#qi-hardware IRC log for Tuesday, 2011-02-08

Mauro_RHello Ayla :-)12:12
Mauro_Rnot so much action today in this room ....12:13
kristianpaulACTION !!!!!12:15
kristianpaulhi Ayla12:15
AylaI have a question for you guys12:16
Aylathat'll give some action12:17
kristianpaulgo ahed, just drop it ! :-)12:17
Ayladoes the linux kernel used by the nanonote support suspend?12:17
AylaI almost got it to work on opendingux (the new kernel for the dingoo a320)12:18
Aylabut there's one bug left12:18
kristianpaulYes it suspend, but there is a bug..12:18
Aylaafter a resume, the SD card behaves strangely12:18
Aylaah, ok :)12:18
kristianpaulIn dingux, how long it survive suspend?12:19
Aylaafter a resume you mean?12:19
kristianpaulI mean:12:20
kristianpaul1. You set supend12:20
kristianpaul2. you do something else for certain time12:20
larscbut thats a hw bug12:20
kristianpaulIn the nanontoe it may not resume after some hrs12:21
kristianpaulyes larsc :(12:21
kristianpaulI wonder if is the same with dingux12:21
kristianpaulis the same chip as the nanonote isnt?12:21
larscmore or less12:21
Aylawell I succeded to resume it after one hour, using the RTC12:21
Aylabut I tried only once12:21
kristianpaulone hour is okay12:22
larsctry a 10h12:22
kristianpauli realized it after more than that12:22
Aylathat's because of a lack of power somehow?12:22
larscbut luckly we can workaround the bug in sw, by using the rtc to wake the device up from time to time12:22
kristianpaullarsc: How you can confirm is hw bug? Ingenic confirmed? or you tracked the suspend state some how in the board?12:23
kristianpaulhmm rtc, nice trick :-)12:23
kristianpaulB_Lizzard: ^12:23
B_Lizzardkristianpaul, heard that, I could work something up12:24
larsckristianpaul: i assume its a hw bug, no confirmation12:25
kristianpauldamn i dont, but is good idea12:25
B_LizzardSince I suspend by doing "echo mem > /sys/power/state" I could probably fit a small script in there.12:25
kristianpaullarsc: good asumtion  :)12:25
kristianpaullarsc: there is a linux driver for rtc?12:26
Aylaanyway, except with this bug, does the suspend work correctly?12:26
larscand Ayla added wakeup support some time ago12:26
kristianpaulbut not for too long12:26
kristianpaulAyla: :)12:26
larscAyla: whats you problem with suspend?12:26
kristianpaulwell the SD part i dint tested12:26
Aylathe issue I'm experiencing is that the SD card behaves strangely after a suspend12:27
kristianpaulso it dint behave well after wakeup?12:27
Aylait is possible to read/write to the card,12:27
kristianpauli/o error?12:27
Aylabut for a defined time12:27
larschm, thats rather strange12:27
Ayla(so I guess when I read/write the card it just read/write the kernel buffers)12:27
kristianpaulAyla: did you tried with a single card?12:27
Aylait depends what/where I read/write12:28
AylaI tried with two cards12:28
larscAyla: which linux version is dingux based on?12:28
larscthere were some issues with suspend/resume in the linux core in <= .3612:28
Aylaand the errors I get are not I/O errors, it's more like if the card was not responding anymore12:28
Aylalarsc: it's 2.6.37 now12:28
larsci'll have to catch a bus, i'll be back in 30 minutes or so12:29
kristianpaulis it posible reload the mmc driver just in case after resume? (i think)12:29
AylaI suppose the card is not reinitialised correctly after the resume12:29
kristianpaulme too12:29
kristianpaulAyla: if you can try the 12h suspend will be awesome to confirm the _posible_ hw bug12:32
Aylabut that wouldn't confirm the hw bug12:32
Aylawe're both using linux :)12:33
kristianpauldinguz use JZ4720 as well?12:34
kristianpaultuxbrain: kind expensive the gta04, i saw the guys with the t-shirt12:51
tuxbrainkristianpaul: I have started a lot of thing this year, I can't afford right to be involved in that right now, I wish them the best of lucks13:09
wpwraktuxbrain: so gta04 is still in the stage where they have an IO board in the gta02 case but the cpu board sticks out ? i thought they had everything inside the phone by now ?13:53
tuxbrainI have not see it working but yes they have some gta02 sized boards on the table13:54
tuxbrainI have not see it doesn't mean it's not working , is just that there were a lot of people to attend13:55
tuxbrainand I have taken no pictures of one of those boards13:56
wpwraktuxbrain: ah, i see. maybe mention it in in the commentary ? it's not nice to present their yesteryear's stuff in a way that make it look as if it was the latest and greatest13:57
tuxbrainok going to fix it.13:58
urandom__just an idea i had, when i read about the inferno os from bell labs:  milkymist has no MMU but people want still run some kind of unix on it? why not port Inferno? it needs no MMU, requires few resources, isnt really an unix but looks like an useable os13:59
tuxbrainbtw, Dr Niko has see the post and he likes it :P but I'm go to try to fix it anyway :)14:00
kristianpaulurandom__: there is qemu  in case you dont have a mm1 :')14:00
kristianpaulurandom__: but i think rtems is fine now and couting there are bugs to fix14:01
Action: kristianpaul will read about inferno just for intelectual curiosity14:02
urandom__does rtems work well? i have no experience with it14:03
kristianpaulsure it is fast !14:03
kristianpaulis real-time and real-free14:04
kristianpaulurandom__: flickernoise work out of the box and it uses rtems14:05
wpwrakurandom__: don't design around a problem - fix the problem ;-)14:05
urandom__nah wpwrak i dont want any MMU for Milkymist, we shouldnt design it for running linux14:07
wpwrakurandom__: what would be the point of that ?14:07
wpwrakurandom__: make it more exclusive ? ;-)14:07
kristianpaulurandom__: how indefero kernel culd be a virtual machine? i dont get that14:07
urandom__i would prefer garbadge collection in hardware :P14:08
wpwrakurandom__: write cleaner programs ;-))14:08
urandom__wpwrak make people try new ideas is the point of it14:09
wpwrakurandom__: btw, in a way, an MMU does give you "garbage collection" in hardware :)14:09
Action: kristianpaul wonder if there is a no-mmu movement14:09
wpwrakurandom__: i don't think there's a shortage of new ideas to try ;-)14:10
urandom__yeah i wasnt that serious wpwrak, just wanted to say we should not design milkymist as a linux maschine14:10
kristianpaulurandom__: i dnot think sebaestien tought in linux when designed the SoC, but you could ask him ;)14:11
kristianpaul /j #milkymist14:11
wpwrakkristianpaul: i read this more as sebastien not having a good idea and/or time for implementing an mmu yet and instead of admitting this lack of omniscience and omnipotence, to claim that none is needed :)14:11
wpwrakkristianpaul: i'm not so much worried about him, because i think he understands the use of an mmu quite well, but of other people building some cargo cult around it14:12
kristianpaulwpwrak: :)14:12
wpwrakurandom__: (linux machine) you said, you'd rather have a lisp machine, yes :)14:12
kristianpaul(he understands the use of an mmu quite well) i dont :(... is on my TODO14:13
urandom__in fact i would like everything that is not designed for running C stuff wpwrak ;)14:14
wpwrakall this reminds me a bit of the amiga. amazing video hardware, os capabilities well ahead of the "industry standard", and all this, but ultimately doomed to die in its niche.14:14
kristianpaulwpwrak: (sebastien), no time, no mmu i agree14:14
wpwrakurandom__: C is hard to kill ;-)14:14
urandom__i dont want to kill it, i just want to live in my fancy lisp world pretending is doesnt exists14:16
wpwrakonce linux runs properly on it, you have a solid base for experiments. when you need some tool you already know, it's readily available. you have a ton of languages to choose from, and so on. you have drivers for lots of pluggable hardware (usb and such)14:16
wpwrakyour baseline just jumps up by a few thousand manyears14:16
kristianpaulat least you want to do SDR, then decide to move to a real time executive os14:17
wpwrakurandom__: ah, that reminds me that, while we have sw, hw, and a bit of mech and artwork, we don't have a drugs division in qi-hw ;-)14:17
kristianpaul(SDR) but i guess it can be done in Linux as well but i dont know yet how :/14:17
wpwrakkristianpaul: with the fpga, you can be more aggressive about SDR - just design your heavily hw-assisted subsystem for it14:18
kristianpaulwpwrak: indeed14:18
kristianpaulwpwrak: but are parts in wich hardware logic is not as efficient as sofware one14:19
urandom__wpwrak maybe wolfspraul doesnt want the drug division to be that public ;)14:19
kristianpaulbut time and coding will show me wich parts14:19
wpwrakkristianpaul: (exaggerating) e.g., a five instruction cpu for gps: RECV, FFT, CORRELATE, PUSH, AGAIN :)14:19
wpwrakurandom__: (hidden division) yeah, could be that it's still too early :)14:20
kristianpaulwpwrak: ;-))14:20
wpwrakkristianpaul: i think there is still very little expertise with finding a good equilibrium between sw and programmable hw. most fpga designers treat their fpga like an asic, which is quite different.14:21
wpwrakkristianpaul: so that's where i see a lot of revolutionary potential for mm1. bring this entirely new world to the people.14:22
kristianpaulwpwrak: http://2009.rmll.info/IMG/pdf/slides_jm_friedt.pdf on slide number 14, just trying to keep my energy about working with rtems for now :-)14:22
kristianpaul(most fpga designers treat their fpga like an asic, which is quite different.) agree14:23
wpwrakkristianpaul: (slide) framebuffer access ?14:24
kristianpaulwpwrak: 16, sorry14:25
wpwrakkristianpaul: yes. linux isn't designed for real-time. there are approaches where you have separate a low-latency rt kernel.14:28
wpwrakkristianpaul: rt also means a lot of constraints. you have to design your applications very carefully or they're not rt, even if your system is.14:29
wpwrakkristianpaul: e.g., you better don't try to access files :)14:29
wpwrakkristianpaul: or if you do, you need to pay attention to the organization of the storage device, how communication with it works, how the file system organizes data, and so on.14:30
kristianpaul(design your applications very carefully) good point14:31
wpwraknice. atusb enumerates :)14:34
kristianpaulcheers !14:35
wpwraknow let's see how i can hook my stuff into those control transfers ...14:36
wpwrakand of course, 473 lines of patches to make it behave so far. e.g., they don't believe in "void" and "static". once you add them, some issues appear that look like potential bugs ...14:40
wpwrak("they" being freakusb)14:42
Aylahi again15:29
Aylaabout that SD bug15:29
Aylaif you don't experience it, that's weird, as we share the same driver15:29
larsccould you compile your kernel with CONFIG_MMC_DEBUG=y and see whats in the systemlog after resume?15:34
Aylayes, but I can't do it now ; next week, probably15:37
larscbtw. you have the rootfs on the sd card, right?15:38
Aylabut it behaves the same if I put the rootfs on the NAND15:41
larschm ok15:46
tuxbrainlarsc: FYI it also hangs booting from nand NN recovers after more than 2 hours of suspend, unfortunately I was trying a very unestable version of jlime just aimed for demo porpouses in Fosdem, and sistems hangs after some time of idle and I was not able to recover any logs.15:46
tuxbrainsorry it can recover form suspend if booting from nand with no SD card inserted15:46
tuxbrainok let me wrote it right this time, booting from Nand and without sd card inserted system can recover after more than 2h of suspend, booting with nand with a card inserted behaves exactly as booting from sd, it fails on recover from suspend unfortunately I was trying a very unestable version of jlime just aimed for demo porpouses in Fosdem, and sistems hangs after some time of idle and I was not able to recover any logs.15:50
tuxbrainok now is at least more clear :)15:50
larscwhich kernel version does jlime ues?15:51
tuxbrainon the version I was using15:51
larscthey should upgrade to 2.6.37,  resume with sd card is known to be broken with 2.6.3615:52
mthAyla: did you try rootfs on NAND with 2.6.37 or only with older kernels?15:53
Aylamth: both15:59
AylaI'm testing this since the second 2.6.35 we had16:00
Aylathere, it was not working at all16:00
Aylawhen booting from the SD16:00
Aylawithout the SD, I was able to suspend16:00
Aylanow on 2.6.37 I'm able to suspend even when booting from SD, but there's still that bug16:01
lekernelinduction heater kits16:39
lekernelwebsite is sloppy, but interesting concept16:47
Jay7rjeffries: you was asking about from where in Russia am I iirc17:08
Jay7rjeffries: I'm from Ulyanovsk, middle Volga17:08
wpwraklekernel: i know this question is non sequitur when it comes to this sort of toys, but what would you use it for ?17:09
kristianpaulhe :-)17:10
lekernelme? gassing out vacuum tube electrodes. but the mentioned "plumbing" uses are worthy too.17:11
wpwraklekernel: interesting uses you have ... following in the footsteps of tesla ? ;-)17:19
lekerneloh, it's a classical use of induction heating. that's just a different domain than playing with microcontrollers...17:21
lekerneland a less explored one, too17:22
lekernelactually the designer of said kits also does that: http://www.johndearmond.com/2010/11/13/using-the-induction-heater-neon-processing/17:23
wpwraklekernel: whee ! :)17:25
rjeffriesJay7 that is interesting Thanks. I live in Arroyo Grande, California, USA18:26
wolfspraulkristianpaul: hey you nuked a spammer in the wiki. thanks!19:31
wolfspraulthat's one of the little adrenalin kicks I get out of some mornings. ahhh, a spammer. nice :-)19:31
wolfspraulbut today the page was already gone (which is good)19:31
wolfspraulI do exactly as you did - block the account indefinitely, delete all contributions.19:33
zrafawolfspraul: xiangfu will be with us soon? I would like to know the curretn layout of openwrt and how to set it inside kernel config (I guess that it is there)20:04
zrafawolfspraul: so I can prepare jlime kernel with the same layout. I am guessing that current jlime booted from uSD breaks qi-openwrt on NAND, and I do not want both fighting.. I want both friends :)20:05
wolfspraulzrafa: xiangfu should be back from his hometown already20:13
wolfspraullet's see whether he shows up in this channel later...20:14
wolfspraulwhat do you mean with 'jlime booted from 8:10 breaks openwrt on nand'? how is that possible?20:14
wolfspraulyou mean it writes something into nand corrupting an existing openwrt installation in nand?20:14
zrafawolfspraul: I am just guessing. I read some users report saying.. "hey, I booted jlime, and after that openwrt can not boot anymore".. some checked the ubifs corruption on nand. Also20:20
zrafawolfspraul: I found a similar behaviour with another environment. Gamerunner (the game distro I did for freerunner one year ago) booted from microSD breaks ubifs on freerunner nand. But well, on this case20:21
zrafawolfspraul: gamerunner tries to mount a jffs2 fron nand.. and if there is some ubifs there it breaks it. Here, in nanonote20:21
zrafajlime tries to mount the ubifs nand on some dir for access, using the jlime layout.20:21
zrafaI am guessing that because layouts differ, jlime does something unknown by me which breaks the openwrt ubifs20:22
kristianpaulwolfspraul: but, spam is again!20:22
zrafawolfspraul: I am not completely sure of course, but al least, having the same layout boths (for nand) would help to the dual-booting idea and also would avoid this problem (until we fix it)20:24
kristianpaul(spam)  done, lets see if show up again20:25
wolfspraulok got it. yeah that sounds like an important bug.20:25
wolfspraulkristianpaul: interesting. autoonline12, carinsurance03, looks like the same bot maybe20:25
kristianpaulwolfspraul: i agree20:26
wolfspraulI'm wondering whether they have something to go past the math captcha automatically, or whether some kid is paid 1 cent / captcha20:26
kristianpaulhaha ( kid is paid 1 cent / captcha)20:27
wpwrakkristianpaul: think "gollum"20:32
kristianpaulwpwrak: noo!!, i'll on ben soon :-)20:36
wolfsprauloh sure. those 'microservices' are booming.20:41
wolfspraulthere is also some quite amazing 'content farms' with tens of thousands of people writing trivial texts full-time.20:42
wpwrakkristianpaul: "tricksy hobbits deleted my precious spam"20:42
wpwrakwolfspraul: rent-a-microvandal ? :)20:42
wolfspraulwpwrak: believe or not, some of those content farm corporations are valued in the billions of USD!20:43
wolfspraulwe are definitely 100% doing the wrong thing here :-)20:43
wolfspraulwe should be organizing kids to bypass captchas and others to write trivial nonsense texts20:44
kristianpaulThat business will progress here as soon english will be mandatory on school i bet20:44
wolfspraulI don't think poor english is even a problem.20:45
wpwrakwolfspraul: sigh. that's what we get for living in a secularized society. in places where cruel eternal revenge inescapably awaits you in the afterlife, such things don't happen ...20:45
wpwrakxMff: exactly :)20:46
wpwraknow, be a good usb stack and let me have these two measly status bytes. it can't be so hard. you've already sent me dozens of descriptor bytes ...20:50
steve|mwpwrak: heh.. you're patching freakusb for the avr? ;)20:54
wpwraksteve|m: yeah. but i think it won't be long before i just throw it away :)20:55
steve|mactually there seems to be a real lack of a good, somewhat hardware independent, real open source usb stack20:56
wpwrakwell, i have one that does enumeration. not hardware-independent just yet, but ...20:57
wpwrakthat is, enumeration, vendor/class control transfers, and a bit of bulk20:58
wpwraknothing on top, though20:58
kristianpaultrought avrs?21:00
wpwrakright now, it only does c8051f32[0167]21:00
wpwrakbut i think the avrs aren't too different. the basic logic is always the same.21:01
steve|mthis is something minimal I found recently http://www.pjrc.com/teensy/usb_serial.html21:02
wolfspraul'demand media' is worth 1.5 billion USD21:03
wpwrakyes, i've seen it. the chip seems to be different from mine, though. freakusb has a driver for the usb162 which is (almost) the same as the 32u221:03
wolfspraulyahoo bought 'associated content' for 100 million USD.21:04
wolfsprauland there are probably hundreds of smaller ones. oh well :-)21:04
wpwrakwolfspraul: demand media = monkeys + keyboard ?21:04
wolfspraulyes. trivial content. written as fast and as cheap as possible.21:04
wolfspraulamazing you can build a 1.5 billion USD business out of it, no? :-)21:05
wpwrakfor spam ? or for fox news ?21:05
wolfsprauljust content, doesn't matter what it is21:05
wolfspraulthe pages are then filled up with google ads21:05
wpwrakaah, i see21:05
wolfspraulthey 'employ' tens of thousands of people writing junk21:05
wolfspraulevery day21:06
wolfspraulI can write pretty fast, I think I'd have a chance.21:06
wolfspraula few typos don't matter, so I should be able to do 100 words / minute consistently21:06
wpwrakhmm, i guess i'd have to settle for premium content :)21:06
larscwolfspraul: will these words also form sentences?21:08
wpwrakaha .. i think i can smell the bug ...21:09
wpwrakU8 status[2]; ... memset(status, 0, 2 * sizeof(status));21:11
wpwrakmuch better :)21:11
wpwraknow i pass lsusb with flying colors. but it still doesn't want to call my part of the stack. grmbl.21:12
wpwrakthat is, lsusb -v. lsusb worked out of the box.21:17
qi-bot[commit] Werner Almesberger: atusb/fw2: firmware for the AVR-based atusb (in progress) http://qi-hw.com/p/ben-wpan/259400f21:19
qi-bot[commit] Werner Almesberger: atusb/fw2: moved board-specific functions (SPI, RF reset) out of main() http://qi-hw.com/p/ben-wpan/1c004ee21:19
qi-bot[commit] Werner Almesberger: fw2/usb/patches/correct-array-size.patch: correct memset in ctrl_get_status http://qi-hw.com/p/ben-wpan/06f617421:19
wpwraktuxbrain, rjeffries: how are things going with the pcb fabs ?21:23
kristianpaulfab fab fab !21:26
wpwrakthe UBB experience will also be useful for atben. if we run into any problems, it'll be much easier to solve them on UBB than on atben.21:27
wpwrakyeah ! more good bytes coming out of the board :)21:33
wolfspraulsooo :-) gossip time :-) will nokia ditch meego already on friday?22:17
wolfspraulsymbian is dead, that's clear. meego not sure.22:17
wpwrakhow long will nokia survive whatever they decide to do next ? :)22:18
wolfspraulbut it's the last chance for meego for sure. either nokia gets behind it now 100%, or it will either die fast or slow.22:18
wolfspraulnokia still has momentum, but it's a huge turnaround case now. let's see whether the new ceo has the necessary power for the bloodshed :-)22:19
wolfspraulfire 50% of people, kill 80% of products, and so on22:19
wpwraktheir panicky changes in the last months are eerily reminiscent of openmoko ...22:19
wolfspraulI have the biggest respect for the guys at nokia, it's still a great company. managing is hard, as you know I have many years experience managing smaller groups from 10-40 people, and even that is one hell of a job. cannot imagine what it must be like to come in as new ceo for nokia now. must be a little mao :-)22:20
wolfspraulwpwrak: did you read the 'leaked' memo?22:21
wolfspraulat least their PR is getting better :-)22:21
wpwraknot sure if a big bloodbath is really what's called for. the problem is rather that they need to redefine themselves. and this won't happen within one quarter, no matter what they kill.22:21
wolfspraulyes it's possible I think22:21
wolfspraulshort-term financial improvements are a must, you cannot argue against it with a publicly listed company22:22
wolfspraulso you have to kill 80% of products, really22:22
wolfspraulalso massively lay off people, make one big severance package and get over with it22:22
wolfspraulon the technical side, cut features/variations across the board22:22
wpwrak(leaked memo) reading ...22:23
wolfspraulget out all the old cruft of managers and the pile of people they accumulated around them22:23
wolfspraulget a famous 150 million USD symbolic investment from Microsoft :-)22:23
wolfspraulwe can all learn from Steve godfather Jobs22:23
wolfspraulthey have enough technology I think, they just need to cut and focus on the sales force. the sales force at Nokia is probably in auto-pilot mode for many years already.22:24
wolfsprauljust keep track of incoming orders :-)22:25
wolfspraulI'm mostly only interested in the fate of Meego.22:26
steve|meven MediaTek is mentioned :)22:26
wolfsprauland finally a realistic market share22:26
wolfspraul(for mtk)22:26
steve|m"this ecosystem now produces more than one third of the phones sold globally" - yes22:26
wolfspraulnot that any of this would have been a secret22:26
wolfspraulbut for an organization to bubble up it takes time, understandably22:27
wolfspraulthey have long ignored mtk entirely, basically just 'censored' mtk sales, to make the nokia market share look bigger.22:27
wolfspraulI can see how this works inside such a large corporation, where you pad on each others shoulder, and nobody says that the king wears no clothes.22:27
wolfspraulsteve|m: so what's your take on meego this friday?22:29
wolfspraulwhat will nokia decide about meego?22:30
wpwrakbasically three choices: go meego, go android, or go windows. or does a parallel approach seem likely ?22:31
wpwrakandroid could work, capitalizing on the brand recognition.22:32
wolfspraulit depends on the power of the new ceo. sometimes what you want and what you are able to accomplish is not the same thing.22:32
wolfspraulI'm sure he wants to focus on one system.22:32
wolfsprauland in fact he wants to get away from the pure software comparison thing22:32
kristianpaulparallel aprouch i heard from nokia before22:32
wolfspraulkristianpaul: the only company who is really executing parallel well is Samsung22:33
wpwrakwindows would have some probability considering his background. however, there aren't many success stories among companies that have gotten into such a close technological partnership with microsoft.22:33
steve|mhm, good question.. don't know if meego would work out in the end22:33
wolfsprauland they are brutally honest about it. they say 'we do them all' and 'we care about none of them'. in that combination it can work.22:33
wolfspraulyes but at least microsoft is a strong business partner22:34
wolfsprauland microsoft needs to stop selling software as in the 90's as well22:34
wolfspraulso if you combine it with bing, kinect and what not, maybe they team up?22:34
wolfspraulmicrosoft is sitting on huge cash, and keeps making more cash each quarter (was it 6 billion in q4/2010?)22:34
wpwrakmeego would depend on the time to market. it's tempting to try to copy the approach of apple and google, namely to establish a new platform. the problem there is that you have such a huge company to run.22:35
kristianpaulis for nokia hardware a strong point to just just swich to android may be?22:35
wolfspraulwith meego they have no partners22:35
kristianpauli mean they value more hardware experience than sofware it self?22:35
wolfspraulthe big problem of true foss - no capital22:35
wpwrakand just demolishing most of nokia to bring it down to the right size is probably too demoralizing22:35
kristianpaulsome GSM networks are deploy by nokia arent they?22:36
wolfspraulit's a big corp, not sure what they do. but we are talking about the consumer devices here, or consumer perception/brand value.22:36
steve|mkristianpaul: that's Nokia Siemens Networks22:36
wolfspraulI don't think they will go the Siemens way and totally retreat into the (very profitable) infrastructure backend.22:36
wpwrakwolfspraul: M$ is good as a cash source, not as a tech source22:36
wolfspraulwpwrak: yeah, but cash is quite important for nokia now, I would think.22:37
wolfspraulto turn around such a huge slowly sinking titanic you need HUGE capital.22:37
wpwrakwolfspraul: probably. but they'd have to find a way to get the cash but not the technology22:37
wolfspraulnotice how he mentions the downgrading of their debt facilities22:37
wolfspraulor rather the potential downgrading22:37
wolfspraulif it's a capital driven decision, meego is dead22:38
wolfspraulor maybe something with Intel?22:38
wolfspraulIntel just bought the Infineon cellphone chip business.22:39
wolfspraulyeah another option :-) team up big time with intel. kill everything but meego. switch to x86 phones.22:39
wpwraki wonder how broad meego is within nokia. meego sounds more like a "medium group" activity to me. maybe 100 heads, tops. (good people, of course)22:39
wpwrak(x86) that could be an interesting approach22:40
wolfspraulthat would be a very anti-ms move though, a bit hard to imagine for an ex-ms guy22:40
wolfspraulthen the intel-ms alliance is definitely over for good22:41
wpwrakyeah. i'm a bit afraid that he's MS' agent.22:41
wolfspraulnot 'agent'.22:41
kristianpaulmegoo is good aprouch, but for my point if view was too early for tablets and late for handset22:41
wolfspraulat this level everybody knows inertia. so when they scout people, there is plenty of time for any candidates to share their core beliefs with the board etc.22:41
kristianpaulandroid history is just inverse22:41
wpwrakde facto mole :)22:41
wolfspraulmeaning that, at the time of hiring, the board already knows in which direction it will probably go22:41
wolfspraulI don't think that's fair. I am sure there are no moles or spies or agents or anything like that.22:42
wpwrak(board) oh, they better do :)22:42
wolfspraulyou also cannot argue with someone's 20+ years experience.22:42
wolfspraulif you don't like the particular experience of the person in front of you, then get someone else22:42
wpwrakthat's the point where i'm worried22:42
wolfspraulwhy worried?22:43
wolfspraulmicrosoft needs some transformations too at some point, they may become a better supporter of foss than others22:43
wpwrakif you look at his profile, doesn't it seem likely they got him for a switch towards windows ?22:43
wolfspraulI was surprised to see how bing cooperates with openstreetmap, for example22:43
wpwrakhmm. MS going meego ? now that would be something22:44
wolfsprauloh no, not ms going meego :-)22:44
wpwraknokintsoft ;-)22:44
wolfspraulif they go ms, then it will be windows 7 or whatever future version they do22:44
wolfspraulmaybe meego only has a chance if they make a big intel announcement22:45
wpwrakof course, intel+ms would also work for windows on x8622:45
wpwrakfor courting strong partners, that may be the most convenient compromise. they all hate apple. they all hate android. they have a lot in common.22:46
wpwrakwell, intel may hate android a bit less than the others. doesn't help much with 86 sales, though22:47
wolfspraulI just think financial. those are all businesses (=money making machines)22:47
wolfspraula big intel move could work, true?22:47
wolfspraulthat would be great for meego :-)22:48
wpwrakso the real question may be how independent nokia want to remain22:48
wolfsprauland a lot of good new foss sources to be developed...22:48
wolfspraula big ms move could also work I think22:48
wolfspraulwith Android not so sure, I don't know where the strong business partners (financially) could come from in that move22:48
wpwrakhow about the intel+ms move ?22:48
wolfspraulso Intel has to kill meego too? :-)22:49
wolfspraulwell then, Friday is near22:49
wpwrakgoogle has some cash too. they've seen stuff that had no use for them but that could hurt them in the hands of others get snapped up by enemies. they may not want this to repeat.22:49
wolfspraulfrom my experience I'd say that one is unlikely22:50
wpwrakit would be a very indirect move22:50
wolfspraulgoogle setup android so that the manufacturers go kill each other, and google picks the winners. why should they help this ailing manufacturer now?22:51
wpwrakkeep the IP amassed there under control22:51
wpwrakof course, such a move would sour the relationship with other android manufacturers22:52
wpwrakwell, could. you never know.22:52
wolfspraulunlikely. the android play is made, and is now being executed.22:53
wpwrakit wouldn't be so much a move towards getting another manufacturer but to secure assets that are tied to a manufacturer. secure them before someone else does.22:55
wpwrakbut i'm not sure if google is sufficiently concerned about such things that they'd move in this direction22:55
rjeffriesGood Evening. Since you asked, yes, the Nokia internal memo is fascinating22:56
wolfspraulbtw, don't overestimate nokia. looking at it financially, it is only worth 40 billion USD now, a fraction of google/ms/apple/intel22:57
wolfsprauland then it has big downward inertia attached to it, like the memo says and we all know competitors taking business from them everywhere, high-end, mid-range, low-end22:57
wpwraka similar buy-to-scavenge move would also be possible for others, like apple22:57
wolfspraulgod no, never22:57
wolfspraulif meego doesn't get big support this friday, it's dead22:58
rjeffriesMy prediction is Nokia will team with HP WebOS or Microsoft Windows Phone 7. Android is a NO for multiple good reasons22:59
wolfspraulron has another idea22:59
wolfspraulrjeffries: I like a big intel move.22:59
wolfspraul100% meego, x86 phones, capital investment by intel.22:59
wpwrakof course, the very best they could possibly do would be to liquidate all assets immediately and invest everything in qi-hw ;-)22:59
rjeffriesRon? who is this "ron" who supposedly "has an idea?"22:59
wolfspraulrjeffries: we were discussing this before you joined, but HP is a new one :-)22:59
rjeffriesIntel can't find it's ass with both hands (in mobile)23:00
kristianpaulwpwrak: ;)23:00
rjeffriesI saw the discussion23:00
wolfspraulnot sure. they just bought infineon cellphone business.23:00
rjeffriesI'd put HP at top of the list23:00
kristianpaulwpwrak: they just will die i think23:00
wpwrakrjeffries: MS isn't much better (in mobile)23:00
wolfspraulwhen I think of hp I immediately think of overpriced ink23:00
wpwrak(than intel)23:00
rjeffriesInfineon is uninteresting. this is not about who can make baseband chips23:00
wolfsprauland refilling tricks with mp3 drm songs in the cartdrige and what not23:01
wolfspraulso I probably cannot rationally think about HP at all23:01
rjeffrieswolfspraul I am not suggestion you can or should like it. it doesn't matter23:01
wpwrakHP do a lot of stuff. but i don't really see them there either.23:01
wolfspraulit's gossip anyway, we can be free. it's a good idea.23:01
wolfspraulI am mostly concerned about meego.23:02
wolfspraulif meego doesn't get a shot in the arm soon it's dead23:02
rjeffrieswpwark the layesy Windows Phone 7 is damned good. has a unique and good GUI23:02
wolfspraulthe switch from maemo to meego was already very destructive, just picking the .deb to .rpm switch as an example23:02
rjeffriesmeego is prolly toast. I had hopes for it, bt they have screwed around too long23:02
rjeffriesMeeGo is DOA.23:03
wolfspraulyeah, that's possible. sad.23:03
rjeffriesand I liked the idea a LOT23:03
rjeffriesnew topic23:03
wolfspraulI bought the Nokia 770 enthousiastically many years ago.23:03
wolfspraulI was one of the few who got the first batch at compusa :-)23:03
rjeffriesI put out a copule of emails about a possible griup buy of Ben NN23:04
wpwraki would agree with a pessimistic view for meego. it's the kind of thing that should grow organically, not something two elephants cook out together.23:04
rjeffriesBy the way, I have a friend who works with linix on the Palm platform and lies it  alot23:04
rjeffriesback to the Ben group buy23:05
rjeffriesso far two people besides one for me23:05
wolfspraulwpwrak: I agree with organic, but it is very very hard to establish a competitive business around that.23:05
wolfspraulmaybe impossible, don't know23:05
rjeffriesI will present Ben overview at a local Linux group (not the SBLUG)23:05
wolfspraulat least from how I learnt and know how a business is run. I wouldn't know how to marry that with an 'organic' approach, even if it's confined to technology.23:06
wolfspraulrjeffries: great! [mails and presentation and group buy]23:07
wpwrakwolfspraul: (organic) you'd basically need something that can operate like a startup. kind a like openmoko did. just with an experienced team.23:07
wpwrakwolfspraul: grow the base system without too much pressure. when it's starting to look good, bring in the big guns.23:08
rjeffrieswell, so far it looks like I'll simply buy one myself when the next s/w load is ready (I understand I will have to reflash)23:08
wolfspraulI still don't know how to do that organically. Competition is about who has a solution for the most profitable customers first. It's kinda the opposite of organic.23:08
rjeffriesso whether sane or not, I plan to buy when the next stable bimage after the Dec 14 one is available23:09
wolfspraulI would never dare to do something like we do at Nokia.23:09
rjeffriesthat is because small scale can do things that large scale can not dare to do23:09
wolfspraulrjeffries: come on it must be sane. Once you have one we can finally support you better too!23:09
kristianpaulwolfspraul: so you'll kill megoo if you were that CEO?23:10
wolfspraulnothing rings in my ears louder than something coming from a paying customer.23:10
wolfspraulkristianpaul: no idea. from the outside it's all just gossip. entertaining though :-)23:10
rjeffrieswolfspraul I did not realize I was asking for anything. ;)23:10
kristianpaulI want free cotent for my nanonote !23:10
wolfspraulI am anticipating.23:10
rjeffriesI am waiting patiently as software matyres it is almost there.23:11
wpwrakrjeffries: you should advertize the idea of a ben/owner ratio > 1 :) for those who like to tinker, a second ben is quite desirable23:12
rjeffriesit I set up a build of some BBB cards. I think wpwrak has layed out the plan pretty well23:12
kristianpaulI gossip a new shiny device coming from nokia as one last try to call at least more atention i guess...23:12
rjeffriesso far one univesity hacker wants one23:12
rjeffriesanother perso is thinking about it23:12
rjeffrieswhat does gn8 mean?23:13
wpwrakwolfspraul: (organic) you'd do this in parallel to normal operation. have a number of "risky" but relatively small projects. explore new ground. when something works, jump on it.23:15
wpwrakwolfspraul: of course, once you're beyond a certain point, you don't have that liberty anymore. the tone of that memo suggests that this line has been crossed for meego.23:16
wpwrakwolfspraul: or at least that very good arguments will have to be brought forward very fast23:16
wpwrakrjeffries: the ben won't change for a while. so there's little point in procrastinating :)23:18
wpwrakrjeffries: something to bear in mind: each ben sold means no only a return of money to wolfgang's pocket but also a slight strengthening of the case that can be made in favour of qi-hw23:19
wpwrak(gn8) g = good, n8 is self-explanatory :)23:19
wpwrakrjeffries: (two other people) i saw this and like it :)23:28
rjeffrieswpwrak I have a lot of other things going23:28
wpwrakthis seems to be reboot hour. had to reset my cable modem just a few minutes ago, too.23:28
rjeffriesI have a couple of friends who have recently had PCBs fabbed I willcheck with them on what house they like23:28
rjeffrieswpwrak one at fisrt misunderstood my email and thought 802.15.4 was ready now23:29
rjeffriesI got back and apologized for the confusion23:29
rjeffriesshe still wants a Ben23:29
wpwrak(pcbs) perfect. if they've done something with similar characteristics, that's best.23:29
rjeffriesanother guy runs a hackerspace I thought they'd be interested so far just a yawn23:30
wpwrak(wpan) yeah, still taking a little while23:30
wpwrak(hackerspace) toy overdose already ? :)23:30
rjeffriesmy buddies work with a guy who knows pcb fab inside and out23:30
rjeffriesI'll send him your specs and see what he says23:30
rjeffriesYou had it as one document and then there was one more thiing.23:31
rjeffriesPersonallyI hope Tuxbrain kcks in gear and does a run of UBB23:31
rjeffrieshe has a sales channel. I manly one a few for personal use23:32
wpwrak(one more thing) yeah, the drill file. forgot that, sorry. the process still isn't quite scripted. (and such mistakes are precisely why i hate manual processes)23:32
wpwrakafaik, tuxbrain already contacted a pcb fab. let's see what comes back from them.23:32
rjeffriesNo, I have not had time to start learning Kicad. I hope to find an eager engineering student at local23:33
rjeffriescollege to dig in. we shall see23:33
rjeffriesthis may be much less of a learning exercise for him than me. The only reason I may go forward23:34
rjeffrieswould be for the experience.23:34
wpwrakheh :)23:34
rjeffriesyes I have a LOW IQ. ket's just stat ethat right up front23:34
wpwraki think it will be useful to know a place that can do small stuff. not all of the pcb houses do such things.23:34
rjeffriesbecaus etrying to fab a PCB just to learn is stupid beyond words. ;)23:35
rjeffriesactually I am being less than honesy23:36
wpwrakwell, it's a simple board. and the board itself is the result. so you do't have component stock, smt, assembly, all depend on the pcbs. takes a lot of pressure away.23:36
wpwrak(simple as in "you see immediately if anything went wrong")23:37
rjeffrieswolfsproul may, someday, offer a distributor discount that would allow distributor to make money23:37
rjeffrieseven if he does not there may be a Sparkfun sty;e play with Ben 8:10 gadgets23:37
wpwrakyeah, distributors like accessories :)23:38
rjeffriesnone of this makes economic sense when the world population of Ben Nanonoye is what, maybe 1,600 by now, tops.w23:38
wpwrakalso one of the reason why i want to get the whole 8:10 thing going23:38
rjeffriesoh I agree UBB is a sweet and simple product23:39
rjeffriesif Tixbarin fabs them he will sell a fair number in Europe23:39
rjeffriesI have no idea what they willcost landed in USA23:39
wpwrakdepends a bit on the price structures of the pcb houses. if they have a big setup fee and then a very light increase (which is the most honest calculation), then fabbing at many places doesn't make sense23:41
wpwrakif they try to make the lower end more attractive, then it might23:41
wpwraki think you need to make at least 1000 units of UBB to reach a good fixed/variable cost ratio. but again, how much of this you see depends on the pricing of your pcb fab.23:42
rjeffriesagree that 1,000 qty will be needed23:43
wpwrak1000 is not a lot in this context. i would calculate 10 units as MOQ for end customers. they should be considered as an inexpensive item that your experiments consume23:43
wpwraknot some precious tool you build rituals around23:44
wpwrakalso, the recommended way of soldering, with gluing the cable to the board, does not encourage much reuse23:44
rjeffriesin BIG qty I am sure this board could cost $00.50 USF or less23:46
wpwraklet me put it this way, if you have N customers, it will cost you almost the same to make N UBBs or 10*N UBBs. except for very large numbers of N.23:47
wpwrakand if N gets very large, the cost of 10 UBBs will be very small anyway.23:47
wpwrakso there's no point of even pricing a single UBB23:48
rjeffriesand totally understand why UBB does not catch wolfspraul imaginatiom23:48
wpwraki think it does23:49
wpwrakit's just not something he needs to do now23:49
wpwraklet someone else take the lead. if there's a lot of interest, he can always jump in, too.23:50
wpwraki would consider the current structure of qi-hw with everything that needs "making" inevitably going through wolfgang as a degenerate case of the model23:51
wpwrakinstead, there should be a loosely coupled net of manufacturers who complement each other23:52
wpwrakwith some amount of competition of course. that's part of the game23:52
--- Wed Feb 9 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!