#qi-hardware IRC log for Thursday, 2010-10-07

wolfspraulwhy too late?00:00
aisaI thought you wouldn't be around :-)00:00
wolfspraulmirko verified the permission/owner issue you brought up, and can't seem to find it00:00
aisaas in, verified it isn't there?00:00
wolfspraulit seems00:01
kristianpaul7win 1900:01
wolfspraulso maybe you can walk us exactly through your steps and we see where things go wrong?00:01
aisaHow about this: I'm going to be writing a new image probably over the weekend,00:01
aisaI have a package to check in before.00:01
wolfspraulthe permissions and ownership are supposed to be copied over as they are00:01
aisaand I will take meticulous notes the the process I used to build an image.00:01
aisaIt is good to hear that the problem was not recreated.00:02
wolfspraulaisa: I finished the news/feed cleanup I was talking about00:03
aisaI had a good conversation a couple days ago with someone on openwrt, and have since spent more time exploring how to build packages.00:03
wolfsprauland wrote a rather lengthy mail to the discussion list yesterday describing the process in the wiki00:03
wolfspraulmostly for you :-)00:03
aisaAh, I have not seen it yet.00:03
aisaI have a project due tonight that I've been focused on,00:03
aisaso my e-mail has backed up.00:03
aisahowever, this is great news.00:03
aisaI will follow up over the next two days.00:04
wolfspraulsure, take your time00:05
wolfspraulthings are moving00:05
aisaI can't find this e-mail, what was the subject line?00:05
aisais it: upstreaming progress, xburst-tools (a little bit) in Debian00:05
wolfspraulno reply to Allin Kahrl, open source hardware summit debrief00:06
wolfspraulI should have changed the subject line, forgot00:06
aisanow I see it.00:07
aisawow, I had not read this.00:07
aisaI'm glad you brought it to my attention.00:07
qbjectHi all.00:14
aisahello qbject00:15
aisawolfspraul: this is absolutely amazing, I didn't know mediawiki could do this.00:15
qbjectHm? What's mediawiki doing?00:18
aisahttp://lists.en.qi-hardware.com/pipermail/discussion/2010-October/005707.html00:18
wolfspraulaisa: that was the last one?00:19
wolfspraulqbject: you mean a problem with the mediawiki on qi-hardware.com? which problem?00:19
aisaThat links describes how to add news to the wiki.00:19
qbjectWolfspraul: no problem. I just stumbled in on aisa saying "I didn't know it could do that."00:21
wolfspraulmediawiki is great00:21
wolfspraullots of extensions00:21
wolfspraulof course our installation is also well maintained :-)00:21
wolfspraulhttp://en.qi-hardware.com/wiki/Special:Version00:21
aisa:-)00:21
qbjectSo few are!00:21
wolfspraulyes but ours is00:22
qbjectAgreed.00:22
wolfspraulthe setup is documented here http://en.qi-hardware.com/wiki/Server_setup#MediaWiki00:23
wolfspraulaisa: what was the last line you saw from me after the URL?00:23
wolfspraulI wrote a few things, then got disconnected...00:23
aisathe setup is documented here...00:24
aisawith the Server_setup page.00:24
aisaI'm really happy the Server_setup page gets so much attention.00:24
aisaThat is a wealth of information.00:25
wolfspraulI want the servers to be 100% transparent00:25
wolfsprauland downloadable too, still stuff to improve there00:25
wolfspraulaisa: no I meant above, about the news feed stuff00:25
wolfspraulI'll just repost what I think might have been lost00:26
aisakk.00:26
wolfspraulover the next few months, I want to work on some really high quality Qi news00:26
wolfspraulmost importantly monthly community news summarizing what's going on00:26
wolfspraulvery important - with nice quality embedded pictures or videos00:26
wolfspraulthen maybe nice editorials or interviews, if we have something - again quality & originality before quantity00:26
wolfsprauland finally press releases, i.e. something we think other blogs/media can pickup, and written in such a way that they can easily do so00:26
wolfspraulI think the infrastructure in the wiki for all this is now complete00:26
wolfspraulno rush on this, quality (by far) before quantity00:26
wolfspraulbut a nice steady stream would be good00:26
wolfspraulnot too much/frequent, not too infrequent00:26
wolfspraulcommunity news every month or two00:26
wolfspraulpress release once a quarter or so00:26
wolfsprauleditorials or interviews when we have something meaningful to say00:26
wolfspraulwhat do you think? that's the plan... :-)00:26
wolfspraulthat was the continuation of that 'how to write news' post on the discussion list I showed you00:26
wolfspraulI think I can get a number of journalists to subscribe to the Qi RSS feeds, the originality and quality of what we have is not bad, and if we collectively work on this a bit it's going to be very helpful for them00:27
aisaI saw none of this :-)00:27
aisaI feel that I can do my part to contribute news and stories,00:27
aisaand spend some time aggregating,00:27
aisaand that is what I'll have the energy to do.00:27
wolfspraulgreat00:27
wolfspraulno rush, it's about quality00:27
wolfspraulI have good relations with a number of journalists00:27
wolfspraulin the US, Europe, Japan00:27
wolfspraulthey are all under insane pressure, no time to do anything!00:28
wolfspraul20 minutes for a great article with pictures and all00:28
wolfspraulhow can anybody work like this...00:28
qbjectWolfspraul: i'd like to help with this too.00:28
wolfspraulso what we need to do is to create a nice steady stream00:28
wolfspraulno need to do a lot, not at all00:28
wolfspraulquality is far more important than quantity00:28
wolfspraulthey are flooded with crap anyway00:28
wolfspraulevery day00:28
wolfspraulI am not so worried that our stuff will not stand out in the flood of PR/marketing blurbs.00:29
wolfspraulI'm only worried that we:00:29
wolfspraula) do have original and high quality news00:29
wolfspraulb) do a bit of work to make the life of journalists easier, with high quality reusable pictures, for example, and well written texts00:29
wolfspraulc) do not post too much, overloading their capacity to follow, and also not too little, making them forget about our project00:29
wolfspraulthat's all00:29
wolfspraulif we can do this, I am quite optimistic journalists will love to subscribe to our feeds, follow, and reuse our content in their publications00:30
aisaI like this guidance.00:30
qbjectSounds reasonable, doable, and effective.00:30
wolfspraulmany journalist like content where they feel their readers can make a genuine contribution to00:30
wolfspraulbecause readers want to read stuff that let's them get active/contribute in some way (whatever way that is)00:31
wolfspraullet's keep that in mind00:31
wolfspraulso we don't need to throw our greatness in front of their feed00:31
wolfspraulfeet00:31
wolfspraulthings can be unfinished, unpolished00:31
wolfspraulbut the readers must somehow feel how they can contribute to this, in which way, no matter how small00:32
wolfspraulafter that it's a bit of luck to get published, it depends on what is happening on any single day00:32
wolfspraulthat is uncontrollable by anyone00:32
wolfspraulso you can plan whatever you want for day X, if on that day another Tsunami hits, forget about your stuff00:33
aisaheh.00:33
wolfspraulon the other hand, on a slow day, they may be under pressure (they always have to publish something), so even a bit obscure news may make it to the headline on such a day00:33
wolfspraulthat part is uncontrollable by anyone00:33
wolfspraulso let's not focus on that00:33
wolfspraulwe focus on a steady, readable stream of high quality, original content00:34
wolfspraulover time that should sort itself out00:34
wolfspraulnow that I think about it, maybe I should make the News page in the wiki editable by admins only?00:34
wolfspraulthat's the one that aggregates wiki articles into the news feeds (Qi & planet)00:35
wolfspraulI mean anybody we know to be a reasonable person can get admin rights immediately anyway.00:35
wolfspraulbut right now, I just realize, even anonymous people could push things through the news feed :-)00:35
wolfspraulmaybe that's a bit too extreme00:35
aisaGiven the intent of the newsfeed, I agree with you.00:35
aisarestrict the access.00:35
qbjectIf your going to court serious folks, it would probably be wise. No need to be embarassed by publishing vadalism.00:36
aisathe option option is to wait until it is a problem,00:36
wolfspraulaisa: is your account an Admin account already?00:36
aisaif you don't like the idea right away.00:36
aisaI am an admin, yes.00:36
aisaI've used my powers once :-)00:36
wolfspraulqbject: what is your wiki account? I'll make it an admin too00:36
wolfspraulno I think inclusion in the feeds should be restricted to admins. we are very liberal about the admin rights anyway.00:36
qbjectWolfspraul: Allin_Kahrl00:36
wolfspraulahh! :-)00:36
qbject;)00:37
aisahaha.00:37
aisaI had no idea.00:37
aisayou jerk.00:37
aisathough, I guess you might not know aisa is Alan Post.00:37
qbjectNo, but I suspected. :D00:37
aisaqbject, what are you even doing awake, you have work tomorrow ;-)00:37
qbjectYou're telling me. But look what I would've missed.00:38
qbjectAnd I just came on to see if the openwrt version of imgv could do slideshows.00:39
wolfspraulalright, I set the News page to only be editable by admins00:40
wolfsprauland I gave admin rights to qbject00:40
qbjectThanks wolfspraul.00:41
qbjectAisa's right. I should bed down for the night, but I look forward to discussing this further maƱana.00:43
aisaI just finished a data analysis myself.  After I summarize the results I'm way overdue to sleep myself.00:44
aisagoodnight qbject.00:44
qbjectG'night aisa, wolfspraul.00:44
wpwrakwolfspraul: a/b) have news that stick out ("original" in the way it's used in German - novel, exceptional). the journalists are only the gatekeepers. the news also have to pass the "spam filters" of the readers.01:53
wpwrakwolfspraul: interesting that journalists "get" the readers' desire to participate. that's encouraging.01:53
wolfspraulwpwrak: what is the killer application of the Xue camera?03:35
wolfspraul:-)03:35
wolfspraul(stealing your question...)03:35
wpwrakwolfspraul: wish i knew ;-)03:35
wpwrakwolfspraul: a distinguishing feature is the openness03:35
wolfspraulmethril_work was asking about the difference to Elphel a long time back on the mailing list, and I never answered (nobody did)03:36
wpwrakbut is there more ?03:36
wolfspraulhonestly I don't know03:36
wolfspraulso I can also just guess/try to dive into the idea03:36
wolfspraulif you look at a CMOS image sensor, my understanding is that it is always basically made out of 2 parts03:36
wolfspraulthe cells/pixels that are capturing the light, and additional logic that processes the data03:37
wolfspraulso in terms of freedom innovation, the idea is to lower the entry point at which free software starts03:37
wolfspraulwe wouldn't want a sensor that just spits out complete jpegs03:37
wolfspraulnow, we wouldn't want a sensor either that has zero processing logic at all, first because afaik such sensors don't exist, and then also because that would make the level so low that we wouldn't have a functioning camera for years03:38
wpwrakyup, understood. but is it "just" an openness improvent or is there anything else xue can do/do better than others ?03:38
wolfspraulanother way to look at it is from the application side03:38
wolfsprauloh many things03:38
wolfspraullet's look from the app side03:39
wolfspraulI'm in beautiful China03:39
wolfspraulthe country leading on spying everywhere03:39
wolfspraulwhen I drive around Beijing, I am followed by hundreds of cameras03:39
wolfsprauleverywhere03:39
wolfspraulthey are also at every toll booth03:39
wolfspraulat every entrance and exit of a public parking lot03:39
wolfsprauland at every other place you could imagine putting one03:39
wolfspraulafaik those cameras are now able to identify up to 30 license plates in parallel03:40
wolfsprauland they are working on fingerprinting faces as well03:40
wpwrakand xue can do more ?03:40
wolfspraulhe he03:40
wpwrakor identify occupant(s) ? :)03:41
wolfspraulwell you cannot do any of this with a sensor that can only give you 3, or 5, or even 20 jpegs per second03:41
wolfspraulso there are many applications that are, at least today, only possible if you move the application logic lower03:41
wolfspraulfor example into an fpga03:41
wolfspraulElphel could probably give you a long long list of such applications03:41
wolfspraulessentially all their customers come to them for that03:41
wpwrak(20 fps) why not ? have mor than one plate in the image03:41
wpwrakso, faster license plate recognition is the killer app ?03:42
wolfspraulno03:42
wolfspraulany application that is only possible to moving the processing into an fpga03:42
wolfspraulby moving03:43
wpwrakalright, but what's the proof of concept ?03:43
wolfspraulwell first take any picture, even just a still picture03:44
wpwrakalright. that's R&D :)03:44
wolfspraulthe problem on the freedom side is that the CMOS image sensor will always have pixels + processing03:44
wolfspraulwe will not be able to make our own CMOS image sensors for xxx years03:44
wpwrakwe were talking about marketing. why should i, a hypothetical prospective customer, want a xue ? and not happycam.cn ?03:44
wolfspraulso essentially we are still coding onto a proprietary API, the one for whatever CIS we are choosing03:45
wolfspraulwell again, I think there are many applications03:45
wpwrakwell, many sensors provide fairly raw images03:45
wolfspraullet's think more high level03:45
wolfspraulhow programmable/scriptable is the typical mass-market camera?03:45
wolfsprauleven for simple things, say you want to synchronize 3 cameras03:46
wolfspraulto take a picture at exactly the same time, say within 1 ms03:46
wolfspraulhow do you do that?03:46
wolfspraulwith our camera, no problem at all03:46
wolfspraul(I would think :-))03:46
wpwrakokay, synchronization03:46
wolfspraulwe can do real-time encoding in Ogg Theora, or other formats, even new formats03:47
wolfspraulwho else has 'hardware accelerated Ogg Theora encoding'03:48
wolfsprauland new real-time/on-the-fly formats can be developed with our camera, where else can you do that?03:48
wolfspraulhow about that?03:49
wpwrakokay, we;re getting closer. real-time ogg sounds good03:49
wolfspraulalso programmability/scriptability03:49
wpwrak"future-proof" is always tricky :)03:49
wpwrakyou don't needs scripts in the camera. send the stream to a pc. figure out there what to do.03:50
wolfspraulnot practical in many cases03:50
wolfspraullet's say you want to build a stereoscopic camera, or panoramic camera03:50
wolfspraulyou can beat many of these arguments with 'theoretically' we can do it such and such03:51
wolfspraulbut then I can also say "yes, but equally theoretically we can do it with Xue"03:51
qi-bot[commit] Bas Wijnen: mass storage sort of working, but not nice http://qi-hw.com/p/iris/5f9fd7c03:52
wolfspraulif we open up the CIS with an FPGA, and make it easily programmable at that level, that will open up a lot of interesting applications for many people03:52
wolfspraulit's all about how easy the access to the CIS gets03:52
wpwrakhmm. stream plus timestamps ? that way, you can easily compensate for timing differences03:52
wpwrakso, new algorithm development ? R&D or just R ?03:53
wolfspraulI do believe in the power of free software :-)03:53
wolfspraulso I'm not always asking myself "how can I do this with proprietary software"03:53
wolfspraulbut rather "what new things do we make possible with free software"03:53
wolfsprauland that's a lot then, I think03:53
wolfspraulthe CPU is already a free design, Milkymist. Now we hook a CIS up to it, with full fpga access.03:54
wolfspraulwho is the competitor?03:54
wolfspraulElphel? Xue will be much cheaper, and at least my strategy is to drive the price down, whereas Elphel is happy with a high-end 2000 USD and up camera.03:55
wpwrakso you're coming back to the killer feature being openness ?03:55
wolfspraulthat's the enabler, sure03:55
wolfspraulif that is not interesting, why start the project?03:56
wolfspraulyou can buy tons of excellent cameras03:56
wpwrakenabler for what ? ogg ? R&D ? ...03:56
wpwrakno no. the question is not that much was convinces us but what can convince others.03:56
wolfspraulagreed03:56
wolfspraulso after the first boards are made, we need to go find out "which applications become possible with this design"03:57
wolfspraulof course we can try now, but once we have the boards and people can start playing, I think we have a better chance of finding something really interesting03:58
wpwrakyup. that usually works best. make toys, watch 'em play :)03:58
wolfspraulI wouldn't underestimate the attractiveness of something simple like 3D (stereoscopic) or panoramic cameras03:59
wolfspraulalso there are almost no nicely programmable cameras in the market03:59
wolfspraulat least not at the level of 'ease of programming' I am imagining03:59
wolfspraulI give you an example :-)04:00
wolfspraulwhen Google started the book scanning project, all the big corps competed for this customer04:00
wolfspraulCanon, Kodak, everybody04:00
wolfsprauland... tiny 1-man Elphel from Salt Lake City, Utah04:00
wolfspraulhow did Elphel win?04:01
wolfspraulit came down to a few modifications in the firmware04:01
wolfspraulKodak and the others said they would need at least 6 months04:01
wolfspraulso even though this customer had huge political cloud, and the big guys gave it top priority, in the end Elphel won simple because it would not take 6 months to make a small modification to the firmware04:02
wpwrakalright, so flexibe image processing04:02
wolfspraulI think many many customers are in this position.04:03
wpwrakthis either needs examples, even if partially developed, or a "blue sky" target04:03
wolfspraulalthough it will be a challenge to craft a clear message, and get that message to them.04:03
wolfspraulyou mean the exact features Google needed back then?04:03
wpwrakyup. they'll probably see examples04:04
wolfspraulsee I think when you have a big project, like book scanning, small problems can quickly grow exponentially04:04
wpwrak"if it can do X then it might also do our Y"04:04
wolfsprauland then the inability to hack/script/modify/customize _ANYTHING_ in the firmware becomes equally nasty04:04
wolfspraulyou will probably admit that in general, the hackability of common cameras is really really bad04:05
wolfspraulthere are some projects actually04:05
wolfspraulit's like the smartphone world04:05
wolfspraulthey are juggling binaries04:05
wolfspraulreverse-engineering04:05
wolfspraultrying to find datasheets04:05
wolfsprauletc. etc.04:05
wolfspraulthen suddenly Canon changes CPUs and everything starts over :-)04:05
wpwrakwell, that would be camera setup and image output04:06
wpwrakbut that's just the basics04:06
wpwrakyou can get that by buying a documented sensor04:06
wolfspraulvery rare04:06
wolfspraulthe one we will choose also has no open datasheet04:06
wpwrakthe xue has to deliver more :) but yes, some nice examples can help04:06
wolfspraulAptina MT9M033...04:07
wpwraksigh04:07
wolfspraulwell of course we have the datasheet04:07
wolfspraulon the Aptina website you need to register and then ask for it, need to try what that means04:07
wolfspraullooks like Samsung04:07
wolfspraulI think the key thing that Xue will hopefully unlock is to hookup an FPGA to the CIS, and then open that whole interface up for free software04:08
wolfspraulespecially since the SoC design on the FPGA is already a free design04:08
wpwraknow you just need the fpga developers :)04:08
wolfspraulI see it like an additional peripheral04:09
wolfspraulwe can take Elphel sources, to a degree (don't know how compatible Xue and Elphel are on that regard)04:09
wolfspraulthe goal is compatibility with Elphel to be able to reuse at least some of their (GPL licensed) Verilog design04:09
wolfspraulthey even have an old Ogg Theora in Verilog already :-)04:09
wpwrakoh, compatibility is interesting04:10
wolfspraulfor sure04:10
wolfspraulthat was the original idea04:10
wolfspraulmilkymist compatibility, elphel compatibility04:11
wolfspraulboth may have been compromised a little by now, I don't know04:11
wolfspraulthe less the better, for sure04:11
wolfspraulit's about free software, reuse, etc.04:11
wolfspraulwpwrak: I totally agree with you that we need to define a concise marketing message04:13
wolfspraulkiller apps04:13
wolfspraulwhat is special about Xue?04:13
wolfsprauletc.04:13
wolfspraulit's at best grey right now04:13
wolfspraulwe are just building a toy04:13
wolfspraulif you look at Xue from the perspective of Milkymist One, it has the following differences:04:14
wolfspraula) adds a CMOS image sensor as a new 'peripheral' (the 'eye')04:14
wolfspraulb) in KiCad instead of Altium designer04:14
wolfspraulc) designed for energy efficiency (unlike Milkymist One)04:14
wolfsprauld) tried to retain as much as possible Milkymist compatibility, and adding Elphel compatibility (probably to a lesser degree)04:15
wolfspraule) board is substantially smaller (5x9cm), compared to 10x15cm for Milkymist One04:16
wolfspraulthat's it I thnk04:16
wolfspraulroh: you probably know that www.openmoko.org is down, right?04:26
wpwrakenergy efficiency and small size are probably not very "newsworthy". that it'unitl they're extreme04:27
wpwrakMM1 compatibility matters when the MM1 is popular :)04:28
wpwrakor when it's more convenient for prototyping. but that's probably not the case, given that applications would typically do something with the image04:29
wolfspraulfirst I'm describing the differences to you04:30
wolfspraulthose differences are the reason Xue exists, without them it would not exist.04:30
wolfspraula camera in most cases needs to be small, mobile, close to the object that is to be captured04:31
wolfspraulMilkymist One may grow into a 'desktop' or 'table' computer, but it will always be a media computer that is either somewhere under the table or in the corner (Milkymist One), or maybe at the hands of a media operator (Milkymist Two)04:32
wolfspraulone day Milkymist One/Two may also be done with Kicad, so that difference may go away04:32
wolfspraulI doubt Milkymist One/Two will ever add a CMOS image sensor, it would totally not make sense04:33
wpwrakagreed04:33
wolfspraulenergy efficiency is not important when you are plugged into the wall04:33
wolfspraulfor fashion/green reasons we may one day optimize it too, but it will never be a design criteria that tops other features like it might be in Xue04:33
wolfspraulsame for size, it just doesn't matter as much in a product like Milkymist One as it does in a camera04:34
wolfspraulso in prioritizations, size will never trump other features, like anything related with high media performance, in Milkymist One/Two04:34
wolfspraulbut in Xue it might04:34
wpwrakenergy efficiency would ave to be compared with other imaging solutions, not the MM1 :)04:35
wpwrakyup. they're very different there.04:35
wolfspraulthe Xue board can be the base for the Milkymist-based NanoNote04:35
wolfspraulor a phone, or any other mobile product04:35
wolfspraulit's much closer to those, although I prefer to do things one step at a time, and right now it's about camera, and camera only04:36
wpwraksure. a board that done some of the useful things.04:36
wolfspraulfor camera applications, a bit market are also industrial uses04:36
wolfspraula camera as part of a machine that is watching something04:36
wolfspraulin many many cases machines are becoming faster and faster04:37
wpwrakyup, right04:37
wolfspraulso when I gave you this example of the traffic cameras and how many cars they can identify at the same time now, of course that's not something I consider worthwhile competing in04:37
wolfspraulbut let's say you have a machine that cuts cheese into slices04:37
wpwrakso it needs fast image recognition apps. kinda a toolbox.04:37
wolfspraulor hundreds of other industrial machines04:37
wolfspraulthey need machine vision all over the place04:37
wolfsprauland our camera is programmable, and it has an fpga meaning that expensive image processing logic can be implemented in a high-performance way04:38
wpwrakyes. so, if the customer is to be a developer of such vision systems, there would be the toolbox04:38
wolfspraulyes04:38
wolfspraulsmall board04:38
wolfspraulsuper documented04:38
wolfspraulfree software04:38
wpwrakif the customer is the devloper of the toolbox, then you need less :)04:38
wpwrakif the customer is the factory, more :)04:38
wolfspraulwait cannot follow04:39
wolfspraulwhat?04:39
wolfspraultoolbox? developer? factory?04:39
wolfspraulI was just describing a use case.04:39
wolfspraulindustrial machine, doing whatever, at high speed04:39
wolfsprauleven in our industry, automatic optical inspection systems after SMT04:40
wpwrakyup. what i mean, the respective customers need different levels of existing infrastructure before they can do "their" thing04:40
wolfspraulwell, I would hope we would make the start as easy as posible04:41
wolfspraulso there's Ethernet04:41
wpwrakso, freaks first, the rest of the world later :)04:41
wolfspraulpower cable is separate04:42
wpwrakkinda like the ben04:42
wolfspraulso you need 2 cables with xue, then you can start04:42
wolfspraulElphel is using Power over Ethernet04:42
wolfspraulit's a looong story, but in the end we think a separate cable is better, at least for now04:42
wpwrakpatents ? :)04:42
wolfspraulyes, that's one big problem04:42
wolfspraulPoE is a patent mine04:42
wolfspraulpatent minefield rather04:43
wpwrak"cherchez le brevet" ;-)04:43
wolfspraulanother one is that the PoE standard is geared towards long distance cables etc. 40-70V etc.04:43
wpwrak70 V is pretty inconvenient04:43
wolfspraulso it doesn't go along very well with the power efficiency goals of Xue04:43
wolfspraulPoE is about powering all sorts of stuff through the Ethernet cables04:44
wolfspraulso you can focus on one cabling instead of two04:44
wolfspraulbut of course the whole network is then not really trying to hook together devices that squeeze out every last mW04:44
wolfspraulit's more heavy-duty plumbing :-)04:44
wolfspraulso you can power the big fat Cisco routers and what not04:45
wpwrakin such cases you also don't care. you have mains.04:45
wolfspraulit just doesn't fit very well with the power efficiency goals of Xue04:45
wolfspraulalthough just needing one cable is an interesting proposition04:45
wolfspraulbut then, PoE in general is very expensive and never got widely adopted, because of the patent situation04:45
wolfspraulit's a good idea though04:45
wpwrakwell, if your application doesn't need power efficiency doesn't mean that it wouldn't benefit from the xue04:46
wolfspraulfirst time I had my Elphel camera I couldn't believe that there was only 1 connector :-)04:46
wolfspraulwpwrak: did you know about PoE before this conversation?04:46
wolfspraulI seriously had _never_ heard about it :-)04:46
wpwrakthat it exists ? of course04:46
wolfspraulthen you tell me04:46
wpwrakit's not that obscure :)04:46
wolfspraulwhat do you think about it?04:46
wpwraknice idea. simplifies cabling. avoids potentially error-prone power supplies. like USB.04:47
wpwrakof course, you can spoil it all with patents and a clumsy design. don't know enough about it to tell if this is the case.04:48
wolfspraulit is04:48
wolfspraulvery expensive04:48
wolfspraulpatented04:48
wolfspraulplus the design philosophy04:48
wolfspraulPoE is all about how _much_ power can I get04:48
wolfspraulnot about how _efficient_ can the whole system be04:48
wpwrakwell, usb often is too ;-)04:48
wolfspraulthat is totally fine, but just not in line with what I believe are good priorities for Xue04:48
wpwraki think if you have a system that wants PoE, efficiency often isn't a concern04:49
wpwrakotherwise, use USB. or something specialized.04:49
wolfspraulthe latest PoE standard guaranteed 25.5W, and some vendors are proud to already offer 51W...04:49
wolfspraul(according to Wikipedia)04:50
wolfspraulso anyway, no PoE on Xue, at least not RC104:50
wolfspraulthat's another thing about Xue04:50
wolfspraulit is designed to be in the middle of 'vision performance'04:50
wolfsprauljust to get the basics (vision) working at all04:50
wolfspraulafter that we can specialize, if we have paying customers04:50
wolfsprauleither go to higher-grade vision applications, c-mount lenses04:51
wolfspraulor go to lower-grade vision apps, plastic lenses04:51
wolfspraulXue is the first foot in the 'vision' door04:51
wolfspraulhopefully we can then see more clearly :-)04:52
wolfsprauland specialize, turn it into a product04:52
wpwraksounds reasonable. high-end is too expensive, low-end doesn't save too much in the total system cosr04:52
wolfspraulcorrect04:52
wolfspraulI hope to be able to sell xue boards for 200-300 USD or so04:52
wolfsprauleven less04:52
wolfspraulI will try to squeeze out as much as possible04:52
wolfspraulafter we get that up and running, we can specialize and turn it into a product, chasing low-cost, or chasing high-end optics, etc.04:53
wolfspraulwell, I hope others join too at some point04:54
wolfspraulmaybe some company that takes this and turns it into a product04:54
wpwrakso the first run is for hackers, with open-ended application. more focus on infrastructure. similar to ben.04:58
wpwrakwhee, motorola sues apple over some patents. the "everyone sues everyone else" matrix in the phone market is filling up ;-)05:00
wpwrak"who's suing who in the mobile business": http://www.guardian.co.uk/technology/2010/oct/04/microsoft-motorola-android-patent-lawsuit05:05
wpwrakconsidering all the attacks originating from nokie, you'd almost think they have some issues with their products ;)05:06
wolfspraulif we don't sue someone really soon some people will start to say that prooves that we have no IP05:07
wolfspraulabout 'first run for hackers'05:08
wolfspraulthat's not how I see it05:08
wolfspraulalthough maybe I would not agree because it's obvious :-)05:08
wolfspraulany Xue run is about finding the shortest path to a product one can sell to someone who wants to use this for a non-programming application05:09
wolfspraulI do not believe in selling developer tools.05:09
wolfspraulnow, talking about the first run, of course those boards will not even be offered to such a person05:09
wpwrakso the customers buil their tools from scratch ?05:09
wolfspraulits' the first run of a new design with endless unknowns!05:09
wpwrakor do you plan to sell complete solutions ?05:10
wolfspraulno one by one05:10
wolfspraulI react to your 'for hackers' statement05:10
wpwrak"one by one" would be "hackers first" :)05:10
wolfspraulthe first run is about verifying the design05:10
wolfspraulhow many we do? I don't even know.05:10
wpwrakalright. the first run you actually sell :)05:10
wolfspraulI asked Andres how confident he is. He says "totally 100% absolutely confident"05:11
wolfspraulwhat does that mean?05:11
wpwrakokay, make the number low ;-)05:11
wolfspraulI haven't worked with Andres before.05:11
wolfspraulI am putting my money on the line.05:11
wolfspraulof course, I am 90% sure the boards won't even boot :-)05:11
wolfspraulso maybe 2?05:11
wolfspraul4?05:11
wolfspraulmaybe 205:11
wolfspraul:-)05:11
wolfspraulok 405:11
wolfspraul:-)05:11
wolfspraulI don't know.05:11
wolfspraulso of course these boards are 'for' hackers05:12
wolfspraulbut none of us can live on this kind of stuff05:12
wolfspraulshould I sell it to you?05:12
wpwraknot more than 10 :) depends a bit on how the cost structure works. probably not fewer than 4, just considering the overhead05:12
wolfspraulto Andres?05:12
wpwrakto andres. he'll need at least 2.05:12
wolfspraulwe need to sell to outsiders, people who use our technology for their applications05:12
wolfsprauland I try to find the shortest path to such people.05:12
wolfspraulnow, with the first run this question (whom to sell to) is not on the table yet.05:13
wolfspraultoo many unknowns05:13
wpwrakbut all that's pre-production. find the bugs.05:13
wolfspraulsure05:13
wolfspraulbut I will always try to find the shortest path to sell to non-hackers05:13
wolfspraulif you say "this is for hackers" I am just reacting allergically :-)05:13
wolfspraulbecause that's unsustainable05:14
wolfspraulso even if in reality RC1, RC2, RC3, RC4, are all just for hackers, I will always insist that I am on the shortest (!) path to non-hackers05:14
wolfsprauljust clarifying my thinking so we don't need to go over it again and again05:14
wolfspraulprobably we mean the same thing...05:14
wpwraknon-hackers need infrastructure. you have none. hackers can create their own. so i think it's pretty clear whom you can sell to ;-)05:14
wolfspraulok, but we don't undertake the whole endeavor to 'sell to hackers'05:14
wpwrakit may well be the shortest path.05:15
wolfspraulyes of course05:15
wpwraknaw, i hope not :)05:15
wolfspraullike I said we agree, just we use those words slightly differently05:15
wolfspraulso yes, RC1, and most likely also RC2, RC3 will, in reality, be sold to hackers05:15
wpwrakone question is when you define who else should buy it05:15
wolfspraul:-)05:15
wolfspraulI have to admit.05:15
wolfspraulBut i try to get away from it asap05:15
wolfspraulthe moment I am able to find such a customer05:15
wolfspraulselling to a non-hacker is always preferable over selling to a hacker05:16
wpwrakdo you make it a platform for "blue sky research" ? or do you try to steer things in a certain direction ?05:16
wolfspraulfirst let's get it to the point that it boots and can take images05:16
wpwrak(non-hcker vs. hacker) why ? if it's just one, the hacker should be more useful. it gets interesting when it's 100 non-hackers vs. 1 hacker05:16
wolfspraulthen, for any run larger than let's say 100, we need to have an actual customer/marketing message in mind05:17
wpwraksounds good05:17
wpwrakthough the first "public" run would still have to be "for hackers". unless you find the infrastructure somewhere. or you find companies that don't mind doing everything from scratch.05:18
wpwrakpioneer work on an unproven design. difficult. they must be very desperate to try this.05:18
wolfspraullet's make the first steps05:18
wolfspraulwe are not totally unknown, and totally incapable, of getting a message out that we have a product there05:19
wolfsprauland then we need to see which customers we can find05:19
wolfspraulbut right now I think you agree, it's just too far out to do this now05:19
wolfspraulI mean we have nothing but unknowns05:19
wolfspraulMilkymist, KiCad, CIS05:19
wolfspraulwell05:19
wolfsprauleverything actually :-)05:19
wpwrakparticularly, find those in the field of computer vision and image processing. the probably don't know you (yet)05:20
wolfspraullet's build something, then we see what actually works, then we find customers to whom that is valuable05:20
wpwrakthen find out what excites them. make one or two cool proof of concept applications.05:20
wolfspraulyes we are moving there05:20
wolfspraulno boards now05:20
wolfspraulso what's needed? at least 1 board05:20
wolfspraultrue?05:20
wpwrakall the while, you can already try to sell to "raw" hackers, of course05:21
wolfspraulwell, when I say I prefer to sell to non-hackers, of course that doesn't mean I don't sell to hackers05:21
wolfspraulI love a hacker customer as much as I love a non-hacker customer.05:21
wpwraka least 2 boards for design verification. but andres would know what number he needs. probably more. if you want to put adam or someone else on it, 2 more for that guy.05:21
wolfspraulyeah05:22
wolfspraulmaybe 405:22
wolfspraulor 5 if you want one :-)05:22
wolfspraulwe'll see05:22
wolfspraulit's a bit early now05:22
wolfspraulI need to focus on the Milkymist One RC2 run first.05:22
wpwrakdivide by production yield :)  that's 5 even without me ;)05:22
wolfspraulmy gut feeling is the first run will be a disaster05:22
wolfspraul:-)05:23
wpwrakany i'm not sure i want one before they've verified the design ;-) it can be completely broken05:23
wolfspraulbut let's see05:23
wpwrakexactly :)05:23
wpwrakwith my very simplistic and limited DIY things, i always make them in pairs. if one acts up, i try the other. if both have the same problem, then i have somewhere to start the search. otherwise, you're just chasing your own tail.05:24
wolfspraulI will try to enjoy the disaster as much as possible05:25
wolfspraulat least I think I deserve that for my money05:25
wpwrakand first versions tend to be bad. of course, i can fix a lot of problems in my DIY boards by just cutting up traces and adding wires and stuff :)05:25
wolfspraullearn something from it...05:25
wpwrak(having no inner layer helps ;-)05:26
wpwrakthen i may have another version with the V0 errate plus size and buildability improvements.05:27
wpwraketc.05:27
wpwraksometimes, the first try it useful. but often there's some stupid mistake. xue is complex, so the risk increases05:28
wpwrakof course, thigs do't have to take the gta01 route ;-)05:28
wpwraksomething like 3 board runs before the thing would even come out of reset. ten more before it would be considered sellable. well, something like these numbers. i think that's fairly extreme by any standard :)05:29
wolfspraulMilkymist One: first run boots, a few major errata but 80% works05:32
wolfspraulSIE: first run (by me) fully under control, 80% boards fully functional05:33
wolfspraulAVT2: first run, 20 boards, about 8 or 9 fully functional in the end (after reworks)05:33
wolfspraulthat's my track record05:33
wolfspraulall of these runs had issues, but this is about how it should be, imo05:34
wolfspraulI do want the first Xue run to produce boards that boot and can take images.05:34
wolfspraulreworks are OK05:34
wolfspraulwe'll see05:34
wolfspraulmaybe Xue will finally bring my track record down...05:34
wolfspraultuxbrain_away: are you there?05:35
tuxbrain_awaypartially05:35
wolfspraulI need your help with a few bits on the Nanowar announcement05:35
tuxbrain_awayshot05:35
wolfspraulhttp://en.qi-hardware.com/wiki/Press_release:_Nanowar_of_Steel,_copyleft_hardware_featuring_copyleft_music05:35
wolfspraulyou don't need to format it nicely, just dump at the bottom05:35
wolfspraulyour personal contact info, david@tuxbrain.com, plus phone number05:36
wolfspraulofficial Nanowar homepage05:36
wolfspraulofficial (?) download location for the new album05:36
wolfspraulnames of the 3 members05:36
wolfsprauldirect link into your shop with the special edition05:36
wolfsprauljust a few URLs by the way05:36
tuxbrain_awayok  give me 10 mins and I will do it asap05:37
wolfspraulsure no rush05:37
wolfspraulI meant URLs to help people find their way...05:37
wolfspraulI googled a bit but I am not sure what the official homepage of Nanowar is, or even the official download location for the album, if one exists05:37
wpwrak(track record) pretty good. no mass grave.05:37
wolfspraulyou do learn a thing or two as a software engineer about engineering methodologies05:38
wolfspraulwe are not complete idiots either...05:38
wolfspraul:-)05:38
wpwrak(nanowar) considering a "re-launch", PR-wise, after the first one fizzled out ?05:38
wolfspraultuxbrain_away: another thing which is more work and maybe not so important, but still wanted to feed it back05:39
wolfspraulwerner thinks the nanowar video should be cut into a 'best-of' which is maybe 1 minute long05:39
wpwrak(complete idiots) reminds me of that saying about standing on the shoulders of giants ... ;-)05:39
wpwrakwolfspraul: (video) YES !05:39
wolfspraulI haven't watched it yet, and my sound is broken, but Werner is normally a reasonable person so he may have a point05:39
wpwrakwatch it :)05:40
wolfspraultuxbrain_away: but if you could fill in those URLs and contact info/names for me that's great05:40
wolfspraulthe video cutting is more work, and maybe not worth it. i don't know.05:40
wpwraki wonder if there's a good cutting tool. i use kdenlive but i end up spending 95% of the time fighting bugs (and eventually losing) than doign the actual editing05:41
tuxbrain_awayyes, I was think on do a short version, more an announce like with some sincro with music, kino is quite good y you have enough HD to support the conversion to dv05:42
wpwrakthigs like video and audio out of sync after cutting. the getting more out of sync if i try to adjust it. then a crash when i try to fix it in another way.05:42
tuxbrain_awaybut I didn't find the time to do so yet05:42
wolfspraulthe album name is "Into Gay Pride Ride", correct?05:43
wolfspraulI had to google it, wasn't in your text, I think05:43
wpwraki've once tried kino too. also had troubles. but maybe they're less bad now. the odd thing is the kdenlive seems to have worked better some time ago.05:44
tuxbrain_awayyes that the album title05:44
wpwraktuxbrain_away: short version would be great. you have a lot of good base material there. but i think you could put the whole message into a 30 s clip. 3 s per video game, etc.05:47
wpwrakfrom http://www.guardian.co.uk/technology/2010/oct/04/microsoft-motorola-android-patent-lawsuit05:56
wpwrakmicrosoft: "[...] we've spent over 30 years developing cutting-edge computer software." reader comment: "When will it be released?" priceless :)05:56
wpwrakhttp://news.designlanguage.com/post/125203920906:00
wpwraka better chart of who is suing whom. sheer madness.06:00
wolfspraulthere is actually a big risk in these lawsuits, which is that when they are over, the general opinion/press will be "nobody can be in the phone space unless they are one of the big guys"07:12
wolfspraulwhich of course it not true, but it will be a massive wall of FUD to go against07:12
tuxbrainand openmoko failure is not good weapon to use against that wall :P, in spite of there were other reason of it's failure07:15
wolfspraultuxbrain: I am trying to finish those urls07:15
wolfspraulgot the band's homepage now07:15
wolfspraulproblem with the member names is that there are 5 members07:15
tuxbrainyes07:15
wolfspraulbut I need to know the names of the 3 on the picture07:15
tuxbrainok07:16
tuxbrainlet me search07:16
wolfspraulhttp://en.qi-hardware.com/wiki/Press_release:_Nanowar_of_Steel,_copyleft_hardware_featuring_copyleft_music07:17
wolfspraulthe picture in the middle, for the subtitle under it (maybe names from left to right would be great)07:17
wolfspraulI'll try to dig up some more urls07:18
wolfspraulis there an official download location for the album?07:20
tuxbrainupdated the names07:21
wolfspraulgreat07:22
wolfspraulcan I post the phone number I have from you? (the one ending in 87)07:22
wolfspraulor what is a good phone number to reach you07:22
wolfspraulit seems the official spelling of the band is NanowaR ?07:24
tuxbrainhttp://www.jamendo.com/en/album/75948 link to dowload the album07:25
tuxbrainno is Nanowar of Steel07:25
wolfspraulis the 'R' capitalized?07:25
wolfspraulthey write it like that themselves07:25
tuxbrainis a parody of the Manowar original group covers07:25
wolfspraulok, so just Nanowar for us then07:26
wolfspraulhow about your phone number? is it OK to publish?07:26
tuxbrainyes go ahead07:26
wolfspraulI am trying to get the press release to a bit more normal level07:26
wolfspraulalso then we have a template for the future07:26
tuxbrainfull name of the band Nanowar Of Steel07:27
wolfspraulgot it07:27
wolfspraulif a journalist takes a press release for their publication, in almost all cases they will have some small (or large) question before they do that07:27
wolfspraulso a quick way for them to contact you is deal07:27
wolfspraulideal07:27
wolfspraulphone, instant messaging, email07:28
wolfspraulpersonal email07:28
wolfspraulvery important07:28
wolfspraulcheck the press releases even of the largest companies, there is always an individual at the bottom, with clear contact info07:28
methril_worknice!! press releases at qi-hw!!07:30
methril_workmorning07:30
methril_workwolfspraul, nice conversation wit wpwrak about Xue cameras, i'm reading right now because my nick was there :)07:32
wolfspraulyes that's why I nicked you :-)07:36
wolfspraulyou sent that mail months ago07:36
wolfspraulasking about Elphel vs. Xue07:36
methril_workyes07:36
wolfspraulI think about it every day07:36
wolfspraulmy todo list management is really strange07:36
wolfspraul:-)07:36
methril_workbut the conversation gets more interesting :)07:36
methril_workhehehe07:36
wolfspraulI would have so many things to say07:37
wolfspraulbut then never get around writing that one, short, meaningful mail07:37
methril_workyour conversation motivates the devs to use xue07:37
wolfspraultuxbrain: ok check it out http://en.qi-hardware.com/wiki/Press_release:_Nanowar_of_Steel,_copyleft_hardware_featuring_copyleft_music07:38
wolfspraulI think I'm almost done07:38
wolfspraulI haven't found a way to embed the video nicely yet07:38
wolfspraulalso, we should say somewhere that all pictures are free to be reused for publishing in any medium or format07:38
tuxbrainmmm about a nanonote for review.... what will be the policy?07:39
tuxbrainI supose discuss case by case, isn't it?07:40
tuxbrainyes the comment about the pictures must be there07:41
wolfspraulof course case by case07:42
wolfspraul99% don't get anything07:42
wolfspraulthose guys with a useless website who write you a long email how they need a free device, bla bla bla07:42
wolfspraulbut that's not the point, this is a standard offer that just has to be there07:43
wolfspraulof course it is up to you to decide which journalists to work with07:43
wolfspraulmethril_work: are you still interested in the diff between Elphel and Xue?07:45
wolfspraulmaybe we should try here to clarify it...07:45
methril_workit's pretty clever now07:45
wolfspraulI love Elphel, went to visit them for 3 days in May07:45
wolfspraulI want to cowork with them but so far the opportunity has not come up.07:46
methril_worki see a lot of potential in Open Cameras07:46
wolfspraulAndrey is definitely thinking about high-end stuff.07:46
methril_workwell, qi-hw philosophy in cameras07:46
wolfspraulcurrently the 353 will already cost above 1000 USD in almost all cases07:46
methril_workuf!! so expensive07:46
wolfspraulI fear, but I guess Andrey has no problem with that, that the successor, the 373, will be even more expensive07:47
wolfspraulElphel and I can only cowork if he sees a point in driving prices down07:47
wolfspraulthat is the only thing I can add to his product07:47
methril_workwell, maybe the focus is different07:47
methril_workhigh-end could be used for R&D07:47
wolfspraulbut if he doesn't see the point of that, then: a) I cannot cowork with him on 373, because I cannot add anything and b) 373 will be around 2000 USD07:47
methril_workand later on Xue cameras (cheaper) could be used for production07:48
wolfspraulhardware and price is a very difficult and long topic07:48
wolfspraulbasically 'low cost' is something you have to go for, and make very aggressive decisions for07:48
wolfspraulmany people don't understand the true dynamic of hardware costs07:48
wolfspraulso... I will continue to work wtih Elphel07:48
methril_workmost of hackers think that is cheap as sw07:48
wolfspraulin the sense that we already collaborate on the 373, and I am open to manufacture it07:48
methril_workbut it's hw07:49
wolfspraulbut what can I add... not much, unless Andrey gets the whole 'drive price down' approach, and sees value for his business07:49
wolfspraulactually it's also risky07:49
methril_worki like you help open projects07:49
wolfspraulimagine you have a product with a margin of 500 USD, and you sell 40 a month07:49
wolfspraulso you have a gross profit of 20,000 USD / month07:49
methril_workwell, but if you lower prices, you could sell 200 / moth07:49
methril_workonly getting 100 USD, you have the same result07:50
wolfspraulnow you have this option: you can invest 200,000 USD, and then you can manufacture and sell your product for 1/3rd of the price07:50
wolfspraulyes, but:07:50
wolfspraulare you sure there will be more customers?07:50
methril_workuhm!! interesting question07:50
wolfsprauldon't do the easy armchair general thing "yeah, sure"07:50
wolfspraulit is not sure!07:50
wolfspraulmaybe you just successfully shot yourself in the foot07:51
methril_workin relly specialized products, is hard to said how you could scale your market07:51
wolfspraulyou are selling your product for less now, but you have no guarantee on the other side that your volume goes up accordingly07:51
Action: methril_work nods07:51
wolfspraulyes07:51
wolfspraulit's difficult07:51
wolfspraulvery difficult to make those decisions07:51
wolfspraulbut I am saying to Elphel, I can only add one thing to their great product07:52
wolfspraulCOST REDUCTION07:52
wolfsprauland whether they want to add that or not, is up to them07:52
wolfspraulthey don't need me on the hardware side, software side, optics, mechanical, anywhere07:52
methril_workwe hope to see a hackable camera with lower price07:52
wolfspraulso, Elphel is there, and Elphel is great07:52
wolfsprauland you can buy the 353 today07:52
wolfsprauland you can help Andrey improve the software for the 35307:53
wolfspraulI am talking with them about how I can participate in the 373, but maybe they want to do it all themselves, because they feel this cost reduction thing is too risky or not necessary for their customers07:53
wolfspraulthat means the 373 will cost even more when it comes out07:53
methril_worki would love, but i'm lack of time ;) i'm fighting to get 2/3 hours for pending tasks07:53
wolfspraulis it clear now where Elphel stands, from my perspective?07:54
wolfspraulalso, Andrey is a quality fanatic07:54
methril_workit's07:54
wolfspraulhe uses a proprietary tool for the electrical design07:54
methril_workthis is not always bad, quality is iumportant07:54
wolfspraulof course I could say "let's switch to KiCad for 373"07:54
wolfspraulbut most likely he would like to continue with the tool he feels most comfortable with07:55
wolfspraulalso keep in mind that he is a self-taught electrical, mechanical, and software engineer :-)07:55
wolfspraulhe was originally a physicist, working in Moscow's excellent state sponsored physics labs in the 80's :-)07:55
methril_workhe is nice, and their blog posts are really good07:56
wolfspraulyes totally07:56
wolfspraulI would love to 'help' them on the 37307:56
wolfspraulif there is anything I can do that they think helps them07:56
methril_worki love to learn optics and why the lens works as they works in his blog07:56
wolfspraulwe are in touch about that07:56
wolfspraulyes, he is coming at the camera problem originally as a physics guy07:56
wolfspraulthen he built his perfect camera around that...07:57
methril_worki'm also happy that this is not diverging so much (the Ephel - Xue), because joining efforts could help both projects07:57
wolfspraulI have no clue about physics or optics.07:57
wolfspraulI'm a software guy.07:57
wolfspraulwell OK I know that the Apple falls towards the earth.07:57
wolfspraulsomething like that07:57
wolfspraulso I come at the camera problem differently, and I think Andres shares that vision07:58
wolfspraulI want hardware to be as cheap as possible07:58
wolfspraul10 USD07:58
wolfspraulfor the whole thing07:58
wolfspraulthat's my goal07:58
methril_workhehehehe i'm a sw guy too (or something in the middle)07:58
wolfspraulthen have great free software on it to bring it alive07:58
wolfspraulfully scriptable, programmable, the best documentation in the world07:58
wolfsprauldo you see 'optical quality' anywhere in my wet dreams? no...07:58
wolfspraulso it's different07:58
wolfspraulof course I want 'as good as possible' optics07:59
wolfspraulbut it's more like the iphone thinking07:59
wolfspraulI mean the cameras in the smartphones are getting better and better07:59
wolfsprauland those modules cost a few USD, including lense07:59
wolfspraulI want to participate in that kind of innovation.07:59
wolfspraulnothing against phenomenally good optics, in some industries it's necessary07:59
wolfspraulnot only military, also more worthwhile goals07:59
wolfspraulbut the Xue angle is different08:00
wolfspraulhigh optical quality yes, but driven from a reasonable cost basis, and trying to leverage free software as much as possible08:00
wolfspraulthe power must come from free software first08:00
methril_workwell, i would love to see both products being sell08:00
wolfspraulthen from really neat designed, low cost hardware08:00
wolfspraulI want to avoid to compete with Elphel over their customers08:01
methril_workth good thing is that you approach could scale to other fields, like a phone with a "xue" camera inside08:01
wolfspraulbut I think most of them need the Elphel cameras for specific performance features of their cameras08:01
methril_workbut this is far away08:01
wolfspraulyes totally08:01
wolfspraulno absolutely08:01
wolfspraulthat's how I think08:01
wolfspraulfirst: 10 USD08:01
wolfspraulthat's the goal08:01
wolfspraulthe fancy lense for Martin Scorsese where the lense alone is already 20,000 USD is a way way way after-thought for me.08:02
methril_worki remember a work that i made where the HW decision was taken cause the price08:02
wolfspraulthe question for Xue is - how much camera/optical performance can we squeeze out of 200 USD, or less later, and only using free software and free licensed hardware08:02
methril_workthe sw i made, has to be really challenging, due to hw limitations08:03
wolfspraulok did the message come across somehow?08:03
methril_workyes08:03
wolfspraulabout the differences between Elphel & Xue?/////////08:03
methril_work:)08:03
wolfspraulto me they are massively different08:03
wolfspraulXue will start at 200-300 USD, and then down08:03
methril_workyes, now i see the difference08:03
wolfsprauland the performance has to come from free software, and squeezing every last bit of hardware performance out of standard components, and a free design, KiCad, etc.08:04
methril_workthank you so much for all the clarifications08:04
wolfspraulno thanks for asking08:04
wolfspraulnot only do we need your help in making this a reality, but we also need a strong and clear message08:04
wolfspraulfirst within the believers, and then later to spread08:05
methril_workthis is what i love about qi-hw and you08:05
methril_workthe messages are clear and loud, and we could see the main focus08:05
wolfspraulhmm08:06
wolfspraulwe need to make it more clear still :-)08:06
methril_work:)08:06
wolfspraulanyway08:06
tuxbrainand more loud :)08:06
wolfspraullet's see about Xue08:06
methril_worklet's see08:06
methril_worktuxbrain, you are getting the press attention, so.. you are speaking loud :P08:07
methril_worki started to read a brazilian electronic community, i'll try to post and talk about qi-hw08:08
tuxbrain<yoda mode> do or not do it but not try08:10
methril_worklol08:10
methril_workwell, time to back to work08:10
methril_worki read you later08:10
methril_worktuxbrain, <young padawan mode> yes master, i'll do ;)08:11
wolfspraultuxbrain: ok done, I leave it like this08:13
wolfspraulhttp://en.qi-hardware.com/wiki/Press_release:_Nanowar_of_Steel,_copyleft_hardware_featuring_copyleft_music08:14
wolfspraulI don't expect a big anything now anyway, but I wanted to clean this up so it looks good in our news feed history, and maybe some people will still find it later, and we have a template...08:14
wolfspraulmethril_work: subscribe to http://en.qi-hardware.com/feed/rss20.xml08:15
wolfspraulor the entire planet right away :-) http://en.qi-hardware.com/planet/rss20.xml08:15
wolfsprauland spread those links among any brazilian journalists or blogs08:15
tuxbrainon for me well done wolfi :)08:17
wolfspraultuxbrain: btw, what you would normally do now on the release day is that you email, one by one, all journalists known to you08:18
wolfspraulmaybe repeat the text in the mail08:19
wolfspraulmany websites have 'submit news' forms (not email)08:19
wolfspraulyou need to use them08:19
wolfspraulI suggest you work yourself through the world from east to west, because of timezone08:19
wolfspraulso you start in japan, australia, then india, then europe, then us08:20
wolfspraulyou can spend one entire day emailing journalists whose contact info (email) you have, or to general 'submit_news@' addresses of media, or to web forms08:20
wolfspraulit's quite boring stuff, but it needs to be done08:20
wolfspraulone full day is enough, after that it's not news anymore anyway08:21
wolfspraulyou do not have to be worried about spamming anybody08:21
wolfspraulthey _want_ people to write to them08:21
wolfsprauland to submit news08:21
wolfspraulwhether you are able to make your stuff stand out is another question08:22
wolfspraulyou need to think very carefully about the subject line of your email08:22
wolfspraulor the first/second line in the email body08:22
wolfspraulor the subject line in web forms you fill out08:22
wolfspraulin many cases they are scanning through hundreds of mails/posts, and pick out some they think are interesting08:22
wolfspraulthe subject line needs a lot of attention!08:23
wolfspraul(for example the one on this press release is not so catchy, imho, but I also don't have a better one right now)08:23
wolfspraulideally your news relate to something current that is going on08:24
wolfspraulfor example maybe you remember the FreeRunner launch when Steve did the "Openmoko declares independence" news on July 4th08:24
wolfspraul:-)08:24
wolfspraulthat works!08:24
wolfspraulof course it's also risky, let's say Bin Laden decides to attack on the same July 4th08:25
wolfspraulbut you have to take those risks08:25
wolfspraulyou need to connect your news to something that is CURRENTLY HOT08:25
wolfspraulnot something that can also be published a week or month later08:25
wolfspraulok, so far my general guidelines :-)08:26
wolfspraulhmm, maybe another thing08:36
wolfspraulwhen you work yourself through the world (east to west), allocate fixed time for the timezones so you don't spend too much or too little somewhere08:37
wolfspraulmaybe 2h japan/australia/india, 4h europe, 6h north and south america08:37
wolfspraul12h :-)08:37
wolfspraulyou can handle that...08:37
wolfspraulthen you need to be able to handle 2 types of work08:37
wolfspraulfirst, you may get emails or phonecalls back from some people who have questions before they can carry the news08:38
wolfspraulthose need to be answered very fast, say within 5 or 10 minutes08:38
wolfspraulthen, if someone carries your news, there will be reader comments, and ideally you have some spare energy to manage those comments as well08:38
wolfspraulby saying thanks, by adding more facts, by pointing to your shop, etc.08:39
wolfspraulso if you can, you could allocate an extra help for that release day, say someone who handles incoming mails and phone calls, as well as reader comments, while you are working down the list of countries and publications and journalists to push the news to08:39
wolfspraulworst case you do all this, and nobody picks it up. that can still happen of course, but this is the normal release procedure and release work, so I think it should always be done nonetheless.08:40
wpwraknice press release !10:41
xdpirateso I heard there was some drama here, why haven't you done anything about it zear, you lazy fuck?10:50
xdpiratewrong channel =(10:50
larscthe drama is still going on ;)10:50
wpwrakdo i smell dirty laundry ? ;-)10:52
wolfspraulwpwrak: well, the next thing is that the text needs to be written in such a way that you can cut it off easily10:55
wolfspraulbut that's too much work now, and we have a somewhat solid base, if only for the archives10:56
wolfspraulthen we would need a really strong title/subject line10:56
wolfspraulnext time... :-)10:56
wpwraki don't find the title so bad. "copyleft hardware" in itself is an unusual item.10:58
wpwrakthe band name takes care of most of the catchiness anyway :)10:59
wolfspraulbut it's not in the title10:59
wolfspraulneither is the album name10:59
wolfspraullike I said, the title needs some real thought10:59
wolfspraul(in general)10:59
wpwrakhuh ? the title i see is "Nanowar of Steel, copyleft hardware featuring copyleft music"10:59
wpwrakhttp://en.qi-hardware.com/wiki/Press_release:_Nanowar_of_Steel,_copyleft_hardware_featuring_copyleft_music11:00
wolfspraulah OK, sorry11:00
wolfspraulgot that wrong11:00
wolfspraultrue true11:00
wpwraknot sure if you need to cram the album title in there too. may have to be one or the other. else you have an inro paragraph, not a title :)11:01
wpwrakinTro11:01
wolfsprauloh sure11:01
wolfspraulthat's the tough part of a good title/headline :-)11:01
wpwrakand if you just concatenate names/titles, then it becomes unparseable11:01
wolfspraulin the end in most cases any journalist picking it up will make his own title anyway11:02
wolfspraulhe has to11:02
wolfspraulso the purpose of the title is only to stand out in the flood the journalist is looking at, and making him want to read the rest of the text11:02
wolfspraulso the audience of _our_ title is actually the journalist, not the reader11:02
wolfspraulalthough of course the journalist can pickup some idea he got from our title into his title11:03
wpwrakthere are no links to qi-hw and jlime11:03
wolfspraulbut more likely he will make his own anyway11:03
wpwrakyup, the title should work for this. tells them what area it is about and suggests something unusual11:03
wolfspraulmy favorite title in the NanoNote launch was '3 adorable inches of Linux'11:05
wpwrakheh heh ;-)11:05
wolfspraulnot my title of course, made by some Italian publication11:06
wolfspraulbut it was my favorite, definitely11:06
wolfspraulbeautiful11:06
wpwrakyup. very cool.you can't not want to read what the heck they mean :)11:07
wpwrakhmm, compared to kdenlive, kino seems to be a bit limited in terms of editing and mixing. also crash-happy :-(11:25
wpwrak@#%*! when i run eeschema on fidelio with a set of files that usually create "ghosts", i get them, too. when i run the same things under strace, poof, gone.12:25
wpwrakQBJECT: he won't want it here. in fact, he won't even like the link you just posted in the IRC log12:26
QBJECTaw crap.12:27
QBJECT[sigh] noted for the future.12:28
wpwrakhe'll be up and deleting in a few hours :)12:29
QBJECTAt least it can be remedied.12:29
Action: QBJECT is looking up previous logs to get a better sense of where The Line is.12:32
wpwrakno links to packages or instructions for circumventing mp3 on qi-hw12:35
QBJECTArgh. Just found that.12:36
QBJECT"wolfspraul:nothing on the qi-hardware.com or affiliated servers must point towards the patented stuff"12:36
wpwrakno links to places that are strongly associated with such things. that would exclude jlime.com12:36
wpwrakyup12:36
Action: QBJECT contemplates the dangers of unbridled enthusiasm.12:37
wpwrakah, countless possibilities for that :)12:39
wpwrak(ghostbusting) it gets better. on the next run, they're different again.12:40
wpwraklet's see what turning off heap randomization does ...12:41
qi-bot[commit] carlos: Fixing some errors on SAKC Examples http://qi-hw.com/p/nn-usb-fpga/568f21812:54
tuxbrainwpwrak: [kino] I supouse is matter of luck, due for me kino is the most stable by far, and cutting, is quite straight forward, on effects you can do lot of kind of transitions and such, but well, that's magic  of freedom stay on what you feel more confortable/ productive.16:32
methril_hi17:03
methril_worktuxbrain, i sent the message to some brazilian groups, and a so important hw blog call "Laboratorio na Garagem", had a really good aceptation17:05
methril_worki hope to get some qi devices here soon ;)17:05
tuxbraincool that means you will retake the be brazilian distributor idea?17:06
kristianpauldamn i cofused TSSOP with SOIC >(17:26
kristianpaulnoo17:31
kristianpaulorder again from TI...17:31
kristianpauland wait they aprove it first..17:31
kristianpaul:(17:31
kristianpaulwpwrak: had you soldered TSSOP chips with wires? :)17:32
wpwrakkristianpaul: no, but why wires ?18:08
wpwraktuxbrain: (kino) it seems that you can't have multiple tracks that you mix. e.g., audio tracks. also, it seems that you can't edit effects after applying them.18:09
wpwrak(but maybe i'm wrong about the latter. that one')18:09
wpwrak.. one's from memory18:10
kristianpaulwpwrak: (wires) just kidn hurry here to hook thi ship to the SIE and wee what can i read/draw18:17
kristianpaulwires jsutkidding, i can get abreakout board but for SOIC here in colombia18:18
kristianpaulTSSOP by ebay and will take time18:18
kristianpaulkicad, fped, yes i'm larning but i need mroe time,so considering more options18:18
methriltuxbrain, re18:19
wpwrakkristianpaul: (breakout board for soic) why wouldn't just just etch it yourself ? tssop and friends are easy18:19
methriltuxbrain, about the distributor... maybe18:20
tuxbrainre18:20
kristianpaulwpwrak: i cant find TSSOP in kicad, would you help me?18:20
wpwrakkristianpaul: (kicad, fped) learn them now and you'll have your board today or tomorrow ;-)18:20
methriltuxbrain, i'm tgrying to find a workaround for the bureocracy18:20
kristianpaulohh18:20
kristianpaulok i'll dig on this on weekend18:20
wpwrakkristianpaul: what sort of tssop is it ?18:21
kristianpaulwpwrak: Texas tlv156218:22
wpwrakyou could add it to f32xbase/hw/modules/ssop.fpd18:27
wpwrakssop.fpd implements packages according to NXP's http://www.nxp.com/#/page/content=[f=/packages/SOT341-1.xml]18:28
kristianpaulok soon get home :)18:28
kristianpaulas soon**18:28
tuxbrainmethril: I hope you find a way :)18:28
wpwrak(it's untested, that's why it lives in a project, not in kicad-libs)18:28
kristianpaulk18:29
methriltuxbrain, i hope it too :)18:30
ezdagorI think I bricked my Nanonote.18:36
ezdagorI was flashing a filesystem to it, and forgot to flash the kernel. Now the unit will not turn on or boot, or do anything. Not even USB booting will work.18:37
wpwrakezdagor: usb booting = with the rubber button or with U+power ?18:39
ezdagorWith the rubber button.18:39
wpwrakezdagor: have you done the process before ? (successfully)18:39
ezdagorYes.18:40
wpwrakezdagor: hmm. just a reminder of the two most common pitfalls:18:40
wpwrak1) make sure the device really resets: remove battery and usb, wait about half a minute, then plug USB. (don't put the battery)18:41
ezdagorOk. Will try that.18:42
wpwrak2) make sure you press the rubber button before inserting USB and to keep it pressed for a few seconds after connecting USB.18:42
wpwrakalso, the process is a bit fumbly, so it may take a few tries to get it right.18:43
qbjectDoes anyone have a Dingoo w/ Dingux? I'd like to try something and I need some lib files.18:46
wolfspraulwpwrak aisa he, I saw the little mp3 circumvention thing here21:05
wolfsprauljust to keep in mind, there is still a concept of free speech, which imo by far trumps patent law21:06
wolfspraulso: everybody in this channel can say and give advice for whatever they want, restricted only by the limitations to free speech as we typicall know and accept them21:07
wolfspraulonce in a while, me or other businesses who may be the target of a patent infringement claim, and who use or point to resources on the qi-hardware.com servers, will cleanse the irclog archives (grep mp3 ...), to defend ourselves from the patent infringement claim21:08
wolfspraulbut that's about it21:08
wolfspraulI hope nobody feels that their free speech rights on those issues are somehow restricted here. they are not!21:08
wpwrakwolfspraul: alright, so anything goes here and we trust that you'll catch them later when checking the logs21:09
wolfspraulme and others who are likely targets of infringement claims will take care of this, sure21:10
wolfspraulit also has to do with the 'quality' of our copyleft hardware project21:10
wolfspraulthat we can tell new companies or individuals that want to produce or sell some of our stuff that these things are professionally dealt with, as much as that is possible21:10
wolfspraulplus it's pretty easy21:11
wolfspraulall logs are in one folder21:11
wolfspraulgrep -i mp3 *.log21:11
wpwraksure, due diligence21:11
wpwrak;-)21:11
wolfspraulwell, werner would write one sed line in his sleep21:11
wolfspraulwe just need to keep the different concepts separate21:11
wpwrakhow to circumv3nt emm-pee-three  ;-)21:11
wolfspraulfree speech is one thing21:11
wolfspraulpatents another21:11
wolfspraulpatent infringement another21:11
wolfspraulenforcement of patent infringement yet another21:12
wolfspraulwpwrak: I'm sure your uber-sed would catch that one too... :-)21:12
wpwrakalas, courts seem to consider IP claims to trump pretty much anything else21:12
wolfspraulno I don't believe that21:12
wolfspraulthe press is also a bit sensationalist at times21:13
wolfspraulso let's say some patent troll sues Megacorp ABC for 25 fantastillions21:13
wolfspraulit surely makes it into all the blogs and tech press21:13
wolfspraulwhen that suit it thrown out 3 months later, nobody reports about it21:14
wpwrakthere are some extremely broad interpretations of what constitutes appropriation of content. perhaps they'll get reversed if you take it to the bitter end of the appeals process, but that's also very expensive21:14
wolfspraulwe've already outlined the strategy - focus on real threats, not perceived threats21:14
wolfspraulin the messy patent law that's the only chance if you still want to get work done and still want to serve real customers21:15
wpwraknaw, i'm talking about court rulings. particularly the lower ones can do crazy things.21:15
wolfspraulso that's sisvel, mpeg la, others as we identify them one by one21:15
wpwrakyup21:15
wolfsprauland those guys will not log in here, track the channel, and make some compliated conspiracy case out of it21:15
wpwrakalright, i like the "no gag" rule :)21:15
wolfspraulit just won't happen, I am not interested in spending my time thinking about such things.21:15
wolfspraulan end-consumer device that can play mp3, yes, they try to find the source they can plug into for taxation21:16
wpwrakhow about posting a policy ?21:16
wolfspraulwell21:16
wolfspraulyou wrote a great one21:16
wpwrak(log on) naw, but they may google21:16
wolfspraulI have too many mails in my inbox right now, but will get to it (within the next few hours)21:16
wolfspraulyes sure21:16
wpwrakokay, no rush :)21:16
wolfspraulit it's on the server, and especially if there is a link, and if the link is not taken down (whenever exactly we leave open), then one can legally argue that whoever operates the server infringes21:17
wolfspraulthere is no doubt about that, this is a fact21:17
wpwraknot sure if a policy thingy like this would be more good or more bad, but i think people have a right to know what the rules ares21:17
wolfspraulit's good21:17
wolfspraulvery god21:17
wolfspraulgood21:17
wpwrakkewl. let's see how to proceed with it when you've read it.21:18
wolfspraulI think if I walk into a media-markt (similar like best buy) in Germany today, I can find pirate mp3 devices from China21:19
wolfspraul'pirate' in the sense that nobody paid sisvel21:19
wolfspraulthey are fighting the floods (from their perspective), worldwide21:19
wolfspraulthey have no time for complicated legal cases21:19
wolfsprauland that is not their business21:19
wpwraksecond item: the ghostbusting. my band-aid so far would be to just disable heap randomization. how do you feel about that ?21:19
wolfspraulin the whole world, every day, hundreds of thousands of mp3 devices are manufactured and sold21:20
wolfspraulhow many pay the sisvel duty?21:20
wolfspraulwhat do you think?21:20
wolfspraul20%?21:20
wolfspraul50%?21:20
wolfspraulI don't think more than that for sure21:20
wolfspraulmaybe only 10%?21:20
wolfspraulTHAT is their opportunity21:20
wolfspraulthe other xx%21:20
wpwrak(mp3) yeah, one approach to defeat them would be to create a gazillion of little importers with very shallow pockets. then let them waste money on useless lawsuits ;-)21:20
wolfsprauland it's about time, stuff is made, sold21:20
wolfspraulyou have to get the money now, otherwise it's gone21:20
wolfspraul(not theoretically, but practically)21:21
wolfsprauloh sure, they cannot21:21
wolfspraulI am telling you, for someone like dealextreme.com patent problems probably do not exist21:21
wolfspraulI think sisvel would already be happy if no large chain would carry unlicensed mp3 players in 'specials'21:21
wpwrakthe small chinese have the advantages that they're a) far away and b) have no investment in a brand name21:21
wolfspraul10 EUR, huge pile of them in the store, bought quick, sold quick, forgotten quick21:22
wolfsprauland who remembers sisvel? nobody! :-)21:22
wolfspraulsure21:22
wolfspraulthey change company names faster than I can update my address book21:22
wpwrakso they can also rename quickly. as long as the name changes before sisvel has their confiscation order for customs, they miss that one21:22
wolfspraultotally21:22
wolfspraulnow you are slowly in sisvel's world21:22
wolfspraulnot about logging into this channel and coming after a few crazy free whatever guys21:23
wpwrakbut that model doesn't work for us. we want to build a name, a reputation21:23
wolfspraultrue true21:23
wolfspraulbut like I said, I am just trying to explain that free speech is still free speech21:23
wolfspraulinfringement is another, separate, thing21:23
wpwrakno ... they wouldn't come here to find a target but to find evidence once they've already decided on the attack21:23
wolfsprauland the enforcement of that infringement yet another21:23
wpwrakgoogle site:qi-hardware.com mp321:24
wpwrakthis sort of thing21:24
wolfspraulyes21:24
wolfspraulagreed21:24
wolfspraulthat's why I was talking about the links, and the need to 'police' them to avoid a successful infringement claim later21:24
wolfspraulbut that doesn't restrict anybodys rights to free speech21:24
wolfspraulwe can even install a 'mp3 cleansing' script to run daily21:24
wolfspraultakes 10 minutes to setup21:24
wpwraknow you're making "mp3" a forbidden word ;-)21:25
wolfspraulif there is mp3 and http:// in the same line, delete the log line21:25
wolfspraulno21:25
wolfspraulI automatically delete it from the logs21:25
wolfspraulthat's different21:25
kristianpaullets use mpe21:25
wolfsprauland why do I do that?21:25
wolfspraulso that qi-hardware.com is a 'quality source' for people who want to manufacture copyleft hardware21:25
kristianpaulwolfspraul: i'm at home :)21:25
wpwrakwiki.where.ver/why-you-cant-play-mp3 ;-)21:25
wolfspraulwe have the same problem, though with less aggressive people, over microSD21:25
wolfspraulactually any 'SD'21:25
wolfspraulSecureDigital is a trademark21:26
wolfspraulwe own no rights to it21:26
wpwrak(uSD) yeah, your favourite :)21:26
wpwrakis SD a trademark, too ?21:26
kristianpaulyes21:26
kristianpauli think21:26
wolfspraulideally on the entire tuxbrain.com/qi-hardware.com/any other copyleft hardware guy universe there would be no *SD, anywhere21:26
wpwrakthat would suck21:26
wolfspraulI see it as a sign of quality21:27
wolfspraullet's say you are a small shop in the US21:27
wolfspraulyou want to see the designs on qi-hardware.com as an opportunity to manufacture something21:27
wpwrakwell, quality and relatively pointless annoyance at the same time21:27
wolfspraulyou ask "have you guys infringed on anybody's rights with this stuff?"21:27
wolfspraulwhat's the answer21:27
wolfspraulthe answer is "we are in China, we don't care"21:27
wolfspraulnot good21:27
wolfspraulso I feel these issues about IP 'cleanliness' are an important measure of quality for our project21:28
wpwrak"we are in the US, there's no way of knowing" ;-)21:28
wolfspraulso others are not worried to use our stuff21:28
wolfspraulwpwrak: makes sense?21:28
wolfspraulI don't think we have to work on this every day, but in general that's how I see it.21:28
wpwrakof course21:28
wolfspraulit's a _service_ that we remove all these patented or otherwise appropriated things from our universe21:29
wpwraka good angle to include in the policy. the current motivation is in a more reactive sense21:30
wpwrakyesh, give it a positive spin21:30
wolfspraultotally21:30
wolfspraulI also don't buy into this patent FUD many people are spreading, but that's another story.21:30
wolfspraulwe are working on high quality free content.21:30
wpwrakwe don't smash shops. we sell protection. we help business thrive ;-)21:31
wolfspraulthat 'smartphone patent lawsuits' map is hilarious, isn't it?21:31
wpwrakon the FUD scale, you surprised me with not wanting to link to jlime.com (front page. inside would be a different story)21:32
wpwrak(lawsuit map) yeah. sheer insanity21:32
wolfspraulI would need to ask a lawyer.21:32
wolfspraulyour lust for finding the corner case that is not yet thought through drove you to finding that one, which I don't have an answer for.21:32
wolfspraulthe question is whether a 'real' (<a href>) link to the homepage of a free software project can be used to uphold an infringement claim21:33
wolfspraulwhy is the link there?21:33
wpwrak;-) well, jlime's is about a lot of other things than spreading evil codecs. so it would seem fair (to me) to link to them. of course, i'm neither lawyer nor judge and it wouldn't be my head or wallet on the chopping block either.21:34
wolfspraulwhat does your company have to do with this free software project?21:34
wolfsprauldo you know that when this software is used on your hardware, you are infringing on these patents?21:34
wpwrakit uses some of their output.21:34
wolfspraulsee it would get close I think :-)21:35
wpwrakhmm, you mean there should be a warning telling customers NOT to install that package ? ;-))21:35
wolfspraulif I answer 'no' to the last one, they only need to find a small snippet from me somewhere21:35
wolfspraulha ha21:35
wolfspraulthat kind of thing won't work21:35
wolfspraulin fact if you want to play with them, they can get really angry :-)21:35
wolfspraulif I answer 'yes' to the last one, well it's close21:36
wolfspraulI know that I link to the homepage of a software project that could make my device infringe.21:36
wolfspraulwell I guess I can only pray then what the ruling will be :-)21:36
wpwrak(if this software is used) in the case of jlime, this would be the case if you either used all the software there or selectively the problematic bits21:37
wolfspraulas I've said before, from my experience and real legal advice. they just want to make it 'impossible' (for regular people) to play mp3 on that damned thing21:37
wpwrakif you pick a random package, chances are that it's perfectly harmless21:37
wolfspraulthat's all21:37
wolfspraulthey dont' want to argue at infinity about what 'impossible' means21:37
wolfspraulthey mean for 99% of normal people21:37
wolfspraulremember what I said above about the percentages of devices they are even catching each day21:37
kristianpaulwpwrak: ssop.fpd i dint understand where got it..21:38
wolfspraulit makes no sense to focus on that 1% of software engineers when there is a whole flood of unlicensed devices being sold to regular people every day21:38
wpwrakkristianpaul: http://projects.qi-hardware.com/index.php/p/f32xbase/source/tree/master/hw/modules/ssop.fpd21:38
wolfspraulif the device cannot play mp3 out of the box, and if it cannot be easily installed, we should be safe already21:38
wolfspraulif in addition there are no instructions or links on websites affiliated with the sellers, we are uber safe21:39
wpwrakwolfspraul: there must be some limit in how diluted the infringement can be21:39
wolfsprauleverything above that is totally hypothetical and a waste of time, imho21:39
kristianpaulwpwrak: hey when you create that proejct??...21:39
wpwrake.g., by making any device with a sort of cpu, you're creating the possibility of an infringement21:39
kristianpaulahh the siliconlabs thing21:39
kristianpaulgreat :) (for the ssop)21:40
wpwrakkristianpaul: quite a while ago. i'm now making a firmware programmer that doesn't have the chicken and egg problem. (or other excessive prerequisites)21:40
kristianpaulwpwrak: oh you want it self reproduce?21:41
wolfspraulyes but they cannot claim infringement for the cpu21:41
wolfspraulpuh I think we walked through so many times now21:41
wolfspraulthe infringement is for a whole machine that they can show to play mp321:41
wolfspraulthey need to show it!21:41
wpwrakwolfspraul: the term "affiliated" could use some clarification, too21:41
wolfspraulif you sell devices with an empty nand: no infringement21:41
wolfspraulthat's just how it is, legally21:42
wolfspraulno need to argue with me logically21:42
wolfspraulthe reason that (in most countries) selling software does not constitute an infringement is that software by itself cannot run21:42
wpwrakwolfspraul: not so fast :) empty nand plus link to firmware with codec = infringement21:42
wolfspraulwell sure, yes :-)21:42
wpwrakempty nand plus link to site where you can pick a codec or not = infringement ?21:42
wolfspraulinstallation, instructions for how to make it work _IF_ those things are coming from the seller/importer of the device21:43
wolfspraulyes of course, infringement21:43
wpwrakempty nand plus claim of compatibility with other company infringing = infringement ?21:43
wpwrakand so on21:43
wolfspraulif it's on my website, a website I control or link to21:43
wolfspraulhey I have a new question here21:44
wolfsprauldoes the patent stuff also apply to selling used goods?21:44
wolfspraulsay I sell my used mp3 player on ebay21:44
wpwrakit probably should21:44
wolfspraullet's imagine, hypothetically, that I sell it at a profit21:45
wolfspraulmaybe I even have a business of buying used mp3 players in bulk, and selling them on ebay21:45
wpwrakmay be just too disperse to hunt21:45
wolfspraulsure21:45
wpwrakthen you may no longer be disperse ;-)21:45
wolfspraulbut I never thought about this case, or discussed it with a lawyer21:45
wolfspraullike I said, we should stay real21:45
wolfspraulSisvel is hunting the big guys21:45
wolfspraulthey have a flood of unlicensed devices that are bypassing their taxation every day21:46
wolfspraulthe reason they came after Openmoko, most likely, was the corporate parents they thought they could drag into the system21:46
wpwrakany. pulster was probably more of an accident. or maybe they thought OM got so much media attention that a warning shot was on order.21:46
wolfspraulno it was the corporate parents they knew about and Sean talked about at every opportunity :-)21:46
wpwrakah well, maybe that, too21:46
wolfspraulthat's the first thing they wanted to know more about :-)21:47
wpwrakah ! :)21:47
wolfspraullike I said21:47
wolfspraulI'm all for trying to increase the quality of our free content21:47
wolfspraulalso the 'free IP' quality21:47
wolfspraulbut we also should think wisely how to use our limited resources21:47
wolfspraulendless discussions about totally hypothetical problems harm us more than they help21:48
wpwrakyup. hence avoid doing the reckless things. but there's no need to write the book on IP legal philosophy.21:48
wolfspraulwpwrak: so your patent policy is great21:48
wolfspraulyes!21:48
wolfspraulfully agree21:48
wolfsprauland let's keep exchanging experiences, factor the real world in21:48
wolfspraulif a patent owner contacts one of us somewhere, we'll work it from there21:49
wpwraki have that :)21:49
wolfspraulwe can proactively contact some, maybe Lattice, maybe SD consortium, ZigBee, etc.21:49
wpwrak(a clause about reacting to new attacks)21:49
wolfsprauland for some we just have to wait and see when reality hits21:49
wolfspraulof course we are already trying to do our best, 'avoid doing the reckless things' like you said21:50
wpwrakyup. there's no effective general defense anyway.21:50
wolfspraulmp3 is a patented/taxed technology21:50
wolfspraulwe all know it, so let's get over with it21:50
wolfspraulour products can't play mp3, done21:50
wolfspraulin a few years it's (hopefully) over21:50
kristianpaulwee :)21:51
wolfsprauluntil then we use reasonable defense mechanisms as we learnt, in order to avoid a successful infringement claim21:51
wolfsprauland first and foremost that means our devices must not actuall, to a regular person, be able to play mp321:51
wolfspraulit's actually quite simple :-)21:51
wolfspraulif that is the case, Sisvel has far more juicy targets to go after than hunting down some bizarre circumvention mechanism we are trading under the table here21:52
wolfspraulguaranteed!21:52
kristianpaulwhat about pantented video codecs?21:52
wolfspraulmpeg4, h.26421:52
wolfspraulsame thing21:52
kristianpaulahap21:52
wolfspraulnot all of mpeg421:52
wolfspraulI use wikipedia as my reference to determine which areas have active patent enforcers in them.21:53
wpwrakyeah. they're running out of time anyway, so they have to focus.21:53
wolfspraulcorrect!21:53
wolfspraultheir enemy is time21:53
wolfspraulthey need to make money fast, and in reality21:53
wpwrakat least we have one powerful ally ;-)21:53
wolfspraulnot theoretically/hypothetically21:53
wolfspraulSisvel is in business because they are able to enforce these patenst against microsoft, apple, motorola, etc.21:54
wolfsprauleverything else is just noise anyway21:54
wolfspraulit wasn't setup to catch pulster, and it will not stop whether they catch pulster or not21:54
wolfspraullet alone some individuals21:54
kristianpaulso after mp3 what is next , i think is widespread in all the world, so hard to replace?..21:54
wpwrakbtw, i thought about your statement that we're doing their work and globally make the patent stick. of course, one could say is that we do the GIF thing and devalue the patent.21:55
kristianpaulwpwrak: great !!! you have all kind of SSOP pins :D21:55
wolfspraulkristianpaul: don't understand. mp3 should be patent free in a few years. there is mp2 now, and Ogg Vorbis21:55
wpwrakkristianpaul: welcome to fped's world of plenty ;-)21:55
kristianpaulwolfspraul: yeah sure, i wonder what tjem Sisvel will look for money ig no more mp3 patent fees21:56
wolfspraulwpwrak: yes we are too small to make a difference either way.21:56
wpwrakkristianpaul: patented video codes will be with us for a while. well, google may be sabotaging that cartel a bit21:56
wolfspraulkristianpaul: ahh21:56
wolfspraul:-)21:56
wolfspraulyou start thinking about working for them?21:56
kristianpaulno21:56
wolfspraulWerner found some cruel links the other day...21:56
kristianpaulNO :)21:57
wolfsprauldon't worry I am sure they find more fuel for their engine21:57
wolfspraulthey have a number of 'programs'21:57
wolfspraulto them those are like products21:57
wolfsprauland of course they are looking for more input21:57
wolfspraulsubmitpatents@sisvel.com21:57
wolfspraulthey will continue to be with us in consumer electronics, I'm sure21:58
wolfspraulonce you have a successful business relationship with Microsoft, Apple, and all the others, you will try to find new ways of doing business together :-)21:58
kristianpaulhey you missed g00gl3 !21:59
wpwrakkristianpaul: if you're the kind who eventually snaps under work pressure and does a little massacre, perhaps then you SHOULD consider working for them :)21:59
wolfsprauldoes Google pay them? I don't know21:59
wolfspraulthey publish a list of licensees on their website21:59
kristianpaulwpwrak: no thanks :)21:59
wolfspraulof course the exact individual arrangements are not published, so you can see that list as propaganda as well22:00
wpwrakthere are so many people who function outside the usual system of detection and deterrence. someone who'd figure out how to use them would rule the planet :)22:00
wolfspraulhey I have some good news22:01
kristianpaul:) ?22:02
wpwrakthey're always welcome ;-)22:02
wolfspraulit seems Fedora Electronics Lab will jump onto the Milkymist wagon and try to support Milkymist development out of the box22:02
wolfspraulSebastien and Chitlesh had a good phone call, and setup a big wishlist for Milkymist in FEL22:02
kristianpaulah yes (old news for me)22:02
wolfspraulok sorry, if guys are following the milkymist-devel they know already22:02
kristianpaulreally BIG, but is great :)22:02
wolfspraulyes but Chitlesh is a very serious guy22:03
wolfspraulwork on that list will start, it will happen22:03
wpwrakwolfspraul: very nice indeed !22:03
wolfspraulso I'm very happy about that22:03
wpwrakwolfspraul: hopefully also generates some visibility22:03
kristianpaulso you can ship FEH with Milkymist as a full devel kit i the future22:03
wolfspraulI am thinking about a nice press release for 27c322:03
kristianpaulwpwrak: yes that too22:03
wolfspraulwe still have time for that22:03
wpwrakkristianpaul: i've been bitching about crappy PR for a while. now wolfgang is all set on PR ;-)22:04
wolfspraulhmm22:05
wolfspraulkristianpaul: good idea :-)22:05
wolfspraulon a bootable USB stick maybe?22:05
kristianpauloh yes :)22:05
wolfspraulI think most bioses support boot from usb nowadays22:06
wpwrakdiskless, with remote storage in a repository on project ;-) boot right into your CC-BY-SA project ;-)22:06
wolfspraulwell it's all a lot of work, let's see what we can get done22:06
kristianpaulanyway i relly want to see that FEL, so far i'll stick on debian22:06
wpwraks/project/projects/22:07
kristianpaulrelly/really22:07
wpwrakwe need a edit-bot :)22:07
wpwraks/a/an/ # argh22:07
wolfspraulwhen you swear at the edit bit it will stop working for 5 minutes22:07
wolfsprauledit bot!!!22:07
wolfspraul:-)22:07
wpwrakit's contagious :)22:08
wolfspraulhow can we visualize Milkymist?22:10
wolfspraulwe need one or two really strong pictures/photos22:10
wolfspraulwe have a nice logo, people like it22:10
wolfspraulbut it's very abstract22:10
kristianpaulqemu...  may work22:10
kristianpaulahh that visualize !22:10
wpwrakoff to the kitchen, gotta do some etching22:10
wolfspraulkristianpaul: yes, for the press release22:11
wolfspraulmaybe some silicon/gate-level stuff?22:11
wpwrakwolfspraul: show some cool VJ application ? and the box overlaid ?22:11
wolfspraulbut that's too app-related22:11
wolfspraulthis is a Verilog SoC project22:11
kristianpaulwpwrak: there is no such us thing yet22:11
wolfspraulit's about Intel, ARM, MIPS22:11
wpwrakan app EXAMPLE :)22:11
wolfspraula picture of a retro-looking DIP chip?22:11
kristianpaulhmm22:12
wolfspraulI don't know what works.22:12
wolfspraulof course Sebastien has to like it too.22:12
wolfspraulbut we can help him with some suggestions to pick from22:12
kristianpauloops?22:12
wolfspraulon one hand it's software, source codes22:13
wolfspraulbut screenshot? what screenshot?22:13
wpwraknaw, "see what you can do with it". and it's adaptable. adaptability is hard to put into pictures. http://www.heise.de/open/artikel/Die-Woche-Nichts-geht-verloren-1095937.html22:13
kristianpauli dont know if the NO ARM campaing is not too nasty (for then )22:13
wolfspraulhave no browser right now, will look at it later :-)22:13
wolfsprauldon't be negative unless you have to22:13
wolfspraulwe don't have to22:13
kristianpaulok22:13
wolfspraulplease nothing negative about others22:13
wolfspraulARM is great, we are better22:14
wolfspraul:-)22:14
wolfspraulmy problem right now is a really strong picture22:14
kristianpaulhehe22:14
wpwrak("you don't need no stinking connectivity" giggle)22:14
wolfspraulit's not urgent22:14
wolfspraulif you google for milkymist you will see the problem right away (images.google.com)22:14
wolfspraulthe best you can find are those milkdrop effects22:14
wpwrakask sebastien if he has some impressive-looking VJ screenshot ? it just has to look good, catch the eye.22:15
wolfspraulbut that's milkdrop, screenshots of a piece of software that you can also run on your Windows PC22:15
wolfspraulno he won't have it, and I still don't like it22:15
wolfspraulif Sebastien likes it fine :-)22:15
wolfspraulI need something that visualizes an IC design22:15
wolfspraulso it fits in a line with Intel, ARM, MIPS22:15
kristianpaulthe IC with the SoC inside is not bad idea22:16
wolfspraulif we do a press release, we can have 2 or 3 good pictures, and they can be very different in terms of their perspective on the project22:16
wolfsprauljournalists picking the story up will choose which one they think works anyway22:16
wolfspraulkristianpaul: what do you mean? "IC with SoC inside"?22:17
wolfspraulhow does that look like?22:17
wolfspraulIntel always has those clean-room pics22:17
wolfspraulor the wafer22:17
wolfsprauleverybody seems to love wafers22:17
wolfsprauland some guy in a white suit holding the wafer22:17
kristianpaulthe IC is semi-trasparent22:17
kristianpauland a small SoC picture inside that22:17
wpwraki'd keep the image to make it clear what it can do. after all, it is optimized for fast graphics.22:17
kristianpaulhmm22:17
wolfspraulthe press release should be about free IC design, free CPU22:18
wolfspraulI think22:18
wolfspraulput it on the map as an alternative to ARM, MIPS22:18
wpwraksure, write about all that. but show the stuff that catches the eye.22:18
wolfspraulkristianpaul: i cannot imagine how what you are describing looks like :-)22:18
kristianpaulhmm may the the IC wafle (yummy) and then a SoC over it22:19
wpwrakyou're unlikely to successfully introduce a sign language for "this is reconfigurable hardware implementing a SoC"22:19
wolfspraul"semi-transparent IC with small SoC picture inside"22:19
kristianpaulhehe22:19
wolfspraulnot 'sign language'22:19
wolfspraulit doesn't need to be a big thing22:19
wolfspraulcan be any shot22:19
wpwraktoo complex. you can have a diagram with the subsystems superimposed on a big chip. write FPGA somewhere.22:19
wolfspraulI'm just saying if we want to do a press release, we need 2-3 nice pics22:19
wolfsprauland we don't have them today22:19
wolfspraulthe best we have are the milkdrop screenshots22:19
wolfspraulwhich I think are totally misleading22:20
kristianpaulindeed the SoC is no where there22:20
wpwraksomething like fig 1-1 of Jz4740_ds22:20
wolfspraulof course we could just take a few lines of Verilog in a large font (like 2-3 lines only in the picture)22:20
wolfspraulbut something that makes people think, something nice22:20
wolfspraulthe Verilog can even be fake22:20
wolfspraulbut in Verilog syntax22:21
kristianpaulmay be the current SoC combined with some IC behind..22:21
wpwrakmilkdrop on a big screen in the background. pcb or just the big chip in the front.22:21
wpwrakpeople who'll recognize the verilog know what an FPGA is :)22:22
wolfspraulyeah some little logic like if (cpu == arm) { switch_to_something_better(milkymist); } in Verilog syntax22:22
wolfspraulof course more interesting that my if()22:22
wpwraki thought "no negative" ? :)22:22
wolfspraulsure sure22:22
wolfsprauljust my idea22:22
wolfspraulcouple lines of large font, verilog syntax, and some text in it with the message22:22
kristianpaulyou said before the MM One was the apple I for this copyleft era22:23
wpwrakfine. overlay the milkdrops with a verilog listing in the classical green CRT style.22:23
wolfspraulif (want_free_cpu) use_milkymist(); else use_arm_mips_x86();22:23
kristianpaul:D22:23
wpwrakput the chip/board in front.22:23
wpwrakdon't make the verilog _say_ anything. just make it look like it's doing something important22:24
wolfspraulsure22:24
wolfspraulthat's another way22:24
wolfspraulwell we have time for this22:24
wolfsprauland maybe we should play with several ideas22:24
wolfsprauljust a heads up22:24
wolfspraulif we do a press release for 27c3, it's still 2+ months out22:24
wolfspraulI think the press release should first introduce milkymist, not Milkymist One22:24
wpwrakhas sebastien killed the stability bug ?22:25
wolfspraulMilkymist One can be mentioned a little into the text, as a way to jump in, together with FEL (quote from Chitlesh etc)22:25
kristianpaulwpwrak: hey that FT232RL dingle is so quite :)22:25
kristianpauls/dingle/dongle22:25
wolfspraulno stability bug still there22:25
wolfspraulhe is trying the new ISE 12.3 now22:26
wpwrak(intro) perhaps both. describe the concept, then the implementation22:26
kristianpaulit seems s LM32 bug22:26
wolfspraulso that is a big risk for RC2, but I'm taking it anyway22:26
kristianpaulSDRAM passes memtest22:26
wolfspraulah not sure22:26
wpwrak(mem bug) argh :-(22:26
wolfspraulcould be anywhere22:26
wpwrakwell, it's reconfigurable. you can fix it in the field ;-)22:26
wolfspraulYanjun Luo keeps saying he suspects power issues.22:26
kristianpaulyeah :(22:26
wolfspraulnot if it's a power related issue22:26
kristianpaulbut is on ML401 too22:26
kristianpaulso...22:27
wolfspraulRC2 has a number of improvements on the power side22:27
kristianpauleven witht the other propietary xilinx tool still the bug present22:27
wolfspraulno more shared GND vias, some additional/better capacitors22:27
wpwrakyou can add heuristics that slow down a little when power may run out ;-)22:27
kristianpaullol :)22:27
wolfsprauleventually we will fix it22:27
wolfspraulI like tough bugs.22:27
wpwrakc'mon, Intel and AMD are doing very similar stuff to keep their chips from overheating :)22:27
wolfspraultough bugs are the best criteria for determining whether a project succeeds or fails.22:27
wolfspraulso the earlier the better, then we know our fate :-)22:28
wpwraki can pass that challenge ;-)22:28
wolfspraulyou know, when Dieter fixed the #1024 bug, after 1 year, on the same day that everybody was fired (coincidence), I knew we were on the right track, just cut off by investors with more money than brain.22:28
wolfsprauleverybody can start with some crap, and the moment they run into a tough bug they give up, start another stupid project, etc.22:29
wpwrakyeah :) i actually felt really good about the state of OM just before the killing.22:29
kristianpaultaking about invetor how freerunner/openmoko got funded?22:29
wolfspraulbut if you are able to fix the toughest bugs, you have a very strong argument to work with real customers.22:29
kristianpaulinveor/investors22:29
kristianpaultaking/talking22:30
wolfspraulso I like the milkymist instability bug, and so far we are doing all the right things to track it down22:30
wpwrakwolfspraul: not sure if customers think like this. as a customer, i'd prefer to never know of any hard bugs you may have fixed :)22:30
wolfspraulfirst and foremost Sebastien of course, but some people can repro it too and are helping22:30
wolfsprauloh no22:30
wolfspraulas a customer, the only thing you care about is that the guy selling you something won't go away when there is a REAL problem22:31
wpwrakwolfspraul: it's a bit like people praising the great warranty repair service of some companies. i never quite understood their logic :)22:31
wolfspraulbecause 9x% of 'partners' are gone in that case22:31
wpwrakyup. open is the key. no single point of failure.22:31
wolfspraulok but that not's a good argument to propose yourself to a customer22:32
wolfspraulbut of course yes, it is an advantage22:32
kristianpaulaso if the customer is an artist..22:33
kristianpauls/aso/also22:33
kristianpaulwpwrak: definitelly i'll fork your work (f32xbase)  have all i need22:34
kristianpaulalso a Makefile :)22:34
wpwrakanyway. time for acid. make the boards really cute :)22:34
kristianpaulwich ones??22:34
wpwrakthe c2usb boards22:45
wpwrakthe design is still untested22:46
wpwrakbut with some luck, this will change tonight :)22:47
kristianpaulbtw i editing the ft232 to fit my tlv1562 by hand (editor) i guess is okay22:48
kristianpauli aslow ill rename it22:48
kristianpaulbest wishes !22:48
wpwraknow, the soldering ...22:58
kristianpaulhey that was fast !22:58
wpwrakyeah, i got a good mix with the acid. very hungry :)23:00
kristianpauli hope do my acid this weekend finally, now that board thickness dont stop me :)23:00
kristianpaulhmm i need flux (added TO DO)23:00
wpwraki recommend water soluble. it's quite conductive but it's easy to get rid of.23:05
kristianpaulhow i ask for that in spanish?23:07
kristianpaulremenber i live in a small town at weekends :)23:08
kristianpaulthat is when have time go shop23:08
wpwrakask them to show you their fluxes :)23:24
wpwrakbut in a small town, they may not have the good fluxes23:24
kristianpaulis a marker23:25
kristianpauli saw it23:25
wpwrakmarker is good23:30
wpwraknow .. what to glue under the board to make it a little thicker ...23:31
wpwrakand thw winner is "melamina"23:37
kristianpaulhmm how i can visually chnage the pin position for this lib..23:40
kristianpaulpins names and number are ok, just messed up over the U11 in a non nicelly order :/23:40
wpwrakand it enumerates. hehe :)23:41
wpwraksomething liek M to move ? or right-click and then it's somewhere23:42
kristianpaulbut M move the shole IC23:43
kristianpaullook http://wikisend.com/download/473800/Pantallazo-13.png23:44
kristianpaulhmm fped..23:44
kristianpaullet see23:44
kristianpaulah no this is not fot that23:45
kristianpauli think is in the component.lib file just move pins in text mode is not fun..23:46
--- Fri Oct 8 201000:00

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