the trick is to use a bigger shopping bag and make a different meal out if it every day. BR, Nikolaus ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
