#milkymist IRC log for Thursday, 2011-10-27

wolfspraulsorry was offline.. ok nothing in backlog :-)01:01
wolfspraulminimilky sounds great01:01
wolfspraulI think I have my work cut out for me for rc4, I can put all my efforts into that.01:02
wpwrakthis looks like a nice place to ... place M1: http://www.soundonsound.com/information01:02
wpwrak"with in-depth reviews of the latest music-making equipment and photos that inspire insane depths of gear lust"01:03
wolfspraulif we think about changing the nor chip, especially if it gets bigger and bigger with microprocessors and memcards etc. then we may also first take a look at the boot options of artix-701:03
wolfspraulwpwrak: yes, thanks! good link01:03
wolfspraulthere were lots of good links here recently, need to catch up01:03
wolfspraulif we find quality magazines, I think we should also try some ads01:04
wpwrakoh, that's soooo traditional ;-)01:05
wolfspraulthat's a way to keep high quality independent journalists fed, so I like it. if we can at least break even, everybody wins.01:09
wpwrakheh, why not sponsor them directly for quality articles. makes sure the money doesn't just fill the pockets of shareholders :)01:14
wolfsprauldon't understand01:17
wolfspraulthat's not how it works01:17
wolfspraulyou cannot 'sponsor' a journalist01:17
wpwrak;-)01:17
wpwraki think they do this with bloggers, though :)01:18
wolfspraulmaybe in China you can :-) but they have no western concept of journalism here01:18
wolfspraulthen there is no reach probably01:18
wolfspraulads in quality magazines are good01:18
wolfspraulthat goes towards the journalists01:18
wolfspraulso far I've found xlr8r.com and de-bug.de01:19
wolfspraulmaybe soundonsound is another one01:19
wolfspraulI read an interview with a driver in a provincial Chinese city who is driving journalists for 20 years. his bottom line01:19
wolfspraul"the foreign journalists come here to work, they want to go to the hot spot right away, are interested in the story"01:20
wolfspraul"our Chinese journalist first want to go to a good restaurant and eat for a long time, then do some sightseeing"01:20
wolfspraulthat's from a driver who is driving journalists for 20 years :-)01:20
wpwrakthe french call this "savoir vivre" :)01:20
wolfspraulwell01:20
wolfspraulthe downside is that the Chinese journalists will write whatever story comes "from above" as their own later01:21
wolfspraulsince they don't know a thing about what really went on anyway, spending their time in a restaurant instead01:21
wolfspraulbut of course, you are right. if they wouldn't do that they would loose their jobs very quickly.01:21
wolfspraulso anyway, I am thinking about ads in quality music magazines01:21
wpwrakwell, why not. i have no idea how they work.01:22
wolfspraulwhat about this minimilky thing?01:23
wolfspraulwas there a renewed discussion?01:23
wolfspraulmaybe I check the backlogs...01:23
wolfspraulI like the idea of making the cheapest possible entry point into the Milkymist world01:23
wolfspraulor pointing to it if it already exists (subsidized dev board)01:24
wolfspraulahh, there's a big thread on the mlist01:27
wpwraki'm a bit worried about M1 and m1 competing against each other. of course, if m1 is in fact M2 (i.e., discard things that turned out to be useless), then there's no problem01:28
wolfspraulnah01:31
wolfspraullet me read the thread first what proposals are swirling around01:31
wolfspraulIf someone is passionate about something, they should just do it.01:32
wolfspraulthe only thing that turns me off is if someone is passionate about someone else doing something, so they have a good time leaning back and watching.01:32
wolfspraulmy focus is on rc4 now01:32
wolfspraulmake it as good as possible01:32
wolfspraullots of ideas :-)01:32
wolfspraulhe, ok01:42
wolfspraulread the thread now01:42
wolfspraulthat feels like DesperateMist01:43
wolfspraul:-)01:43
wpwrakyeah :)01:46
wolfspraulnot even 5% of what is possible with M1 works today01:46
wpwrakexactly01:47
wolfspraullet's just simplify that to "most things people want to do don't work"01:47
rohhm. dont use upper/lowercase to separate the 2. that will fail and lead to confusion01:47
wolfsprauland instead of focusing on that, let's just randomly change every part on the hardware, maybe a better mix comes out in the end? :-)01:47
wolfspraulthe "bom myth" also shows its ugly head here01:48
wolfspraulunbelievabley how you can make people dance around on a "2 usd bom diff"01:48
wpwrakroh: chaos and mayhem, just the way i like it (-:C  btw, i'm still waiting for an opportunity to use "mHz" somewhere01:48
wolfspraulif we wanted to make a cheap milky entry, we could do the exact same board as today, and leave chips and connectors out01:48
wolfspraulI doubt we would sell many though01:49
wpwrakthat would work a lot better. but then, engineers are cheap, right ? :)01:49
wolfspraulI can just say what I do, and that is to:01:49
wolfspraula) continue with rc401:49
wolfspraulmake rc4 as good as possible01:49
wolfspraulb) watch software continuity above all else01:50
wolfspraulnot a 1 USD connector01:50
wpwrakdunno. if you can really make the price collapse to 20% ... (of course, that would probably even be below BOM cost ;-)01:50
wolfspraulc) add features, which are missing on m1 *everywhere*01:50
wpwrakyup. agreed on all 3 points01:50
wolfspraulwe can upgrade hardware (cheaper, better, etc), but without making our house collapse01:50
wolfspraulartix-7 is something we should probably look into01:51
wpwrakone thing you may want to change is the case production. that shop in berlin seems to be just outrageously expensive.01:51
wolfsprauleven just to know the pros/cons of upgrading to it01:51
wolfspraulone thing I know for sure is that I will not swap in shops that use slave labor and poison the environment01:52
wpwrakif you're doing it because you want to direct money to roh, i think everyone will be happier if you just send him a bit of what you're saving by doing it elsewhere :)01:52
wolfspraulnah. there is/was no better way. I am selling M1 for 499 USD and that's a price nobody can beat.01:52
wolfspraulbecause it's subsidized by at least 500 USD a unit :-)01:53
wpwrak(artix) seems they've been announced mere days ago ?01:53
wolfspraulwell, it's being worked on for years01:53
wolfspraula logical next step01:53
wpwrak(subsidized) counting R&D ? or just parts and labour ?01:53
wolfspraulhard to pin down01:53
wolfspraula mix of unpaid work, paid, friends, etc. etc.01:54
wolfspraulyou should know :-)01:54
wolfspraulsince all major M1 hardware bugfixes recently came from you01:54
wpwrakso you really mean that each M1 sold actually makes you poorer. hmm ...01:54
wolfspraulnah01:54
wolfspraulof course not01:54
wolfspraulI have 40 in stock now01:54
wolfspraulif I sell those for 500 USD that's 20k USD and that's an additional 20k USD in my account, of course01:55
wpwrak(hw fixes) naw, adam found a bunch too. he just doesn't write so much about them ;-)01:55
wolfspraulso I think we are on a very good track with rc401:55
wolfspraulI will just continue :-)01:55
wolfspraulcost down - absolutely01:55
wolfspraulwe should accelerate the MMU and Linux01:56
wolfspraulif we want to remove the NOR chip, that will cost *A LOT*01:56
wolfspraullet's say the chip is 10 USD now01:56
wolfsprauland the replacement costs 3 USD01:56
wolfspraulthe difference is 7 USD01:56
wolfspraulahem01:56
wolfspraulshould I even start the math?01:56
wolfspraul:-)01:56
wolfspraulit will take us *thousands* of units in sales to make back our switching costs01:57
wolfspraultotally absurd01:57
wolfspraulso we think we will sell thousands more because our bom cost is 7 USD less?01:57
wolfspraulARGH!01:57
wolfspraulDesperateMist01:57
wolfspraulhow about removing Ethernet?01:59
wolfspraulthe assumption is that people who need it can use a USB-Ethernet dongle?01:59
wolfspraulbut that would require us to massively improve our USB support01:59
wpwrakether is also a size issue01:59
wolfspraulsize?01:59
wpwrakand yes, USB <- Linux <- MMU ;-)01:59
wolfspraulare we looking for things to discuss?01:59
wolfspraulwho was ever worried about size?02:00
wpwrakwell, sebastien brought it up. because he wants to use a small envelope :)02:00
wolfspraulDesperateMist02:00
wpwrakof course, that somehow makes sense. people tend to view product cost and shipping cost as separate items02:01
wolfspraulshipping cost?02:01
wpwrakof course, you could just silently subsidize the shipping and nobody would know02:01
wolfspraulgee02:01
wolfspraulthe #1 way to reduce shipping cost is to work with distributors02:01
wolfspraulforget shipping costs02:01
wpwrakyes, that's even better02:01
wolfspraulno that's the only way02:01
wpwrakbrings up product cost, though02:01
wolfsprauleven if you just ship 10-packs by air, shipping cost goes down a lot02:02
wolfspraulnot to mention how much it would go down with sea shipments02:02
wolfspraulbut that is only possible if you have an active distributor who can stock etc.02:02
wpwrakcustom costs may go up, though02:02
wolfspraulyou can ship 1 container (40 tons) of goods around the world for ca. 500 USD02:02
wpwrakat least in places like argentina :)02:02
wpwrakwow02:02
wolfspraulthe real cost is not shipping, but customs, fees, the last mile, even the unloading at the port because of unions there02:03
wolfspraulsometimes the last 2 miles cost more than the 5000 miles before02:03
wpwraksounds familiar :)02:03
wolfspraulI think we have a good plan for rc4 right now02:04
wpwraknow, how do we get sebastien to channel his enthusiasm into making M1 better ?02:04
wolfspraulany new ideas how to improve it?02:04
wpwrakUSB seems to be pretty high up. or, if he really hates it, maybe an MMU02:05
wpwrakproblem: we lost lars02:05
wolfspraulwhat do you mean?02:05
wolfspraulno I don't think we lost lars02:05
wolfspraulthis stuff just takes some time until a good thought emerges and a good enough code set too02:05
wolfspraulI think lars is very much interested in and following m1, and thinking about ways to contribute/collaborate/hack02:06
wpwraki see two paths towards USB: 1) implement whatever features are needed in RTEMS/FN. 2) make linux work and use that. 2) is of course a lot more complex and riskier, but it avoids sinking work into a dead-end system02:06
wolfsprauloh for sure MMU + LInux02:06
wpwrakwell, lars has a job now ...02:06
wolfspraulbut it's easier said than done - Sebastien cannot do everything himself02:06
wolfspraulhe works for Analog Devices and very much on the same stuf02:07
wolfspraulI'm not sure he wants me to talk about it here, so I'll leave it to him to02:07
wolfspraulmaybe he can even contribute interesting bits and pieces into M1, now in his new ADI function :-)02:07
wolfspraulbut these things take time02:08
wolfspraulI am thinking about what I can and want to do next02:08
wpwraknow, when will ADI make FPGAs ? ;-)02:08
wolfsprauland that's clearly rc402:08
wolfspraulwpwrak: lars_ should speak for himself02:08
wolfspraulbut he's not "lost", for sure not02:08
wpwrakah yes, another "feature": make the HEAD of RTEMS (or such ... not entirely sure in which repo this sits) work. there's a) the hang and b) also some build issue. all this means that you can't work on top of HEAD and have a working system.02:10
wolfspraulsure I know02:10
wolfsprauland we discuss changing this and that chip02:10
wolfspraulbut we can't keep head building02:10
wpwraknot sure if anyone is eager to write new stuff, but in case anyone is, that'll cool them down pretty quickly02:10
wpwrakyeah :)02:10
wolfspraulthere has been no software update in over 3 months02:11
wolfspraulwe hear solid feedback (still pics, usb midi) but have so far not been able to act upon it02:11
wolfspraulnow: let's change some chips :-)02:11
wpwrakthe ability to mess with things on the top would also allow some usability improvements. that would be all sw. maybe someone could get interested.02:11
wolfspraulyes sure, I think that will happen eventually02:11
wolfspraulone by one02:11
wolfspraula review unit is going to a very serious UK music distributor right now02:12
wolfspraulsometimes you have to pick your battles02:12
wolfspraullet's see what those guys say02:12
wpwrake.g., i think M1 should rely a lot more on overlays. basically have a "training" mode where you get immediate visual feedback for all that's happening.02:12
wolfsprauloh there could be so many things02:12
wolfspraulI totally agree02:12
wolfspraulmost likely the feedback will be a long list and mixture of "missing" features02:13
wolfspraulhardware and software02:13
wolfsprauland then what?02:13
wolfspraulwe should pick what is cheapest/easiest for us to do02:13
wpwrakdvi seems to be a done deal. so i guess that'll happen02:14
wolfspraulok, let's go through sebastien's list, my take02:14
wolfspraulyes, the upgrade to dvi-i looks like a really good idea02:14
wolfspraulbackward compatible02:14
wolfspraulopen up new possibilities (of course FPGA will then need to see development)02:14
wolfspraulcheap, simple02:14
wpwrakimage support is a frequently requested feature. that will be messy, because end users expect a certain set of accompanying features. e.g., USB storage, image formats, scaling, aspect ratio handling, ...02:15
wolfspraulof course02:15
wolfsprauland?02:15
wolfspraulit's a pain with those users, right? man. they want stuff to actually work?!02:15
wolfspraulunbelievable02:15
wpwrakUSB MIDI. all the cool kids do USB nowadays.02:16
wolfsprauland they all buy Apple? hmm02:16
wolfspraulwhy?02:16
wpwrakusers are always a pain ;-)02:16
wpwrakwhat was the why about ? users or USB ?02:16
wolfspraulmaybe Apple in the past has sold them stuff that actually worked?02:16
wolfsprauland a lot of the others didn't so they hate all this other crap?02:16
wolfspraulmaybe... :-)02:16
wpwrakit's not just apple. it's the workflow.02:16
wolfspraullet's see in which camp Milkymist is soon02:16
wolfspraulmost normal (non-Apple) tech companies make some random crap, sell it, and when the pain gets too bad they abandon it and move to the next piece of crap02:17
wolfspraulI think it's fair to say that 90% of the consumer facing tech industry functions like that02:18
wolfsprauland they keep themselves fed, so there is some validity in this approach :-)02:18
wolfspraulbut it also leads to the 3 block lines at Apple stores...02:18
wpwrakanother approach would be to implement USB device, run some weird proprietary protocol over it, and have drivers for windows, macos, and so on. then people cold just "drag their images and drop them into the M1 icon" they would like that ;-)02:18
wolfspraulI think we stay calm. What we have today is an awesome product that is great, and people can edit patches and do good things today.02:19
wolfspraulnow we add features one by one by picking the lowest hanging fruits first02:19
wolfspraulon the hardware side, that's what I try to contribute with rc402:20
wpwrakM1 also needs a bit more "best practice information". the closed wiki doesn't help ... but then i don't know if there are already people who would like to post stuff02:20
wolfspraulgradual improvements02:20
wolfspraulyes agreed02:20
wpwrakwhat is almost entirely missing, though, it user community support02:20
wolfspraulno forum02:20
wpwraks/it/is/02:20
wpwrakyup02:20
wolfspraulagreed02:20
wpwrakno place to put your patches. with a preview image. maybe even a small animation. animation with a soundbite. a sound bite that won't get us in trouble.02:21
wpwrakthere's a DEEP dependency chain there02:21
wpwrak(sound bits) idem with video input02:22
wpwraki.e., it would be good to have a collection of WTF/CC0/CC-BY/CC-BY-SA content that people could apply their patches on. then post things in the "patch sharing forum" (whichever form that takes)02:24
wolfspraulsure02:24
wolfspraulall agreed02:24
wolfspraulhe02:24
wolfspraulI just read the original desperatemist mail02:25
wolfsprauldifficult to answer, not sure what the goal is02:25
wpwrak;-)02:25
wolfspraul1. where are we?02:25
wolfspraul2. where do we want to go?02:25
wolfspraul3. how do we get there?02:25
wolfspraulmaybe Sebastien should try with a notebook :-)02:26
wolfspraulRon can help send him some links to cheap ones02:26
kristianpauli think for the developers M1 version a port for others platforms (copyleft hw or no) is not a bad idea, been friendly supporting then well why not?02:47
wpwrakmore platforms = more work02:47
wpwrakof course, if the people there support themselves, then the more platforms, the more glory :)02:48
kristianpaul:)02:49
kristianpaulanyway this is already happening, like the avnet i bougt before M1 was comercially avaliable, and carlobar was porting the soc to its spartan3 board02:50
kristianpaulwell, not consider avnet that was work from lekernel..02:53
kristianpaulbu there is also the papilio board port02:54
kristianpaulwpwrak: well if support themselves are potentials developers, isnt? :)02:56
wolfspraulI think the M1 product approach is right03:05
wolfspraulmake it work, for yourself03:05
wolfspraulmake it fun, then improve. explain to others why you are happy with it, then they may want to join.03:05
wolfspraulany strategizing over tech features won't work03:05
wolfspraulyou can come out with any random amount of random this or that board, with lots of untested and unimplemented features.03:06
wolfspraulM1 implements 5% of what it could do today...03:06
wolfspraulmeanwhile on the hardware side I have 100 ideas how to make things cheaper, easier to manufacture, easier to make improved products, etc.03:07
wolfsprauland upgrading mega-hertz/bits/pixels whenever possible03:07
wolfspraulmore mega more potential (not that the current potential would have been maxed out in any way, imho)03:07
rohi just had a router on my table.. it costs 25E enduser03:09
rohall chips have leads. no bga. no extra fine pitch. 'regular' 05 or 06 micron03:09
rohone side used for components only. etc.. one could really smell the 'cost optimizedness'03:09
rohbut hey.. its easily en par with an old wrt when it comes to performance (well.. duh fresher chips)03:10
kristianpaulopenrisc got linux support, nice for then http://opencores.org/newsletter,2011,10,#n1 "03:17
kristianpauluptream*03:17
wpwrakkristianpaul: (devel) yup, that's the good part. of course, if they stay around and contribute back. there's a certain risk of sinking a lot of time into some student projects with very little return03:28
wpwrakroh: bga is often more expensive than small leaded. sometimes even qfn is more expensive than, say, SSOP. but that's slowly changing03:31
kristianpaulhum patch editor very sensitive to blank spaces..04:31
xiangfukristianpaul, end or begin? there is such code in flickernoise for eat spaces. :)04:35
kristianpaulbetween :)04:36
kristianpaulhum04:36
aw_0x7d: MIDI_RX(TP31) normal level is only 1V which must be 3.3V once powered-on. and dumplock shows no locked; and after used reflashed successfully with './reflash_m1.sh --rc3' version "2011-10-14", then TP31 works @ 3.3V correctly and CRC images tested all pass. It's freakish. My question is: Where's the allocation in NOR about image of midi codes? Inside also in locked block area?07:12
aw_I deleted already my previous reflash script which doesn't fully lock before (0x006c0004) = 0x0001 (1). If xiangfu can write a reflash script without locked function then I can use it again to clarify this weird 0x7d board.07:17
lekernelwolfspraul, so, how can we make a $99 device that supports a decent part of the M1 functionality?07:33
lekernelwolfspraul, looking at consumer devices in this price range, it seems to me that they implement those "desperate" optimizations07:34
wolfspraulthe cheapest way would be to use the existing rc3 production files and leave chips and connectors out07:35
wolfspraulthat can be made even cheaper by improving yield/manufacturability (which we work on as part of rc3), boomification of the bom to make sourcing easier07:35
wolfspraulwho wants to sell such a device btw?07:36
wolfspraulmaybe bearstech could be interested?07:37
lekernelso, it's basically about recouping the investment put into developing a new device?07:38
wolfsprauldon't understand07:38
wolfspraulwho wants to develop something and who wants to recoup investments?07:38
lekernelme. the M1 sales and popularity are disappointing.07:39
wolfspraulstill don't understand - what is the question?07:41
lekernelyou said "the cheapest way would be to use the existing rc3 production files and leave chips and connectors out". this is about not developing new stuff, right?07:43
wolfspraulno that was my answer to your question, I think that's the way to get to the cheapest milkymist entry-level board07:43
wolfspraulwhat do you mean with "developing new stuff" - you mean a new product?07:44
lekernelyes07:44
wolfspraulhow would it differ from m1?07:45
wolfspraulmy action plan is roughly like this:07:47
wolfspraulimprove m1 with every new run07:47
ThihiHow long has the final version of M1 been out?07:47
wolfspraulbetter yield07:48
wolfspraulThihi: 3 weeks :-)07:48
ThihiSo isn't getting disappointed with sales a bit ... premature?07:48
wolfspraullittle things like adv7181c, medium things like dvi-i07:48
wolfsprauloh sure, totally premature07:48
wolfspraulI will develop this product for 5 years :-)07:48
wolfspraulI would like to make the daughterboard a staging area for new extensions07:48
wolfspraulwe can reduce sourcing cost with boom07:49
lekernelyes07:49
wolfspraulwhile trying to improve m1 as much as possible with each run, maintain backward compatibility for all m1 users and deliver updates07:50
wolfspraulif that is not possible, offer upgrade programs07:50
wolfspraulthe most important is to take our customers forward07:51
wolfspraulif they are happy, they will be happy to pay as well07:51
wolfspraulJoerg has proposed a really nice idea to get rid of the metal sheet for ESD safety07:51
lekernelupdates by whom and for whom? there are so few active customers and developers07:52
wolfspraulif you don't believe in your product, then for sure nobody else will either07:53
lekernelalso my email mentioned we keep software compatibility07:53
wolfspraulso that's a chicken & egg problem that only the original guys can break through07:53
wolfsprauloh I like your plan - cheap entry point07:53
wolfspraulI've discussed it in this channel or #qi-hardware before07:53
wolfspraulthe cheap entry point will not make money for anybody however07:54
wolfspraulbecause it's so cheap :-)07:54
wolfspraulbut it's good to have for the community and long-term07:54
lekernelwell, as long as we don't lose money and gain some developers, users and popularity, it's not bad07:55
wolfspraulimagine this: bearstech agrees to do worldwide sales & marketing for a cheap entry board07:55
wolfspraulyes!07:55
wolfspraulthen we move to rc4 now07:55
wolfsprauland in the rc4 run, we make some boards towards this 'cheap entry point' goal07:56
wolfspraulthat is we leave stuff unpopulated, and send a whole batch of pcbas to Bearstech07:56
wolfspraulthat will get really cheap!07:56
wolfspraulsince we can share the pcb & smt & testing costs with the product run07:56
wolfspraulI would leave everything out, even power supply07:56
lekernelI'm not exactly sure what stuff can be left unpopulated though07:56
wolfspraulI wouldn't want to sell it either, that's why I asked that earlier.07:57
wolfspraul"sell it" I mean one by one07:57
lekernelsay we remove the audio chip - then there's no audio at all, contrary to the artix7 solution07:57
wolfspraulyes but your proposed board looks entirely different, why not just go with a Xilinx dev board then?07:58
wolfspraulhttp://www.xilinx.com/products/boards-and-kits/AES-S6MB-LX9.htm07:59
wolfspraulwhat is the problem you are trying to solve with your new board?08:00
wolfspraulfor removing the wolfson codec, I think we should test that first and compare audio quality as Lars mentioned. then we can make an educated decision.08:00
lekernel??08:01
wolfspraulfrom a Milkymist One perspective, of course I like artix-7, evaluating it08:01
wolfspraulalways good08:01
lekernelwe're talking about the cheap entry-level stuff here. audio quality doesn't matter much.08:02
lekernelthe problem I'm trying to solve is how to lower one of the barriers to entry08:02
lekernel_many_ people turn away when they learn it's $49908:02
lekernelit's lamentable of course, but what can i do08:02
lekernelas I said - this entry level thing will share 99% or more of the software with the M108:05
lekerneldevelopments, patches, etc. can be used on the M1 too08:05
wolfspraulI think the way to get there is to focus on rc4, and then branch out into a "cheap devel board" if we find someone who wants to market it08:09
wolfspraulthat cheap devel board would be the rc4 pcba, minus chips and connectors we can identify08:10
wolfspraulif you cannot maintain that compatibility, you need a new board and I think that alone will kill any hope of it being 'cheap'08:10
wolfspraulI think it's extremely unlikely you will arrive at a 99 USD pricepoint, even if you do it's a pricepoint that will block any further innovation, imho08:11
lekernelI can't see any major obstacle to maintaining compatibility08:11
wolfspraulno I mean developing the "cheap devel board" as an offspin of rc408:11
wolfspraulI think it's way too early to judge M1 sales response08:16
wolfspraulmy sales goal was 200 units for 201108:16
wolfspraulwe had a 2-3 months rc3 production delay, unfortunately08:16
lekerneland why not $99? if you look at my proposal, what would cost so much?08:16
wolfspraulso far we sold about 70, 130 to go (we won't make it)08:16
wolfspraulok one by one08:16
wolfspraulyou like the MSL guys in the UK, and Yi had a good impression too08:17
wolfspraulso let's go for it, let's take their feedback serious08:17
lekernelI never said we should stop the current M1 :)08:17
wolfspraulbut you will see - they will most likely ask for a lot of things this and that to be improved in hw and sw08:17
wolfspraulsure sure08:17
wolfspraulwe are just discussing08:17
wolfspraulI like your 99 USD thinking08:17
wolfspraulcost down is good08:17
wolfspraulbut now let's see how we can execute it08:18
wolfspraulI think it should be an offspin of the full product08:18
lekernelwell, it is08:18
wolfspraulalso we can do this: we come out with an improved product, say in January 201208:18
wolfspraullet's say that still costs 499 USD08:18
wolfspraulat that point we can sell the old one for any lesser amount08:18
wolfspraulthat protects our various interests08:18
wolfspraulwith that I mean mostly the customers that went out and paid 499 USD08:18
lekernelthat $99 thing is definitely a M1 offspin. don't be fooled by the artix7 proposal.08:19
lekernelit'll run the same code08:19
wolfspraulyes sure08:19
wolfspraulit's done by you08:19
wolfspraulyou will of course take the same code and port it over :-)08:19
wolfspraulthat's the last thing I'm worried about08:19
wolfspraulif this is what gets us started with artix-7, let's do it just for that08:20
wolfspraulartix-7 eval board08:20
lekernelvideo-in is exactly the same, ethernet too, DRAM too, flash maybe not08:20
wolfspraulwe can look into ddr2 vs ddr308:20
wolfspraulI think we are still safe with ddr208:20
lekernelddr 2/3 will completely break compatibility with the FPGA code using DDR08:21
wolfspraulwith 'safe' i mean that if there is a massive industry 'trend' towards ddr3, we may quickly have to pay a lot of money and face availability problems with ddr208:21
lekerneland it'll be difficult to support the DDR-based M1s08:21
wolfspraulyou mean if we go to ddr3, it will be difficult to support ddr2 boards?08:21
lekernelyes08:21
wolfspraulI think we are still safe with ddr2 for a while08:21
lekerneland the current ddr1 boards08:21
wolfspraul*think*, haven't looked into this in detail08:21
lekernelany DRAM change makes supporting the previous generation difficult08:22
wolfspraulddr1?08:22
wolfspraulok08:22
lekernelyes, the current boards use DDR108:22
wolfspraulah ok08:22
wolfspraulwhat are the goals of the new board?08:23
lekernelreduce cost08:23
lekernelthat's the only goal08:23
lekernelwhile maintaining as much compatibility with M108:23
riaanvddooli have a business idea for cheap milkymist, not sure if you would like the application08:23
lekerneland throwing away some features08:24
wolfspraul"throwing away features" is not a goal, you can do that with m1 already ;-)08:25
riaanvddoolbaby soother - unit playing nursery rhymes with attention grabbing visuals08:25
wolfspraulI think the best strategy is to add new features to m1, and then sell the old versions or pcba only variants at the lowest cost possible08:25
riaanvddoolyou might think i am joking, but baby gadget market is huge08:26
wolfspraultons of markets are huge, but actually doing business in any particular market is not that easy08:26
wolfspraulbecause sooner or later you will notice that there are players in that market already that are damned good at what they are doing :-)08:26
wolfspraul(that's not against your baby soother idea at all, just saying in general)08:27
riaanvddoolso why try any market then? :)08:29
wolfspraulyou try the one closest to what you think you are good at first, 'closest' in terms of technical, locale, channel, personal connections, pure luck :-)08:30
wolfspraulit's a bit naive to think that you can land in any place and any industry, and be 'good' at what you do right away08:30
lekernelwolfspraul, anyway. what is it that you don't like with the entry-level product? competition against the M1?08:31
wolfsprauland I think Milkymist One is on a good path in the creative/video art scene, even if Sebastien is impatient and ambitious, which I can understand08:31
wolfspraulneither "don't like" nor "competition"08:32
wolfspraulartix-7 is good, sooner or later that's an option we have.08:33
wolfspraulI just think super practical about the next steps I am planning08:33
wolfspraulwhich is yield improvements (learn more from rc3), make rc4 as good as possible08:34
wolfsprauldeliver updates08:34
wolfspraulfind serious distribution partners08:34
lekernelso, you see it as a distraction?08:35
wolfspraulphew, good question08:35
wolfspraulnot for me, since I don't work on it. and you are free :-)08:35
wolfspraulwe should be ambitious and explore new ways forward08:35
wolfspraulI cannot answer that question.08:35
wolfspraullast time I brought up artix-7 you dismissed it :-) (it was too early, I admit)08:36
lekernelsure. it will be work to integrate it on the current M1. but for a new design, it makes sense.08:36
wpwrakThihi: (final version) M1rc3 isn't even the "final" version in any sense. there are already non-bugfix changes planned for M1rc408:37
wolfspraulwhat's the difference between the two?08:37
wolfspraulif you make a new design, I would try to integrate it into M1, no?08:37
lekernelhaving it on the M1 would only improve clock frequency a bit at the cost of rerouting PCB, changing power supply, and breaking some compatibility08:37
wolfspraulah ok, you think it doesn't add much to the product08:38
wpwrakThihi: and there hasn't really been a "launch" of M1 either. it'd call this a "fizzle" at best08:38
lekernelno, it doesn't add much08:38
wolfspraulwell, life is never easy, right :-)08:39
wolfspraulwe need a plan to develop Milkymist One the product08:39
Thihiwpwrak, well, for instance, I'm actually interested in the product, might buy one at some point when I have the money, and hang around on the developer irc-channel, and even I didn't know when it launched ;)08:39
wolfspraulfor more than a few weeks :-)08:39
wolfspraulthe feedback we gather so far all point to stuff that can be delivered via software updates08:40
wolfspraulstill pics, usb midi, higher res sometimes08:40
wolfsprauland small bugs like detecting Ethernet cable insertions and lots of similar small improvements08:40
wolfspraulThihi: you made it here, first step :-)08:41
lekerneland do you think those small improvements would solve the large unpopularity problem we have?08:41
Thihiwolfspraul, sure, but that's just since I hang out on other visualizer projects developer channels, and someone talked about this on #projectm. What I mean to say, is that I'm basically very close to the person you want to reach, and the fact that I even know about you guys is by pure slim chance.08:42
Thihiwolfspraul, which is the reason for the "disappointing" sales. People don't know about you.08:42
wolfspraulI am not disappointed about sales.08:43
ThihiYes, but lekernel was in the beginning of this discussion.08:43
wolfspraulin fact we sold them almost at the rate that I was able to make sellable units08:43
wolfspraulyes he is ambitious, and that's good08:43
wolfspraullekernel: there is no guarantee that any feature will solve the "large unpopularity problem"08:43
ThihiI was just pointing out that he fault isn't so much the price (even though it is high, and keeping me from splurging just yet), but that M1 is mostly obscure.08:44
wolfspraulsure08:44
wolfspraulnow...08:44
ThihiHmm, I need to go, I've got a meeting in an hour.08:44
wolfspraulhow to change that? :-)08:45
ThihiYeah, that is the question.08:45
lekernelcheap might help... as it makes viral communication easier08:46
lekernelwe can perhaps use some vaporware too, like the raspberry pi people08:46
wolfspraulI think we have to deliver updates, if you look at successful tech products you will find a massive correlation between the ability and actual execution to deliver updates, and their success.08:46
riaanvddoolit sounds like a case of, if we build it, the users will come08:46
riaanvddoolit will take time08:46
wolfspraulno no, this box works today and is great08:47
wpwraklekernel: which parts of M1 has your cost analysis shown as being critical for causing the USD 500 price ? for the most expensive component, the FPGA, it doesn't even seem to be sure that you will see a significant price reduction.08:47
Thihiwolfspraul, correlation does not imply causation, but I'm with you on the updates anyway.08:47
ThihiDamn, really gotta run.08:47
ThihiLuck to you guys. ->08:48
wpwrak(cost analysis) i.e., do we have a complete cost breakdown somewhere ?08:48
wolfspraulso... I think this strategy has the "highest chance" of success:08:48
wolfspraulspend a lot of energy to develop communication channels with your customers, make ways for them to give feedback, and to spend their money with you08:49
wolfspraulalways always always deliver updates to your users, and find ways to charge for those updates08:49
wolfspraulhappy users will WANT TO spend their money with you08:49
wolfspraulwe are in tech, so we need to aggressively improve our technology, drive specs up, drive cost down08:49
wpwrakone unpopularity problem is that several features are "below par". e.g., i think people who ask for "DVI" probably mean "higher resolution". right now, they get a VGA with a VGA resolution, so they may think (perhaps even unconsciously) that a DVI connector will get them a more modern resolution08:50
wolfspraulthat will allow us to keep the price of our top-end product at a pricepoint that allows us to keep innovating, and discount older generations and sell them to people who really make cost-driven decisions08:50
wolfspraulwpwrak: totally agree08:50
wolfspraulI think it's wonderful how *FAR* we got Milkymist One08:51
wpwrakanother unpopularity problem is that the demo material still looks lame. and we have the bug in the marketing message that we claim M1 can work as a standalone device (i.e., no VJ), which it is boring at.08:51
wolfspraulyes!08:51
wolfspraulso we add features, in hw and sw08:51
wolfspraulthe sw features get delivered to everyone, even first-gen hw buyers08:51
wolfspraulthe hw features require our users to either pay for an upgrade, or buy a new unit. also they allow us to sell older hw generations at lower cost.08:52
wpwrakyou could maybe call unattended operation the "auto-pilot". it won't start, it won't land, and it'll disengage on the least difficulty, but yes, it will allow you to nod off for a second or two08:52
wolfspraulthat is a model I believe in. Does it guarantee success? you bet not!08:52
wolfspraulalso the marketing message of our product needs to be consistent for some period of time (years)08:53
wolfspraulsince we are also building a new brand here!08:53
wpwrakalso, you can't compare the end-user prices of china massware with M1. if you were to manufacture M1s by the millions, they would be cheap, too08:53
wolfspraulwe are building 4 things at once: new brand, new tech, new channel, new category08:53
wolfspraulif we make too many changes, we will confuse everybody about all 4 of those08:53
wolfspraulah, a product like m1 does not exist anywhere08:54
wolfspraultotally new computing architecture08:54
wolfspraulwhether it fizzles or goes on and grows remains to be seen, of course08:54
wolfspraulARM was founded in 1992, I think08:54
wolfspraulwhen did they solve their 'unpopularity problem'?08:54
wolfspraulit must have been pretty lonely at their offices all throughout the 90's08:55
wolfspraulI can only do a small part in all this, and I'm trying08:55
riaanvddoolprobably more of an 'undiscovered problem'08:55
wolfspraulpeople are wisely sceptical of great promises08:55
wolfsprauljust wait a year and see whether that bullshit is still around :-)08:55
wolfspraulI think we all do this with a lot of things08:55
wolfspraulto weed out the nonsense08:55
wolfspraulif we believe in Milkymist, we first of all have to live it and push it forward ourselves08:56
wolfspraulso back to the new board08:57
wolfspraulI learnt something new08:57
wolfspraulartix-7 on m1 doesn't add much value08:57
wolfspraulbut on a totally new board, we could design that board specifically around artix-7 for a better overall mix08:57
wolfspraulmaybe milkymist one is just an over-hyped spartan-6 devboard in the end :-)08:58
wpwrakwhat does artix-7 provide ? is being new all that's to say about it ?08:58
wolfspraulhow can we achieve continuity and growth for the Milkymist SoC on one side, and product and branding continuity for our paying customers on the other side?08:59
wolfspraulno idea, it's just the next-gen from Xilinx08:59
wolfspraulwe could also look at Altera, or even the new ARM-FPGA combos, although I don't know how they would fit into the Milkymist world08:59
wolfspraulsince that next-gen is made in a 28 nm process as compared to the 45 nm s-6, I would think there must be some improvements in there09:00
riaanvddoolsounds like some huge risks in terms of engineering time, going to a new family?09:00
wolfspraulwe are risk takers here09:00
riaanvddoolhehe09:00
wpwraklet's look at more unpopularity problems: where does the request of image support come from ? did rejon make this up on his own ? or did potential customers mention it ? in the latter case, it means that they think ahead far enough that they find that kind of holes in the feature set. and then, of course, reject M1, because it won't scratch their itch09:00
wolfspraulor otherwise we would have never gotten to where we are09:00
wolfspraulwe should watch continuity for the Milkymist SoC codebase09:01
wolfspraulif we believe in the value of that codebase09:01
wolfspraulSebastien says he can keep that with artix-709:01
wolfspraulthat's good news09:01
wpwrakand nobody believes in vendors promising a feature will be delivered after you buy. that meme has been vigorously stamped out by the industry.09:01
wolfspraulno I disagree09:01
wpwrakand in fact, they're right even in M1: lekernel said several times that he won't work on a given feature before there are more sales. classical deadlock.09:02
wolfspraulI repeat from above: *updatability* is one of the top criteria tech investors look at to gauge the chance of success09:02
wolfspraulof course with a web service that's easy09:02
riaanvddoolin terms of communication with users: do you have a mechanism for users or potential users to vote for a feature / bug fix?09:03
wolfspraulno09:03
wpwrak(artix-7) sounds like a bad train to jump on right now then. why be among the early adopters who sink time and money helping xilinx to debug the chip ? will xilinx pay for that ?09:03
lekernelartix7 brings a tad more speed and ADC. that's about it.09:04
wolfspraullekernel: what do you think about the new arm-fpga combos?09:04
lekernelthey bring a lot of proprietary blocks and won't improve much as far as M1 is concerned09:05
wolfspraulok09:06
lekernelsure, they will be MMU and good support from GNU. but it'll make the M1 a boring ARM devboard.09:06
wpwraklekernel: are xilinx looking for companies to work with to make early successful designs of artix-7 ? "work with" as in technical support but also financing ? if you could get xilinx to finance the m1, then the idea would look considerably better09:06
wolfspraullekernel: I fully understand both your cheap entry-level board and "disappointing sales" feedback, but you should also be happy to see how far Milkymist made it!09:06
wolfspraulI am really curious about Simon's feedback, for example09:07
wolfspraulyou can ask yourself what you can do to keep Jon moving forward. Jon is really throwing a giant amount of energy and reputation behind Milkymist.09:07
wolfspraulthere is no guarantee for more success09:07
wolfspraulmore visibility/popularity09:08
wolfspraulthere is not09:08
lekernelthat's actually what is worrying me, because I do not see many returns09:08
wpwrak(arm) an asic core would be nice. but not so much if it weakens the openness of the fpga side09:08
wolfspraulit's difficult to navigate at this stage09:08
wolfsprauleverything is fragile09:08
wolfspraulyou tell me - what is best for the Milkymist SoC?09:08
wolfspraulthe codebase - how to achieve continuity? how to move forward so that we can make the most exciting products in the future, easy & fast? I don't know, learning from you :-)09:08
wpwraklekernel: how good is your german ? do you read technical german articles ?09:08
wolfspraulI can tell you which next steps *for myself* I believe in, and which ones you think I should not do :-)09:09
wolfspraulI roughly listed before. focus on rc4, make rc4 as good as possible09:09
wolfspraulimprove yield09:09
wolfspraulboomify the bom, make sourcing easier09:09
wolfspraulremove metal shield without worsening ESD stability09:10
wolfsprauldeliver updates to existing m1 customers09:10
wolfspraulyep09:10
wolfspraulroughly09:10
wolfspraul:-)09:10
riaanvddoolhttp://en.wikipedia.org/wiki/Diffusions_of_innovations09:10
riaanvddoolwhere do you think this product / brand  is on the adoption curve ?09:10
wolfsprauldaughterboards also, I think daughterboards could be a staging area for experiments and accelerated innovation09:10
wolfsprauland I want to keep selling directly to normal non-technical people who love music and visuals and video art09:11
wolfsprauland I can help people who want to sell & market to hackers and developers, even beyond the "100% free and open files and docs" that we do now09:12
wpwrakfrom a user community side, how to solve the wiki and the forum problem ?09:12
lekernelwith popular products, users sometimes solve this problem themselves by opening their website09:13
wpwrakyes, there are products that have a fan base even against the company's wishes :)09:14
riaanvddoolhardware or software products?09:14
wpwraki guess both09:14
riaanvddooli think it is much easier to form a community arround a opensource software product than a opensource hardware product09:15
wolfspraulriaanvddool: Innovators, definitely09:15
wolfspraulvery early09:15
riaanvddoolyeah, thought so09:15
wpwrake.g., cypress were great at ignoring the customers of their psocs. finally, one guy opened his own site with forum and such. great success. probably helped a lot to get the word out to people who weren't businesses already in contact with cypress.09:15
riaanvddoolso you should probably cater for that market and the next one09:15
wolfspraulthose are multi-billion USD companies09:16
wpwrakyes :)09:16
wolfspraulyeah so the comparison is... :-)09:16
wolfspraulI leave the characterization to you09:16
wolfspraulwe make all sorts of big statements09:16
wolfspraulnow we have to proove them, and we are about to do that09:16
wolfspraulM1 ships09:16
wolfspraulit's real09:16
wolfspraulyou put in your credit card, and the fedex package is on the way to you tomorrow09:17
wolfspraulwe want more? ok09:17
wolfspraulfine09:17
wolfspraulgo for it09:17
wolfspraulMilkymist the new Apple?09:17
wolfspraulsure09:17
wolfspraulgo for it09:17
wpwrakwell, people become fans of the big corps because they perceive there's a benefit for them there. we can play that game too :) M1 is pretty unique, after all09:17
wolfspraulfrom an ambition-standpoint we already made it, that's clear09:17
wolfspraulM1 is absolutely thrilling09:17
wolfspraulSebastien has to guide us on the continuity and what's good for the Milkymist SoC codebase so it remains a world-class starting point to make great products09:18
wpwraklekernel: btw, are you happy with the speed at which M1rc4 is advancing ?09:18
wolfsprauland M1 has to be the first example of such a great product, and I will not abandon it for years, and (try to) deliver updates09:18
riaanvddoolyeah, ride out the S-curve :)09:19
wolfspraulunless say I have a heart attack tomorrow, that'd be bad :-)09:19
wolfspraulforget those curves09:19
wolfspraulfirst time I looked at that now09:19
wolfspraulMilkymist One works09:19
wpwraki wouldn't quite plan on extreme longevity. at least not of the specific device. of the family, yes09:19
wolfspraulyou should get one and enjoy and start learning and being on the ride :-)09:19
wolfspraulwe want to grow sales09:19
wolfspraulfor that it has to be attractive09:20
wolfspraullike you said multi-billion USD corps can be attractive for a variety of reasons09:20
wolfspraulI keep repeating: Milkymist is new on: brand, tech, channel, category09:20
wolfspraulthat's too much for most people, they will dismiss it as some obscure thing with some few strange folks pushing it who will eventually give up and get a life09:20
wolfspraulwhich is an understandable reaction!09:21
wolfspraulmaybe we can remove some of those, if an existing brand or channel would come along, we should consider joining.09:21
lekernelyeah and combine that with a "high" price, and you get the disappointing sales I was talking about09:22
wolfspraul'high' by which standard?09:22
lekernelout of the "cheap tech gadget" category09:22
wolfsprauland you want to be in that category?09:23
lekernelwhose limit is about $100-$150 imo09:23
riaanvddooli dont think the price is too high09:23
wolfspraulyes09:23
wolfsprauland there is no innovation in that category at all09:23
wolfspraulbecause it's impossible09:23
lekernelI think it would be good to have an entry-level product in that category, yes09:23
wolfspraulif it's a subsidized entry point/old version of something more expensive, then yes09:24
wpwrak"cheap tech gadgets" in that price range generally don't have a single USD 30 chip in them ;-)09:24
wolfspraulthat's why I described my approach above09:24
wolfsprauladd features, keep price point at 499, sell old versions for less09:24
lekerneleven if no one makes money with it. as long as we don't lose money, it's fine.09:24
riaanvddooli am siding with wolfspraul09:24
wpwraksubsidizing is an option, yes. another one is huge volume, so you get massive quantity discounts09:24
wolfspraullet's keep a little perspective09:25
wolfspraulin the 80's, people saved money for months to be able to buy their 2000 USD "home computer"09:25
wolfspraulright?09:25
wolfsprauland that was with lower monthly salaries, and then they bought accessories in the 500-1000 USD range still09:25
lekernelyes but "computer" was hype at that time. "milkymist" isn't.09:25
wolfspraulwhy did they do that?09:25
Last message repeated 1 time(s).09:26
riaanvddoolexactly, create the hype09:26
wolfsprauland what are they spending that money on today?09:26
wolfspraulof course 'tech' as such is dead09:26
riaanvddoolthe product is not good enough on its own09:26
wolfspraulpeople don't believe in it, they don't see a bright future somewhere there09:26
wolfspraulmaybe they invest 5k usd in new solar panels on their roof?09:26
riaanvddoolnot this product specifically i mean, in general09:26
wolfspraulmany are surely spending 200-400 USD on a headphone, as I start to understand09:27
wolfspraulI guess they want to treat their ears well09:27
wpwrakM1 isn't really thought of as a "home user" product09:27
wpwrakyou're having marketing blur again :)09:27
wolfspraulI just mentioned some pricing perspective09:27
wolfspraulyou cannot innovate at 50 USD, 100 USD09:27
riaanvddoolpeople wont mind spending $500 for a product that they trust. updates and support will be very important.09:28
wolfspraulamazin is trying at 199 USD, which is a giant bold move, but they have prepared for this for 5+ years, 'kindled', and they are makign this a multi-billion USD bet09:28
wolfspraulamazon09:28
wolfspraulriaanvddool: agreed09:28
wpwrakat least not unless you find someone to sponsor your R&D :)09:28
wolfspraulso no, I think for M1 a 499 USD pricepoint is perfect09:28
wolfspraulnow we need to deliver value09:28
wolfspraulto old and new customers09:28
wolfspraulactually it's amazing about 70 bought already :-)09:29
wpwraki brought up that idea several times. alas, not much has happened there. well, 1.5 contacts. better than nothing, but ...09:29
wolfspraullet's improve m1 asap, and then sell the old version for less09:29
riaanvddooli agree09:29
wolfsprauland we can even branch out a pcba only old version, for the absolut most rock bottom price possible09:30
wolfspraulat or below subsidized dev boards, and without typical dev board bugs, and with Milkymist community support09:30
wpwrakanother way to speed up things would be to find outside investment. that would allow to hire people for software development, maintenance of community services, marketing, trade shows, and so on.09:31
riaanvddoolbut vc companies dont like opensource09:31
wpwrakthat's a problem, yes09:32
riaanvddoolunless you get someone like mark shuttleworth :)09:32
wpwrakhas anyone talked to him yet ?09:32
lekernelmany vc companies I know of have had a lifespan smaller than this project09:32
wpwrakwho cares ? ;-)09:32
wolfspraulthe only way to win over someone like mark is if mark buys a unit or gets one for free, and absolutely LOVES IT09:32
riaanvddoolsure, but trust me if they cant patent it they wont bite09:32
wolfsprauland unfortunately that LOVE cannot be enforced, in no way09:33
wolfspraulwe can only make the best possible product we can, and hope some spark of beauty makes the difference, with mark or anyone else who can help us go to the next level09:33
riaanvddooldon't count on mark though.09:33
wolfspraulbtw, we did another demo of M1 at Lantern today09:34
wolfspraulthe club in Beijing09:34
wolfsprauland it turns out the product is just not ready for the club without still picture support09:34
wpwrakhow did it go ?09:34
wpwrak;-)09:34
wolfspraulit's too boring too fast to just let it run09:34
wpwrakyou can't "just let it run"09:35
wolfspraulI think rejon made the right recision in not pushing the sale through based on personal relationship (aka begging), and isntead opting for what is right in the long run09:35
wpwrakforget that meme09:35
wpwrakM1 is as much an "unattended" device as a hammer09:35
wolfspraulok so what's the result of our discussion09:36
wolfspraulit's difficult to integrate artix-7 in m109:36
riaanvddoolcould having more presets or having them change more often help?09:36
wolfspraulfor low-cost, I propose an offspin of the m1 product, and I need someone else to market and sell it09:36
wolfspraulI suggest we improve m1 the product in hw+sw, keep the pricepoint at 499, and sell old versions for less09:36
wpwrak(artix-7 in m1) extreme design risk for marginal return, it seems09:37
rejonyes, just got off phone with xiaowu09:37
wolfspraulwe deliver sw updates for free to all m1 customers starting from the first rc2 customers, and hw updates allow us to sell new units, or upgrades, or sell old versions for less09:37
rejonyes, better to not have sold to them09:37
rejonthe club wants: image support, basically to cycle through images of their logo and future events...like slides AND we need some solution to make the effects less "winamp like"09:38
rejonless robotic09:38
wolfspraulalso I learnt that Sebastien doesn't think the new arm-fpga combos will be good for us09:38
rejonthe first one needs to happen in software09:38
rejonthe 2nd one maybe can do simply with something like a footswitch and a patch09:38
rejonsoooo, no sale to them until have that solved...09:38
wpwrakwolfspraul: so your take on "m1" is to basically point to the open design files and sebastien's willingness to do all the R&D for free ? that sounds like a salomonic solution :)09:38
wolfspraul"take on m1"? don't understand09:39
wolfspraulI explained my low-cost strategy 5 times now :-)09:40
wolfspraul1. improve m109:40
wolfspraul2. make extra pcba only on the side09:40
wolfspraul3. after we have any improved hw, we can make the older hw versions (without bugs) which allow us to sell them at a lower price without changing the 499 USD pricepoint of our "best" product and keep us in a position to innovate09:40
wolfspraulfor sales, I want to focus on non-tech music & video art lovers, so for a pcba only I need a global sales partner, who buys a whole bunch from me, and then sells them one by one09:42
rejoni agree with wolfspraul09:42
rejonthis type of club sale is fucking huge for us!09:42
rejonfor so many reasons09:42
rejonif we can get the service right, then it becomes much easier sale to others09:42
rejonwe have so much to do on m109:42
rejoni can't even think about something else right now09:42
wpwraki don't really see clubs as a customers09:43
wolfspraulI don't expect any "free R&D" from Sebastien, where do I do that? I am trying to make the best possible product based on Milkymist09:43
rejoni agree wolfspraul09:43
rejonjust need more sales09:43
wpwrakdo clubs own the laptops the DJs use ?09:43
rejonwpwrak: do you go to clubs?09:43
wolfspraulin fact there is a royalty in each M1, and I like to try to pay Sebastien for his trademark & brand09:43
rejonanyway, clubs are great place to sell m109:43
wolfspraulI wish I could find more HDL hackers, trust me09:43
rejonthey see the lights, can slip in m1 logos09:43
wolfspraulJon and I are both trying to find them09:43
rejonits fabulous!09:43
wpwrakevery once in a while. not too often lately, though :(09:43
rejonclubs are HUGE!09:44
wpwrakrejon: M1 is not a "standalone" box09:44
wpwrakit doesn't work09:44
wolfspraulSebastien can learn back as far as I'm concerned. But he needs to understand that right now, he is the only one in the position to quickly deliver a number of software updates.09:44
wpwrakit's boring09:44
rejonwpwrak: that is a software concern09:44
wolfspraulso if he doesn't do that, Xiangfu will try but it will be slooooow, compared to Sebastien. Years instead of days ;-)09:44
wolfsprauland maybe one day we find more active HDL hackers, that would be awesome!09:44
wpwrakI think AI researchers have been saying that for the last 80 years or so ;-)09:44
rejonthe m1 just cycling through images is hard?09:45
wolfspraulbut yes, right now Sebastien is the only one for a lot of features, that's just how it is so that's why we sometimes look at him :-)09:45
rejoni think we can sell the m1 to this club with simply cycling through images09:45
rejonno need for even effects on them09:45
rejonjust like cycling through09:45
rejonthe simplest of support for images09:45
rejonyes, i have lots of ideas on how to make the m1 more standalone and interesting09:45
wolfspraulbut Sebastien did a huge start and we made it to this really unique product and computing architecture, so I'm not complaining.09:45
rejontotally lekernel did shitloads09:46
rejonits just hard work this valley between what we have, and how to connect that up with what people want09:46
lekernelputting them into patches isn't basically much harder, except for the details like decent GUI etc.09:46
wpwrakrejon: why buy an unproven M1 that cycles through images at VGA and it messy to handle if i can just get a netbook for half the price that has DVI/HDMI, 1080p, and runs windows ? :)09:46
lekernelor proper patch language integration09:46
rejonwpwrak: i can't even hear what you are saying09:46
rejoni'm selling m1s09:46
wolfspraulwe try to make the best possible m1 and sell it to everybody! :-)09:47
wolfspraulI want the standy mode removed ;-)09:47
wolfsprauland I want Ethernet cable insertion detection to work so DHCP will kick in by itself at runtime09:47
wpwrakrejon: you're not selling anything but M1s. but your customers can buy other things. that's the disconnect ;-)09:47
wolfsprauland I probably want another 100 things09:47
wolfspraulbut for the time being what we have is great, and sellable, if slowly09:47
rejontotally09:48
wolfspraulthere is nothing even close, anywhere in the world09:48
rejonwpwrak: whatever man, these people are buying the milkymist09:48
rejonthey ahve a great disconnect in the club betweeen club systems and laptops09:48
rejonlaptops are not cool09:48
rejonclub terrain is diff. from outside world09:48
rejonanyway09:48
wolfspraulhomes too09:48
lekernelwell I'm still with you for continuing M1 improvements, but I need to see a clear point in them09:49
rejonthese are completely diff. terrains09:49
wolfspraulwe can sell to people that like video and digital art09:49
wpwrakrejon: you've asked several times product positioning questions of the kind "given that I have an M1, how would i do this or that with it ?". that's a good question for you personally. and for any early adopter who has bought one. but not for someone who still has to decide whether it makes sense at all.09:49
wolfsprauljust needs to be targeted right09:49
wolfspraullekernel: M1 improvements lead to low-cost versions09:49
lekernelI do not want to dig myself into a hole and keep pissing out code 2 people care about09:49
rejonwpwrak: i'm interacting with people daily09:49
rejonwhen i ask questions, i'm usually channeling other people09:49
rejonlekernel: i hear you man09:49
rejonthe image support is most requested by far09:49
wpwrakrejon: i think you need to snap out of that logic. it's good to look a bit on the side, but it's bad to try to sell it as something where it doesn't even begin to be viable competition09:49
wpwrakoh, image support is clearly a good thing09:50
rejonwpwrak: i'm listening to you, but you are not in on all my conversations09:50
wpwrakbut i would be something that enhances the VJ's performance09:50
rejoni save 99% of the speculation from you guys09:50
rejonyou will completely throw me away then09:50
rejonanyway, i don't want to annoy anyone09:50
rejonesp. not lekernel09:50
rejon;)09:51
rejoni got some interviews at an event this weekend about m109:51
rejonnext week is going to be a big m1 week09:51
wolfspraullekernel: I cannot guarantee you that new features will help sales or visibility09:51
wolfspraulbut it's an essential part of the strategy going forward09:51
wolfspraulnew sw features reward our existing customers09:51
rejoni can guarantee that the image support will help fabricatorz who already have one, and will help me sell to clubs09:52
wolfspraulwhen delivered via an effective update mechanism that really reaches them09:52
wolfspraulnew hw features allow us to sell more units to old customers, either as a totally new one or some sort of upgrade program09:52
rejonyes09:52
rejoni want to spam the world everytime we get new features09:52
rejonalso, would love to get more people helping hack09:52
wolfspraulalso new hw features allow us to make the older gen hw and sell them for lower prices *WITHOUT* destroying our pricepoint, business, and future09:52
rejonmore ways to get people hacking, great09:52
rejonnew patches, etc09:52
wolfspraulcorrect, those lower cost (not latest gen) entry points could be really cheap09:53
wolfspraulbecause the latest-gen protects our pricepoint and brand and everything09:53
rejonwhat percentage of our harware is supported in software right nwo?09:53
wolfspraulthat's just how it works, typically09:53
rejonnow?09:53
rejonor rather, how complete is the software implementation to the hardware we actually have on the m1?09:53
wolfspraulrejon: 5% :-)09:53
wolfspraulthat's my number09:53
rejonshit09:53
wpwrakrejon: (club interface) that sounds interesting. why do they resent laptops ? just becasue they're laptops ? or are there more specific reasons ?09:53
wolfspraulthe box, theoretically, could do a hundred times more than it does today09:53
wolfspraulbut where should the engineering come from?09:54
rejonwell, that can be phrased well to customers :)09:54
wolfspraulso we need to pick lowest hanging fruits first09:54
wpwrak(and .. if there's an apple logo on it, would that make it less of an ugly laptop ? :)09:54
rejonby saying: the milkymist is getting better!09:54
wolfsprauland I think we have identified low hanging fruits now09:54
rejonyeah, that is cool09:54
rejonm1 starts up, says, i am 5% ready09:54
rejonhahaha09:54
wolfspraulseriously, I described my proposed strategy 10 times now, it feels. enough. if it's still not clear - check the backlog :-)09:54
rejonupdates can be percentage points09:54
rejonthis is the 6% release09:55
wolfspraulsw updates need to reach existing users09:55
rejon10%09:55
wolfspraulhw updates allow price differentiation09:55
rejonyes, completely09:55
rejonwe can outcompete the others on updates09:55
wolfspraulit all needs to be one product, that needs to be supported in this style (successfuly) for several years09:55
rejonpatches and software09:55
rejonyes, totally agree09:55
rejonnow that the hardware is out, we gotta keep on the grind09:55
wpwrak(sw coverage) i'd say it's maybe 50% from something that's friendly09:55
rejonits not going to be fast09:55
wolfspraul'successfully' includes bigger sales, of course. this idea/approach is meant to establish sales growth.09:55
wpwrakwell, strong and friendly09:56
rejonlekernel: one thing that would be cool, maybe it exists, but are simple ways/guides to get people helping on the milkymist09:56
riaanvddooli wish i could help with stuff like that, but i first need to buy a unit09:56
rejonriaanvddool: cool, get that order in?09:56
rejonare you in netherlands?09:56
wpwrakrejon: each time you install an uopdate, the progress bar changes a little ;-)09:56
rejonwpwrak: totally09:56
riaanvddoolno, south africa09:56
rejonhahaha09:57
rejonriaanvddool: aha, yeah, get one!09:57
rejoni bought a soldering iron today, tried to fix my headphones09:57
rejonbut the shitty headphone plugs were busted09:57
rejonshit09:57
rejonchina!09:57
wpwrakrejon: so, regarding the club IT environment. you do they dislike laptops ?09:57
wpwrakACME - Amazing China Makes Everything ;-)09:58
wolfspraulif we can kick a bit of momentum into the sales, the world is wide open09:58
lekernelI'm fixing a canon ixus65 digital camera I found in the trash right now, it's not that much better09:58
rejonlaptops are an anomaly in dj culture09:58
wolfspraulthere is zero innovation in notebooks09:58
rejonlekernel that is cool09:58
rejonif a dj is worth anything can dj on turntables (not a laptop)09:58
wpwrakrejon: macbooks then. iPad if you must. better ?09:58
wolfspraulthe entire industry except apple has made sure there is no oxygen in the room they live in09:58
rejonat this club, they often use turntables09:58
riaanvddoolrejon: i am going to do a demo at a club this weekend. hopefully will result in sales eventually09:59
rejonthe laptop is just to augment turntables STILL09:59
rejonriaanvddool: that is great! thanks so much!09:59
wolfspraulyou can expect more innovation from smaller companies doing 'obscure' stuff like Milkymist. some of those will become big.09:59
lekernelfinally got the last word with it though :) but it's _full_ of those cheap PCB connectors and soft cables you break soo easily...09:59
rejonwe need and want your help!09:59
wolfspraulbut only if they manage to keep their finances healthy, that's for sure09:59
rejonwolfspraul: i completely agree09:59
rejondid you guys read that article i posted about why apple went with arm09:59
rejonthat was even a few years ago10:00
rejonbut, still intel couldn't innovate fast enough10:00
rejonwell, we are much smaller10:00
rejonwe can pivot much faster than giants10:00
rejonjust have to find the right opportunities10:00
wolfsprauloh totally10:00
rejonso we can pivot uppercut10:00
wolfspraulthe resources that allow us to do what we do are just unbelievable10:00
rejonyes10:00
rejontotally10:00
wolfspraulbut we need more sales and then grow10:00
rejondudes, this club lantern here in beijing, already we spend not really that much energy on10:01
rejonbeen there two times10:01
rejonthey have beceom more of friends10:01
wolfspraulmake Milkymist cool, don't worry about competing with 99 USD cheap gadgets10:01
wolfspraulthey are already dead10:01
wpwrak(laptop = uncool) hmm. but that would be for an un-cool task anyway. i can see that the DJ/VJ wants to be seen doing something. not just be in silent communion with a mysterious box in front of him10:01
rejoni even got tackled in their club last weekend and recovered from that!10:01
wolfspraul(and I am selling 99 USD NanoNotes :-))10:01
rejonstill that is very inexpensive10:01
wolfspraulso the next NanoNote needs to be 99 USD as well, that's tough but that's the goal there. not an easy goal.10:01
rejonso, we try to offer good service10:01
rejonyeah, nanonote as a performance system10:02
riaanvddoolwhat on a laptop competes with m1?10:02
rejonif we could run flickernoise on nanonote10:02
riaanvddoolmilkdrop?10:02
rejonprototype patches, so cool10:02
rejonnanonote is a m1 patching tool :)10:02
rejonhahaha10:02
rejoncool10:02
wolfspraulso rc410:02
wolfspraulanything we can improve beyond what is currently planned?10:02
wolfsprauladv7181c10:02
wolfsprauldvi-i10:02
wpwrakriaanvddool: the case we're studying here is basically a slide show. i.e., when you don't have a VJ10:02
wolfspraulgate & fixed reset ic10:03
riaanvddoolok, because even between songs it is nice to be able to fade out the visuals. and that you cant do with winamp/laptop10:03
wolfspraulwish we could do more, but may too risky or unclear rewards10:03
riaanvddoolso DMX is good feature10:03
wolfspraullekernel: I really understand your artix-7 attack angle. maybe you should try such a board just to learn from it and then we can apply something to m1?10:04
rejonyeah, cool...10:04
rejoni gtg to a meeting10:04
riaanvddoolwolfspraul: why is it called rc4 when you have launched already?10:04
rejonmore m1 showoff10:04
wolfspraulbut please don't think that a few USD bom here and there makes any difference right now10:04
riaanvddoolsorry, i dont have the background10:04
wpwrakriaanvddool: i'm trying to fight the "M1 for slide shows" idea ;-) not that the M1 couldn't do it, but it's a poor choice and distracts from its real strengths10:04
rejonone more thing i think would be cool on m1 is a performance schedule10:04
wolfspraulwhy not10:05
rejonlekernel:10:05
wolfspraulrc1 was a run of 610:05
wolfspraulrc2 was a run of 4010:05
wolfspraulrc3 was a run of 9010:05
rejonlike if can specify general times for things to happen...almost like a sequencer of sorts, or a playlist...fucking cool for artists but also clubs10:05
riaanvddoolok, so rc is not 'release candidate'?10:05
wpwrakhow many VJs are using M1 now ?10:05
wolfspraulin hardware you really always have an infinite sequence of improvements10:06
rejonanother thought for down the line: web server, so can change the playlist, etc, remotely10:06
wolfspraulsince it's analog10:06
rejonanyway, enough ideas10:06
wpwrakriaanvddool: revision code ? ;-)10:06
wolfspraulthe more you study, the more you think you find an even better balance, process10:06
rejonanything i can do to get image support will help...i'm' talking to the air10:06
rejonpre-leave, gtg10:06
riaanvddoolok10:06
wolfspraulthere is no end to this, the higher the volume the more things you discover and find worthwhile improving10:06
riaanvddoolhow about video input for image supprt?10:06
wolfspraulyes, works10:06
wpwrakso, how many VJs are using M1 today ? and what are they saying ?10:06
rejonriaanvddool yes10:06
riaanvddooluse the video input for static images10:06
wolfspraulyou can point the cam to a logo :-)10:07
rejoni even looked at low light cammeras today10:07
wolfspraulwpwrak: maybe 3 or so. it is impossible/impractical to follow up with them one by one10:07
wpwrakrejon: anything interesting ?10:07
wolfspraulpeople are busy10:07
rejonwas seriously prepped to make a halloween doll house with little lights in it to make this sale10:07
wolfspraulno-carrier was in this channel a few times10:07
wolfspraulright after he got his m110:07
wolfspraulwhat is happening now? I don't know10:07
rejonyes, there are some cool solutions, not expensive at all here for cameras10:07
wpwrakyes, no-carrier is the only one i know of10:07
wolfspraulexcitement went away?10:07
wolfspraulthere is one in Munich, one in Australia10:08
wolfspraulbut really, people are busy10:08
rejonbut anyway, too many steps, and they don't want to hire someone to stand there and make a performance in a dollhouse with the m1 camera each clubnight10:08
wpwrakhe's just bee away for days or so ;-)10:08
wolfspraulwe have to make the best possible product, at that point people have to discover their *LOVE* for the product themselves10:08
wolfspraulthis cannot be enforced10:08
rejonyes10:08
rejonok, rushing off 4 real...nice chatting10:08
riaanvddooli think the project needs some sort of feedback aggregation website10:08
wolfspraulthx for continued efforts!10:08
wolfspraulwpwrak: oh sure, I think he is fine10:09
wolfspraulhe planned to take m1 to some festival in Tokyo10:09
wolfspraulsee that's the thing. those guys are not full-time milkymist pr agents.10:09
wpwrakwhat i'm curious about is how people think M1 handles on stage10:09
wolfspraulthey buy this piece for 500 USD, and then they do whatever they like10:09
wolfsprauland that's just fine10:09
wolfsprauland maybe they realize "oh, this is a waste of time", and they silently move it into a drawer and write off the 500 USD because they like us and our spirit10:10
wolfspraulthen what?10:10
wpwraki've more or less come to the conclusion that M1 isn't really useful without an external controller. and there are basically three categories ... hmm, i should actually post this to the list10:11
wolfspraulyes agreed10:11
wolfspraulso we need to talk to those controller companies10:11
wolfspraulreally there are hundreds of business oppotunities10:11
wolfspraulopportunities10:11
wolfspraulbut going after them all needs time10:11
wolfsprauland there is a lot of fake nonsense in the world10:12
wolfsprauland we have a new brand, new tech, new people - all new10:12
riaanvddoolit should be community driven10:12
wolfspraulNOT EASY!10:12
wpwrakhave you looked at faderfox ? some german one-man company.10:12
wolfspraulsure, all good10:12
wolfspraulneeds to be contacted10:12
riaanvddoolwhat is the forum problem referred to earlier?10:12
wolfsprauljust shoot them a 1-line mail if you can10:12
wolfspraulnot installed yet10:12
wolfspraulriaanvddool: yes I agree, best is to build over real users who love the product10:12
wolfspraulthe b2b opportunities emerge out of that10:13
wolfspraulthe real ones10:13
wpwrakwhat i found interesting is that one of his controllers seems to be the only product of maybe a hundred i looked at that seems to come close to what M1 is likely to need10:13
wpwrakwell, or rather, "want"10:13
wolfspraulthe scarcest commodity are users that love your product10:13
wolfspraulonce we have those, many b2b opportunities will emerge10:13
wolfsprauland we have a few right now10:14
wolfsprauland we cannot complain to those few that they are only so few. that's really stupid.10:14
wpwrakthere are all sorts of hacks you can do with other controllers, but that's the only one that has two joysticks. and it still has a decent set of other controls. and it's small.10:14
wolfspraulthat's not their fault.10:14
wolfspraulso let's keep going and grow this thing and get it onto a more sustainable level asap.10:14
riaanvddoolbuild on what you have10:14
wolfspraulsure10:14
wolfspraulk bbiab10:14
wpwrakalas, it's also USB-MIDI. so you still need that PC to translate controller---USB-MIDI-->PC---USB-MIDI-->USB-MIDI-dongle---MIDI-->M110:14
wolfspraullekernel: does this discussion resonate with you in any way?10:15
wolfspraulif not, please holler10:15
wolfspraulwe may be dreamers, and you may be just tougher and that's fine.10:15
wolfspraulI am focusing on rc4 next, and throw all my time and money behind that ;-)10:15
wolfspraulstill need to sell about 40 rc3 units as well, not to forget...10:15
wpwraki'm specifically talking about the LV3 controller: http://www.faderfox.de/mark/index.php?option=com_content&view=article&id=165&Itemid=23910:16
wpwrakhis products get favourable reviews from some: http://www.soundonsound.com/sos/oct10/articles/faderfox.htm10:16
wpwrakothers diss them as "too little, too late" (yes, there's a /. in every scene ;-)10:17
riaanvddoolso what is the special offer price for rc3?10:17
wpwrakbut the main criticism seems to be integration with existing software. not really something we'd have a problem with :)10:17
riaanvddoolwhat is the criticism for M1?10:19
wpwrakM1 ? what i hear most of the time is the price. but then, that's from IT freaks, not from artists10:21
rohin my experience we it freaks are more stiken with cash than the artists10:22
wpwrakwell, one is both. but i'm not sure in which role he responded. also, back then i still tried the "M1 doing standalone effects" meme10:22
riaanvddoolalso: depend if it is a toy, or a business tool10:24
wpwrakroh: or maybe they just have different expectations of what their money buys. a box full of miracles like M1 may look precious to an artist, while the IT geek will just thinks it's yet more cheap china crap10:24
wpwrakriaanvddool: yes, right, exactly !10:24
riaanvddooli think the key is to reach the customer's customer10:24
rohyeah. tricky. maybe its only marketing10:24
wpwrakthe IT guys consider it as a toy. the artists as a tool.10:24
riaanvddoolif they like the effect, then the customer will want to provide the effect10:24
wpwraktricky :)10:25
riaanvddoolhow about trial periods? in clubs?10:25
wpwrakthre's also the problem that effects are hard to make.10:25
riaanvddoolso make the people used to the effects, then take it away, hehe :)10:25
wpwrakseems that many clubs are reluctant to experiment10:25
riaanvddooli can imagine10:26
wpwrakbut then, i can't imagine it to be impossible10:26
riaanvddoolmy pitch this weekend will be for not 'center stage' use, but against some wall or something10:26
riaanvddoolmaybe the ceiling10:26
riaanvddoolor the floor10:27
riaanvddoolnot sure how it will go10:27
wpwraknaw, give the VJ a bit of respect, please :)10:27
riaanvddoolno VJ involved for now, sorry wpwrak, going to try that meme10:27
wpwrakM1 without a VJ is like muzak ;-)10:28
wpwrakriaanvddool: well, get your M1 and try it without affecting the controls, and then with. you'll quickly notice the difference.10:29
riaanvddoolsure10:29
riaanvddoolwill do10:29
riaanvddoolbusy saving :D10:29
riaanvddoolbut still think the price is right10:29
wpwrakshould probably be higher :)10:30
riaanvddoolpossibly10:30
riaanvddoolcould possibly, not should?10:30
wpwrakpossibly both :)10:31
riaanvddooldo you think the see-through case makes it seem more like a tech-toy than a business tool?10:33
riaanvddoolcompare to faderfox's more 'professional' finishing10:34
wpwraki don't know. maybe. for a time, milky see-through was cool. but then it quickly became boring. M1's clear transparency looks nice. (i have the purple version, but it seems blue looks even better)10:37
wpwrakbut it does have a bit of a DIY feel to it, that's true.10:37
riaanvddooli think perceptions count for a lot. it might even 'scare' people off. thinking they would need to know something about the insides to use it10:41
wpwrakdo you really think it conveys that sort of message ?10:41
riaanvddooli think so yes10:41
riaanvddooleven the milkymist catch phrase: eyecandy on a chip10:42
wpwrakfunny :) never thought of that10:42
riaanvddooli think the reference to the chip is in the same category10:42
wpwrakah, indeed :)10:42
lekernelhaha, that catchphrase is very old stuff... actually I think I should update that website logo10:43
riaanvddooli like the idea10:43
riaanvddoolwithout the referenc eto the chip10:43
lekerneldates back from 2007 when I had no other plans than running it on a FPGA devboard (was even Altera at that time)10:43
riaanvddoolsomething like "eyecandy at your fingertips" ?10:43
riaanvddooli think i saw the project for the first time in 2008 :)10:45
riaanvddooli even asked you back then if i would be able to get it runninng on a dsp i had. silly.10:47
wolfspraulthe price of a product is not a joke10:47
wolfspraulbecause people, IT or artists, have alternatives for their money10:47
wolfspraulso the price needs careful 'development' same as tech features do10:47
riaanvddoolsure10:47
wolfsprauland the company that is asking for a particular price needs to spend the money wisely, in the interests of its customers10:48
riaanvddooland itself10:48
wolfspraulno customer wants to see the company they love die10:48
wolfspraullike I said this all needs to be managed and communicated well10:49
wolfspraulI feel 499 USD is just about right for M1, less than that and we cannot innovate, not now or in the future. more than that and we are in a frivolous corner where we don't need to be.10:49
wolfspraulwe can try to bring it down to 399 USD after 12 months, but that's risky. it's better to come out with new better products that can maintain the 499 level.10:50
wolfspraulbecause we believe there is a lot of potential in Milkymist10:50
wolfspraulthat's us first, now we need to get that across to our users10:50
riaanvddoolwolfspraul: what is your take on the packaging?10:50
riaanvddoolie the see-through look10:51
wolfspraulI think it's great10:51
wolfspraulI don't read much into it though, it just looks and feels good10:51
riaanvddoolnot too DIY for end users?10:51
wolfspraulnot at all10:51
riaanvddoolwe should probably ask a non technical person :D10:52
wpwraklekernel: btw, how good is your german ?11:12
lekernellousy11:12
wpwrakcan you read technical german texts ? e.g., articles in c't ?11:13
lekernelgoogle helping, I can usually get the gist of it11:13
wpwrakhehe, okay ;-)11:14
wolfspraulwpwrak: you are around for a long time11:14
wolfspraulhow do you think we can avoid that Milkymist One becomes some sort of one-time obscure hyped devel board?11:15
wpwraki'm asking because i think it may be worthwhile to try to get an article - not just a product announcement - into c't, the largest german it magazine11:15
wolfspraulthe goal must be that Milkymist SoC becomes an often-reused platform, something you look at first when you want to design a great high-performance product11:15
wolfspraulhow to get there? what to avoid?11:15
wpwrakhmm, by creating an active development community and making sure enough hardware is out there. the hardware being open should help also. that's still unproven, though.11:16
wolfspraulthe Linux kernel of SoC designs11:16
wpwrakheh, maybe :)11:16
wolfspraulwe do have one product now, which is amazing11:16
wolfspraulbut...11:16
wolfspraulthat one product is very immature, growing now11:16
wolfspraullet's see where it is in 2 years11:16
wolfsprauland... we need a second, different product. not now, but at some point.11:17
wolfspraulwe = Milkymist11:17
wpwrakfrom my very high-level point of view, M1 is very mature on the SoC side. probably well ahead of all the other (more or less) open rest.11:17
wpwrakat the same time, it's still missing a lot on the "consumer electronics" side11:18
wpwrakright now, you could probably consider it on par with the arduino in terms of accessibility.11:18
wpwrakand yes, product updates are important. my ideal roadmap would go like this:11:19
wpwrak- keep on improving M1: fill the gaps, swap parts we don't like, make small adjustments in the feature set11:20
wpwrak- meanwhile, make the Ya, as a 100% open project with a broad hw developer community from the start. kinda like gta02-core. if we can repeat that alone, that would be a great success.11:21
wpwrak- at some point in time, maybe make a mini2 to try more radical changes. kinda as prototype, but perhaps something you can sell as well11:22
wpwrak- then, do M2 as an open project, drawing from the experience and resources of the Ya, and incorporating features/technology tested in mini211:23
wpwrak- finally, see if we can mix M2 technology and the ya form factor. that should allow for a number of interesting new devices, e.g., for measurement and data acquisition11:23
wpwrakwork with sector-specific specialists to interface with the actual customers (that way, we don't have to learn all the weird bits of strange businesses)11:24
wpwrakon the sw side, run it all under linux. find a major feature-rich distribution to leech on. e.g., debian.11:25
wpwrakonce your community has reached a certain critical mass, it becomes self-sufficient and self-propelling11:27
wpwrakah, and make sure sebastien has plenty of time and motivation to work on llhdl :)11:27
wpwrakopen tools = lower barrier of entry and more flexibility for follow-on developments11:28
wolfspraulsorry will reply in a minute. One other thing that doesn't get out of my mind is whether artix-7 supports new boot options11:29
wolfspraulspecifically, whether we can get rid of the NOR entirely and boot from a memcard, possibly supporting multiple memcards from the SoC...11:30
wpwrakah yes, that could be interesting, yes11:30
wolfspraulanyway, separate issue11:30
wolfspraulok back, reading...11:46
wpwrakone issue is the still extreme barrier of entry. not only is everything based on an exotic operating system (which complicates development but also - and perhaps even worse - deprives you of all the usual system environment)11:46
wolfspraulthe arduino is much cheaper, has far better documentation, easier way to start (IDE), many more samples and starting points, established expansions (shields), training courses, companies making money selling accessories, brand awareness, etc.11:47
wpwrakbut also several of the architectural improvements at the system level require changes to the verilog (e.g., MMU for Linux, USB host controller, ...)11:47
wolfspraulMilkymist is far more powerful, and I think we should develop it along those lines, dealing with music and video data, or other high-performance rf or other signal applications11:48
rohi wonder how complicated usb-host is11:48
wolfspraulother than that I think we can still massively improve documentation, entry points, cheap boards as lekernel suggested, etc.11:48
roheh. sorry.. mean. usb hub support11:48
wpwrakyes, arduino has many differences, too. but what it has in common is that you give people a board, and they're expected to build upon it. not a consumer electronics device.11:48
wolfspraulusb hub is on the "after Linux" list11:48
wolfspraulok I roughly get your thinking, and yes, all sounds reasonable11:49
wpwrak(arduino) if arduino is a kite, M1 is a 747 ;-) but still, i terms of what the customer gets, they are in a similar class for "finishedness"11:49
lekernelwpwrak, have you ever tried building the slightest thing for RTEMS?11:51
lekernelit really goes smoothly. smoother than with linux.11:51
wpwrakonly the FN ball (with the compiler)11:52
wpwrakoh, for building a system image, it may be pretty nice (not sure how much of that i've actually built, though. i used the partially pre-compiled "SDK")11:53
wpwrakbut the thing is that, with linux, i don't even have to worry about system images and all that11:53
wpwraki install a system image that provides the basis for everything else. do i know what exactly my ubuntu does under its hood ? hell, no. and i don't care (well, sometimes i do. but then i do weird things :)11:54
wpwrakonce the base system is in place, i build things with  make CC=whatever-gcc  install them with  scp foo machine:  and run them by logging on and typing  ./foo11:55
wpwrakif i need pass fancy arguments, i make a shell script around it11:55
wpwrakall these things are basically identical for anything linux (i consider the distro-build-centric development point of view an aberration)11:56
wpwraklekernel: (c't) what would you think of trying to have an article there ? someone would have to help with the translation, of course. but if you can more or less read german, that would already help.12:00
wpwrakhmm, i wonder if the NOR is doing the "first block locked" think to me again. ~2700 cycles without corruption. suspicious.12:02
wpwraks/think/thing/12:02
lekernelc't is for little hobbyists, no?12:02
lekernelarduino, makerbot, etc.12:02
wpwraknot at all :)12:02
wpwrakit covers a lot of things in IT. mainly PCs (and macs) and peripherals12:03
wpwrakand software, of course12:03
wpwraktheir focus is quite technical. not extremely hardcore, but still strong12:04
wolfspraulno no. c't is the quintessential magazine for the German tech elite.12:04
wolfspraulor was, maybe. but to a degree still is. maybe the overall audience went down over the years.12:05
wolfspraulit's like the New York Times of German tech journalism12:05
wpwrak(they used to have major hardware projects in the distant past. e.g., building your own 8086 PC, around the time IBM made their 8088-based PC, or, i think, also a modular synthesizer project)12:05
wolfspraulI like ct, and the publishing house. even though they only have ridiculously short rss snippets :-)12:06
wpwrak;-))12:06
wolfspraulI had one of my first reviews that really made a difference in ct, must have been 1988 or so12:06
cxadamsmore data points: my artist friends in Taipei who run a bar like my M112:06
wpwraki think an article on the technology of M1 would fit great into c't. they do cover oddball things from time to time when they like them.12:06
wolfspraulah yes, I wanted to say cxadams ran into a problem last night12:07
cxadamsthey think it's interesting, they want to know more.12:07
cxadamswill they buy one? maybe.12:07
wolfspraulhe had trouble getting the camera to work, but this morning it worked again12:07
wpwrakalso, if they do a tech article, they would probably also put a snippet on the product. plus pimp it all on their newsticker.12:07
cxadamsbut they ask me to bring it back and use it, explore the possibilities12:07
cxadamsright, the camera didn't work last night (i admit I did not try rebooting with camera plugged in)12:07
cxadamsit worked this morning though12:08
wolfspraulso I had a question - how hard is it to detect a composite video signal on all 3 video-in connectors we have?12:08
wolfspraulrebooting should not be necessary, that's a bug that was decisively fixed in rc312:08
wpwrakcxadams: will you do some VJing ? if yes, how will you control the M1 ?12:08
cxadamswpwrak: I'm still learning12:08
cxadamsI can't really find the documentation that describes what I can do with it12:08
cxadamsbut yes, interested in controlling it12:09
cxadamsbut just as interested in being able to set it up and forget about it12:09
cxadamsI don't want to stand behind the DJ station and press buttons all night12:09
wpwraklekernel: btw, regarding control units, yesterday i read the manual of that DMX critter and it looks as if the buttons would basically all be used up by internal functions. with you controller, can you do things like next/prev patch with controller buttons ?12:09
cxadamsjust not my thing :)12:09
cxadamssome patches work way better than others in a club environment12:10
wpwrakcxadams: it's not so much about pressing buttons but about tuning the effects. e.g., the audio levels need to match the music.12:10
cxadamsthe ones that are bright are the best12:10
cxadamswpwrak: aha, tell me more12:10
wpwrakcxadams: if you haven't done such tuning before, you may not even have seen what M1 really can do ;-)12:10
wolfspraulwe need to find a way to rate and category patches12:10
wpwrakit's a radical difference12:10
wolfspraulcategorize12:11
cxadamswpwrak: tell me more12:11
wolfsprauland for those ratings and categorizations to be shared among users12:11
wpwrakM1 standalone = muzak. M1 with someone tuning is = a good DJ in the club ;-)12:11
cxadamswolfspraul: totally12:11
wpwrakwolfspraul: that's step #5 or so ;-)12:11
cxadamswpwrak: how can I learn how to tune it?12:11
wpwrakwolfspraul: step #1: create a place where people can share patches and discuss them12:12
wolfspraulyes12:12
wpwrakstep #2: provide a means to convey their essence to people browsing through them12:12
wpwrakstep #3: make sure the underpinning of that essence is legal ;-)12:12
cxadamswhen we used M1 in Warsaw it was basically the only light source in the room; it was great12:13
wpwrakthat's very nice :)12:13
wolfsprauloh I ran into another question the other day12:13
wolfspraulcan the colors of the video output match the DMX colors?12:13
lekernelwpwrak, no12:14
lekernelI use MIDI12:14
wolfspraulso that when the video output goes red, the lights controlled via dmx also become red?12:14
wpwrakwolfspraul: my reasoning for patch sharing goes like this: you can take a screenshot (and shrink it) for identification purposes. that's sufficient for that.12:14
lekernelwolfspraul, we can do something like that (now), yes12:14
wpwrakbut ... most patches need a bit more to convey what they do. also, if you get incremental developments, the differences may be subtle, which means more need for conveying details.12:15
cxadamswpwrak: so, again, is 'tuning it' documented anywhere?12:15
wolfspraullekernel: what that the answer to my question whether we can detect the composite video-in on all 3 connectors, or about the video-DMX color syncing?12:16
lekernelvideo/DMX12:16
wpwrakso for that you need a short animation. and since M1 visuals coexist with sounds, you also need the music that works with the patch. see how it responds, etc.12:16
wpwrakcxadams: not sure. there are some keyboard controls for video in, but i haven't used them yet. there may be things the remote control can do, but i don't have one.12:17
lekernelI started the "flickernoise handbook" a while ago, but didn't continue lately12:17
wpwrakcxadams: in any case, what you really want is some MIDI or DMX controller, with a lot more knobs.12:17
wpwraklekernel: (DMX vs. MIDI) a MIDI "piano" keyboard ? or something else ?12:18
cxadamsI have the remote control, I haven't even opened it12:18
lekernelyes piano12:18
cxadamswolfspraul: the service that we can offer is that not every M1 customer will be able to figure all this out12:19
cxadamsesp. clubs, galleries12:20
wolfsprauloh you bet12:20
wolfspraulI know12:20
cxadamsso we sell that support12:20
wolfspraulyou are an early pioneer :-)12:20
lekernel...or promote wiimote and laptop based solutions :-)12:20
wpwrak(continuing with the patch sharing) and besides music, you need some video input for showing off the patch, too. now, what sources will people use ? for music, probably some non-open material. for video, it could be non-open (3rd party; e.g., music video or such) or potentially open-able (they captured it themselves)12:23
wolfspraullekernel: is it possible to support plugging the composite camera into any of the 3 connectors and just auto-detecting which one it's in?12:25
wolfspraulcxadams couldn't hit the right spot in the darkness sometimes...12:26
wolfspraulan alternative is to just tape over/lock the wrong ones12:26
cxadamswolfspraul: but I got the right connectors, it just didn't work :)12:26
cxadams'connector' rather12:27
cxadamsnothing to do with darkness12:27
wolfspraulsure, it's a separate question12:27
wolfspraulbut it may come up again, and I don't know the answer :-)12:27
cxadamsthe connector could be labelled more clearly12:27
wolfspraulsince there is an ADC behind, it may or may not be possible depending on the chip12:27
cxadamsthey all say VIDEO IN I think12:28
cxadamsone needs a little camera icon12:28
wolfspraulyeah12:28
wolfspraulwell they are all camera, but all 3 for component video (not sure how well that is tested or working right now)12:28
lekernelwolfspraul, I don't know. check the adv7181 datasheet...12:28
wolfspraulok12:28
lekernelthis isn't tested, needs reprogramming one adv7181 register to switch modes12:28
lekernelno one did it, even though it's rather straightforward12:29
wpwrak(cont'd) so this means that you may not be able to use their example videos, even if they can grab them in some way. but if you could provide known to be free audio and maybe video clips (video could be something coarse, more test image than movie)12:29
cxadamsI'd love to know the median luminance for each patch12:30
wolfspraulI think it will take a long time until we run into the first customer that insists on component video-in12:30
wolfspraulI wouldn't worry about that now12:30
lekernelwe don't have to be anal-retentive about licenses, links to non-free (but OK to use for this purpose) example videos are fine12:31
wpwrakblargh. a water pipe broke.13:13
rohmeh13:14
wpwrakit's not us who are the anal-retentive ones ...13:14
rohwpwrak: your fault or somebody elses?13:14
wpwrakmy pipe's fault13:15
wpwrak(the flexible one feeding the kitchen sink. luckily, only a small pipe.)13:16
wpwraklekernel: hmm, now i'm confused. so you already have OHCI with navre being part of it. and you have a USB stack from BSD. so what's missing then to enjoy all the perverted joys of the world of USB ?13:45
lekernelthe navre firmware doesn't implement OHCI13:46
wpwrak"Navre is here to help you implement this Rube Goldberg protocol called OHCI." ?13:46
lekerneland OHCI will needs a bit more hardware support to generate IRQs (more than the simple system currently in place)13:46
lekernelyes, so you can implement this shit protocol with C programming instead of lower-level Verilog13:46
wpwrakoh, i see. it could potentially do it, but it currently doesn't13:47
lekernelthe idea was to go with OHCI right from the start, but after reading more of the spec I dismissed the idea13:48
lekernelthe current softusb core should let you do OHCI, with some modifications related to IRQs13:49
lekernel(of course, OHCI also made the interrupt controller complicated)13:49
stekernwhat usb stack are you using?13:49
wpwrakah, i see. okay, then OHCI would be the next logical step. does the stack from BSD have things like MIDI and storage ?13:50
lekernelthere's no USB stack, the AVR processor in softusb processes the low-level packets from mouse/keyboard and directly sends the event info to the LM3213:50
stekernaah, ok. wpwrak's comment mislead me13:51
lekerneliirc it has sloppy storage support and MIDI I don't know13:51
lekernelbtw there's a MIDI USB class that isn't more crazy than input devices13:51
lekernelthe fastest way to get MIDI to work is to add support for it into the AVR firmware13:52
wpwraklekernel: didn't you say the BSD USB stack had been ported to RTEMS ?13:59
lekernelit has, but I don't know what drivers14:03
lekernelthe OHCI spec was already a no-go, so I didn't investigate a lot in this direction14:04
stekerndoes it really have to be OHCI compliant?14:04
wpwrakyet you seem to almost have OHCI :)14:04
wpwrakstekern: probably possible to tweak it a little on both ends. though proper OHCI compatibility may be an advantage14:05
stekernI have just written a driver for u-boot for this core: http://opencores.org/project,usbhostslave14:05
stekernand it is not ohci compliant14:05
lekernelwolfspraul, I don't care about artix 7. it would only provide minor improvements. I only suggested it because it can be a bit cheaper than the s6 we use now, and because it's the logical choice for a new design.14:27
wolfspraulI think we should keep the different factors apart14:30
wolfspraulif it's cheaper, then obviously I ask "how much?"14:30
wolfspraul:-)14:30
wolfsprauland the answer is: we don't know14:30
wolfspraulavailability? don't know14:30
wolfspraulso I go back to my booting options. if a-7 allows us to remove the nor flash, that's 10.50 USD less14:31
wolfspraulminus whatever the replacement would cost14:31
wolfspraulbut then I don't understand the focus on bom costs, since our volumes are small and things like yield and pcb/smt overhead are so big in comparison to a few dollars on the bom here or there14:31
wpwraklet alone R&D, which also has an opportunity cost14:33
wolfspraulyes14:33
wolfspraulI'm working towards cheap entry-level board, we get there14:34
wolfspraulI just need hardware improvements so I can breakout a pcba only or last-rev board14:34
wolfspraulif you have a great idea for a super efficient artix-7 board, go for it. we will integrate it back into the whole system somehow, surely...14:34
lekernelI'm not sure a "board" would even be the right answer, even VJs won't use the M1 because they only think about modul8 vvvv etc. and do not want to spend $499 just to try to see if they can learn a new tool14:35
wolfspraulyes go on14:36
wolfspraulyou want to position the M1 differently?14:36
lekernelno, I want a cheap entry level product14:36
wolfspraulyou think VJs are better off with notebook + VJ software?14:36
wolfspraulwho is the product for?14:36
lekernelsame as M1, but for cheap people14:37
wolfspraulhe :-)14:39
wolfspraulI don't understand your argument14:39
wolfspraulother than maybe Milkymist is so bad that we can only sell it if we give a free 10 EUR bill with it :-)14:39
lekernelsimple. $499 for a weird tech product = people do not give a shit14:39
wolfspraulat least those would probably go out quickly :-)14:39
lekernelwhat's the way out?14:39
wolfspraulnah there are tons of products selling for thousands of USD14:40
wolfspraulbut there's a reason for that14:40
wolfspraulhttp://news.cnet.com/8301-13645_3-20123652-47/the-best-headphones-in-the-world/14:40
wolfspraulif you cannot make an attractive product, you are screwed, of course. look for another job?14:41
wolfspraulyou want what? you want to make an innovative product, you want to sell a lot of it, but it's worthless so you can only give it away? huh?14:41
wolfspraulthat won't work14:41
wolfspraulI highly question if M1 sales would be any better if the price were 399 or 299 or 199 now14:41
wolfspraulI just don't know14:41
wolfspraulwe need to find out14:41
wolfspraulit might also make more money if we would sell it for more, 599, 799, 999, etc.14:42
wolfspraulalso we don't know14:42
wolfspraulit's just too early for me. I need to get feedback from lots more vjs, hackers, businesses, journalists, distributors, etc.14:43
wpwrak(we need to find out) a bit difficult with the same product and if you don't have a successor, so that you can sell the old one at a discount because it's obsolete14:44
wolfspraulmeanwhile I can imagine some of the VJs that you think don't like your "weird tech product for 499 USD" are buying, just right now, a headphone for their prized ears north of 1000 USD14:44
wolfspraulpossible? you bet14:45
wolfsprauland why?14:45
lekerneloh, I'm pretty sure of that14:45
lekernelsame with Apple gadgets14:45
wolfspraulbecause they see the value of the 1000 USD headphone, but they don't see the value of the 499 USD M114:45
wolfspraulwell, that's successful business14:45
wolfspraul(for the headphones, not for our neat little M1 :-))14:45
wpwrakit is actually VJs who are complaining ? or hackers ?14:46
wolfspraulnobody I talked to ever complained about price14:46
wolfspraulhow could they when the features are so fuzzy?14:46
wolfspraulif it's boring, they don't want it at any price14:46
wolfsprauland if it's great, they might as well pay more, since they pay a lot for their other tech gear as well14:47
wolfspraulso that's the key question. how good is this product for them.14:47
wolfspraulnotice. emphasis on "FOR THEM"14:47
wolfsprauland there's a gap in m1 here and there. We understand the m1 potential, but our users just press a button and either they see it or not.14:47
wpwrakquite a few people i talked to gave the impression that they may have considered buying it at a lower price. but none of them are VJs.14:47
wolfspraulpeople who say "I would buy at a lower price" may also just use this as a polite way out.14:48
wolfspraulask yourself: which product can you name that you would buy if it were 20%, 30%, 50% cheaper?14:48
wpwraknone of them were asians either ;-)14:48
wolfspraulI could not name a single such product.14:48
wolfspraulevery product I buy is already "price optimized"14:48
wolfspraulotherwise I would not buy it14:48
wpwraktricky question. this isn't how the decision process works anyway.14:49
wolfspraulof course if I can find it cheaper I would save the money14:49
wolfsprauloh sure. that's how you think you can influence people. but it's not that easy.14:49
wolfspraulwe need to do real price testing, not just think "oh, maybe low cost will help"14:49
wpwrakthere are things i'd like to buy but decide against it because of the high cost. then i figure out how to live without them.14:49
wolfspraultotal desperation, this will lead nowhere14:49
wolfspraulsuch as?14:50
wolfspraulI did not say 99% cheaper14:50
wolfspraulof course I would like to fly in a private jet if it would cost as much as my economy ticket14:50
wpwrakso if the price would come down now, i may not reconsider my choice, because i've already spent resources on the plan B.14:50
wolfspraulbut whether the jet is 5 million USD or 4 million USD doesn't really matter to me14:50
wolfspraulwhich product is that?14:50
wolfspraulyou should also think through the other way btw. what if we would increase the price of M1 to 999 USD?14:51
wpwrak(such as) calibrated spectrum analyzer, bigger monitor, a better scope, some extension cards for my USRP214:51
wolfspraulfor sure we would have very nice retail margins then14:51
wolfspraulnah you don't make it real14:52
wolfspraulthat's just a wishlist14:52
wolfspraulsay you have some particular extension card for usrp214:52
wpwrakthese are real purchase-or-not decisions i've made14:52
wolfspraulit costs 500 USD now14:52
wolfspraulbut you will buy it if it were 400 USD?14:52
wolfspraulI would argue you would not.14:52
wolfspraulif it were 400 USD, you would say "I would buy it if it were 300 USD"14:52
wolfspraulin reality you don't need it right now14:52
wpwrakyeah, probably not14:52
wolfspraulif you really needed it, you would also pay 500 USD14:52
wpwrakthat's correct. or i can find excuses for not using it, even though it would be an improvement.14:53
wpwrakbut most purchases are of that kind.14:53
wpwrakthere's almost nothing that CANNOT wait14:53
wolfspraulyou buy something when you feel it's a good thing for you14:54
wolfspraulthat's why people paid 2000 USD for their home computer in the 80's14:54
wolfsprauleven though it could do nothing14:54
wolfspraulanyway14:54
wpwrakin fact, it's more a relation between what i perceive as disposable money, role of the things, and what things cost14:54
wolfspraulreducing the M1 price from 499 now, even dramatically, will kill the product and brand14:54
wolfspraulafter 12 months we can consider a reduction to 39914:54
wolfsprauldepending on feedback from business partners, retail and distributors14:54
wolfsprauland that would be a sign of weakness btw!14:55
wolfspraulthat means we are not able to finance and come up with improvements that allow us to keep the price at 49914:55
wolfspraulour distributors and retailers will not like that14:55
wolfspraulway too early now, I need much more feedback. 499 USD is a fantastic price right now, imho.14:56
wpwrakif after one year you don't have a successor or are at least working on one, then the product is dead anyway14:56
wolfspraulI will think about it more though, maybe I'm wrong :-)14:56
lekernelI never said we should stop the $499 M1. again: I proposed a cheap entry level product IN ADDITION to the M114:56
wolfspraulyeah but how to market that14:56
lekerneland yes, media blackout/unpopularity = massive weakness14:56
wolfspraulI already said - you do that with the LAST GENERATION typically14:56
wolfspraulgee maybe it's unpopular for a reason14:57
wolfspraulyou cannot force others to see the world your way14:57
wolfspraulyou may try, but you see the reality now :-)14:57
wolfspraulfor low-cost, the right way to do that is with last-gen imho14:57
wolfspraullet's ask Simon and other serious potential business partners14:57
wolfspraulI like the "499 USD screen saver" comment on slashdot14:59
wolfspraulmade me laugh14:59
wolfspraulignorance is bliss14:59
wpwrakthat's what you get for marketing it as a device that will do nice things when unattended14:59
wpwrakof course it's a screen saver !14:59
wolfspraulhttp://en.wiktionary.org/wiki/ignorance_is_bliss15:00
wolfspraulk calling it a day15:00
wolfspraulI love my M115:00
wolfspraulI'm sorry for people who think it's not worth 499 USD15:00
wolfspraulmay they happily spend their money elsewhere15:00
wolfsprauland don't tell me what it is, please :-)15:00
wpwrakgiven the current volume, does it actually cover costs ?15:01
wolfspraulhard to say, not really15:01
wolfspraulbut I get very good feedback from people who buy one15:01
wpwrakso the price is already subsidized in a way ? i.e., it costs you more to make one than you earn with it ?15:01
wolfspraulthere just needs to be someone among them who comes back with a bigger order, or bigger real business15:01
wolfspraulno no15:01
wolfspraulnot at all15:02
wolfspraulall numbers are public15:02
lekernelwell... the situation isn't totally dramatic: of course, for a very small minority, it's great15:02
wolfspraulthe free part are the working hours15:02
wolfspraulI am not different from you or Sebastien or Jon or others in that regard15:02
wpwrak(cost to make) bom, fabs, logistics, share of adam's time, share of wolfgang's time, sebastien's share of the revenue, etc.15:02
wpwraklekernel: i think it's way too early to say who things will go15:02
wolfspraulsebastien's share is so tiny and he is so reluctant in even having me pay it, basically we can say it's also zero15:02
wolfsprauleven a Chinese garbage collector would cry at that kind of hourly wage15:03
wolfspraul;-)15:03
wolfspraulthis box needs to sell thousands of units, and sooner or later it will15:03
wolfsprauland now I go to sleep :-)15:03
wolfspraulhttp://en.wiktionary.org/wiki/ignorance_is_bliss15:03
wolfsprauln815:03
wpwraklekernel: also consider this: about half of the M1rc3 made are already sold - within ~3 weeks. it will probably be ~2 months until M1rc4 is available. you probably have too much demand already ;-)15:03
lekernelgn815:03
lekernelwpwrak, the hype surrounding things like the raspberry is what I would call "too much demand"15:05
lekernelunfortunately we're nowhere near it, even though the M1 isn't vaporware contrary to the raspberry15:06
wpwrakokay, but they are toys15:06
wpwrakand 99.9% of the people who think these toys are cool won't buy them anyway15:06
wpwrakbut they like the idea of a computer that costs almost nothing. of course. everybody does. how many OLPC have you bought lately ? ;-)15:07
lekernelnone. but at least OLPC also had a decent share of media exposure.15:08
wpwrakoh, you can have that, too. wait until the pope visits the next time, then run naked in front of the car. voila, you'll be in all the newspapers. with only 10 seconds of work ;-)15:09
lekernelhttp://www.google.com/trends?q=olpc%2Cmilkymist15:10
lekernelhttp://www.google.com/trends?q=%22raspberry+pi%22%2Cmilkymist&ctab=0&geo=all&date=all&sort=015:12
lekerneletc. etc.15:12
lekernelthose graph are galling15:12
wpwrakbreak into steve job's home, put an M1 under a hard to reach but not completely inaccessible place, leave again without leaving other traces. wait. when they find it, the mystery device will be in all the news.15:13
wpwrakhttp://www.google.com/trends?q=%22paris+hilton%22%2C%22albert+einstein%22&ctab=0&geo=all&date=all&sort=015:14
wpwrakyou're in the wrong biz :)15:14
lekernelhaha, good point15:16
wpwrakmedia attention is poor, agreed on that. but i'm not so sure it's the media's fault15:19
wpwrakfirst of all, the last "launch" was weird. confusing exclusive article on the reg and no announcement for others to follow up on. not even on our mailing lists.15:20
lekernelwell, after being ignored by a dozen other news sites, I finally got it reposted on /.15:21
wpwrakalso the marketing message keeps on shifting. once it's a tool for VJs, the next day it's a screen saver or slide show, then a cheap toy for developers, ...15:21
lekernelmh?15:22
wpwrakyes, good to get it into /.15:22
lekernelI only get the last part of what you are talking about, and this was an internal message on a mailing list with ~80 subscribers15:22
wpwrak /. has at least some coverage15:22
wpwrakthe screen saver is what's been making the round lately: try to sell M1 to clubs/bars/etc. as a device that will provide visuals without a VJ15:23
wpwrakthe problem is that it doesn't work like that, as anyone can see who actually tries to demo it :)15:23
wpwrakso i've shown it to a friend who is also a hobby musician, actor, and who knows what else. in my office, i think with sound coming from the radio or my kaossilator. no adjustment of controls. wasn't very impressive.15:25
cxadamswpwrak: have you ever used it with a real sound system?15:25
wpwraka few days later, i showed it to some more friends, on the TV. this time i adjusted audio sensitivity. (messily, by changing the volume of the PC that was playing it, and then compensating on the TV that had the speakers. the M1 in the middle.)15:26
wpwrakthere it already looked much better. some of the effects came to life. but of course, this was senselessly geekish. (the audience were geeks.)15:27
cxadamsin warsaw we had girls dancing in front of the projected light15:27
cxadamsjust saying15:27
cxadamsyou have to know how to throw a party, to use the M115:28
wpwrakcxadams: no. i'm trying to figure out what it takes to reach the point where i could tell a friend "take this to a club and you won't make a complete fool of yourself"15:28
wpwrakyes, girls are important. but before i invite them to dance i want the controls covered. and i'll have to give the camera issue a closer look. i'm not very happy with the images i get so far.15:29
wpwrakplus, i know that without better controls, the camera is very difficult to handle. maybe even with better controls, but i can't tell yet15:34
wpwrakhmm, we'll have new moon on december 24. that could be an opportunity for a M1-enhanced party :)15:40
wpwrakand i also need to do something about the firing control for my firework. but perhaps i won't automate these. remote electrical ignition alone should do.15:41
wpwraklekernel: btw, one big blocker for contributions, at least from my side, is that flickernoise HEAD doesn't work. (if i understand things properly, there's first a build issue, but even if i'd overcome this, i'd then hit the RTEMS hang)16:35
lekernelthere's no build issue that I know of, and I actually find it very annoying that this bug has been around for months simply because I wasn't there to fix it16:36
wpwraklekernel: e.g., i my have tried to clean up the parser, but since this touches a lot of files, wouldn't do such a thing on an old branch. (i don't think i'll have time for messing with the parser now, so it's not that urgent an issue at the moment. but of course, others may bump into the same.)16:37
lekernelthe only new thing in HEAD is RTEMS upstream support (which also introduced the lag bug)16:37
lekernelthere aren't a lot of changes, and none that touches the parser16:38
wpwrak(nobody else fixed it) well, there may not be a lot of RTEMS experts around here :)16:39
lekerneland even fewer Linux experts for this kind of problems16:39
wpwrakso you're saying that one should simply forget about HEAD and all further development will be based on stable_1.016:39
wpwrak?16:39
lekernelno, we should just fix that fucking bug16:39
lekernelI have no idea of the difficulty, I have simply not looked into it16:40
wpwrak(linux experts) hmmm, you may greatly underestimate the expertise available for chasing weird issues in linux :)16:41
lekernelcome on, Linux caused gazillions of problems like that, and then some more. I didn't switch to RTEMS for no reason.16:49
lekernelthe only way you can change my mind about this is by demonstrating a fully working Flickernoise on Linux16:50
wpwrakyeah, that still has some prerequisites. i wish i had the resources to work that, by the way, but my time is running out, too.16:51
wpwrakone thing that helps with linux is that there are a lot more people around who can tell you where things go wrong as long it's not a problem in your platform-specific code16:52
wpwrakplus, many people can fix things. because it's something they've done before. even in the kernel.16:52
wpwrakwith a new OS, even if it's super-friendly, you need to learn many things again and you'll lose time making mistakes because you overlook some differences16:53
lekernelBTDT and now I'm happy with RTEMS.16:53
wpwraknow we just have to perfect the vulcan mind meld and everyone else will be able to benefit from your experience and happiness with RTEMS :)17:01
wpwrakah well, enough bickering. time to go and look for a controller.17:02
wpwrakoh, by the way, one last thing: if you want to bring M1s to developers, maybe consider selling production dropouts that have only a relatively minor flaw. (not sure if there are any, but may be worth a consideration)17:08
stekernare there any plans of selling any "board only" packages? i.e. without the case and the silicon keyboard etc.17:14
stekernI know there was that early adopters version, but I guess I missed that train ;)17:15
wpwrakoh, i found the MIDI bindings ;-)19:13
wpwrakso all file-based ?19:16
wpwrakhmm, and tweaking patch parameters is hard-coded to DMX ?19:18
wpwrakhmm, seems that only patch selection is in the GUI, but no parameter control is somewhere else. need more research :)20:47
--- Fri Oct 28 201100:00

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