**** BEGIN LOGGING AT Thu Mar 26 02:59:58 2015 Mar 26 04:31:43 hmmmmm Mar 26 04:31:47 rather quiet Mar 26 04:51:04 * mr_science tired from 4 months of "conferencing" Mar 26 04:52:12 ds2, good to meet you in person today... unexpected highlight of the trip... Mar 26 06:06:25 vvu: you there? Mar 26 06:12:06 mr_science: ah... same here. Mar 26 06:12:16 now to catch up with commenting the proposals Mar 26 06:14:44 ds2: I am submitting a proposal on the cubesat project. As suggested in the ideas page i want to write code for the PRU to collect and filter magnetometer data for further processing. Can you suggest how it will help the community. I need some feedbacks for my proposal Mar 26 07:01:14 k12: i'll be around, ping me if interested Mar 26 08:21:14 ua8: ping Mar 26 08:21:36 vvu:pong Mar 26 08:22:03 you pinged me earlier a bit Mar 26 08:22:50 hey, i looked at your comment on my proposal, i have created a kext,https://gist.github.com/aroraujjwal8/a7b8a22bcd5a33ee1a78, can you check it? Mar 26 08:23:48 you havea beaglebone black ? Mar 26 08:23:51 have a* Mar 26 08:24:02 nope Mar 26 08:24:23 maybe i have some time tonight to test the kext Mar 26 08:24:42 that will be great, thanks Mar 26 08:25:38 and can you tell me in detail about the problems you mentioned on working with windows Mar 26 08:26:08 the device enumerates but after you attach a driver to it (winusb, libusbK) it stalls Mar 26 08:26:13 which should not happen Mar 26 08:26:48 any ideas, why? Mar 26 08:26:51 no clue Mar 26 08:30:16 does it has anything to do with rndis, i read somewhere that it will need a vendor supplied inf file to operate , have you heard about this before? Mar 26 08:30:42 i tried the default RNDIS driver from windows but same result Mar 26 08:33:26 okay, i’ll do some more research on it, is there anything else i need to work on by the time? Mar 26 08:48:39 Abhishek_:Hi we discussed about my proposal last night regarding the filter implementation on the PRU. So can you suggest for now how to proceed with the proposal? Mar 26 08:50:23 coolsounak101: Submit at least something to Melange, you can edit it later as time is short Mar 26 08:50:38 Less than 48 hours now Mar 26 08:52:23 Abhishek_: Ok I wil upload whatever I have pepared today evening. Mar 26 08:59:46 alexanderhiam: ping !! Mar 26 11:36:58 Hi Mar 26 11:40:03 alexanderhiam: Mighty ping !!! ;-) Mar 26 11:40:16 jkridner: ping :D Mar 26 11:41:24 I have a question about the proposal template. Who should I contact? Mar 26 11:50:30 jkridner: really need your review on my proposal... running out of time :( Mar 26 11:54:09 vvu: any updates with the sniffer? Mar 26 11:56:30 hendersa : proposal submitted for Demo app using BBBAndroid.. have a look.. :) Mar 26 12:02:15 webglider: got you some feedback Mar 26 12:05:30 is there any logs from the meeting available? Mar 26 12:05:40 i don't see it in the one topic linked Mar 26 12:13:04 jkridner: thanks Mar 26 12:13:16 kiran4399: reading yours now Mar 26 12:24:31 jkridner: pls read mine also.. :) Mar 26 12:24:39 will do Mar 26 12:25:02 thanx.. Mar 26 12:29:28 jkridner: could you also please give me your answer for 'if implemented how will it benefit the community?' Mar 26 12:29:55 jkridner: I have got to tell you one important thing... Mar 26 12:30:05 ? Mar 26 12:33:01 jkridner: after discussing with alexanderhiam, I plan to scrap out RPC server... when the respective pruduino object is called by the node interpreter.. the invoked object will build a script and append it to the main script.. after the node interpreter completes the script generating process .. that botspeak script will directly by loaded to the prupseak intpreter via the existing botspeak server.. It's kinda late to tell you Mar 26 12:33:49 kiran4399: does your proposal include this? do you have time to update it? Mar 26 12:35:05 jkridner: Well, it is not so big change... I am ready with the text .. I just wanted to update it now.. Mar 26 12:35:51 jkridner: so basically .. it is like making the pruspeak code generated by the node side only... This implementation will reduce many delays... Mar 26 12:35:59 jkridner: what do you think about this ?? Mar 26 12:39:57 jkridner: I wanted to discuss this with alexanderhiam. ... But most of the content in my proposal will remain the same.. Mar 26 12:40:16 jkridner: I promise.. everything will be perfectly done by tonight !!! Mar 26 12:40:40 I'm OK with pruspeak code generation by the node-side only. That's the way I'd expect it. Mar 26 12:44:15 jkridner: Do you want me to answer the notes here or as a comment? Mar 26 12:44:30 in the notes and in proposal improvements. Mar 26 12:44:52 jkridner: I'm not able to find any reliable node based web terminal emulators Mar 26 12:45:05 jkridner: any ideas on what can be done as an alternative? Mar 26 12:45:15 webglider: Cloud9 IDE now has a very good one. Mar 26 12:45:24 it didn't used to, but it does now. Mar 26 12:46:25 jkridner: ok cool, then nothing extra will have to be integrated Mar 26 12:46:52 jkridner: could you please give me your answer for 'if implemented how will it benefit the community?' Mar 26 12:55:04 jkridner: Shall I remove the project goals and just keep the project deliverables ? Mar 26 13:19:38 alexanerhiam: ping :) Mar 26 13:21:51 Abhishek_: I have uploaded my proposal. Can you review it? Mar 26 13:22:14 my mail id is sounakranjandas9@gmail.com Mar 26 13:40:22 geekswine: what is your ID on melange? Mar 26 13:44:11 jkridner: Can I merge the contents of project goals in deliverables and remove project goals ? Mar 26 13:44:33 jkridner: looks like my proposal is long .. Mar 26 13:46:03 jkridner: by ID do u mean username ?? Mar 26 13:50:36 username : geekswine Mar 26 13:50:37 public name : Vinit Garg Mar 26 13:52:37 jkridner: can you review my proposal? my username is coolsounak Mar 26 13:59:28 geekswine: I like the layout of your proposal pretty well, but I didn't check for all the entries against the template. Mar 26 13:59:41 geekswine: my queries back were pretty high-level. Mar 26 13:59:46 sidbh_: there ?? Mar 26 14:00:03 kiran4399, yes Mar 26 14:01:51 sidbh_: while using starterware to access peripherals... I hope there won't be any I/O conflicts ..right ?? Mar 26 14:02:02 jkridner: i'll try to ans all the queries which u had and will be back to u soon..thanx :) Mar 26 14:02:05 I mean I/O conflicts b/w pru and arm.. Mar 26 14:03:59 jkridner: have u seen my proposal ? Mar 26 14:04:34 kiran4399, yes there will be, If both arm and pru try to access same peripheral it will lead to unknown behaviour of the peripheral, although I don't think system as a whole should fail under any circumstance during such scenario Mar 26 14:05:25 sidbh_: So how should we avoid this ? Mar 26 14:05:53 I think care should be taken during development by the user that the implementation doesn't have any such conflict in there system Mar 26 14:06:51 kiran4399, I think avoiding this at library level would be tricky Mar 26 14:07:03 jkridner: how abt other fields ?? is there something more need to be elaborated on ?? how abt overall proposal ?? Mar 26 14:08:23 kiran4399, although one thing you can do is do all the initialisation, i.e. one time stuff in userspace and leave only iterating parts to be done on PRU Mar 26 14:08:57 initialisation on ARM Mar 26 14:09:05 Iteration on PRU Mar 26 14:09:32 sidbh_: Oh .. OK .. then if any other process on the ARM tries to use that pin.. there will be an error .. right ?? Mar 26 14:09:51 no error won't be generated Mar 26 14:10:02 reading won't be the problem Mar 26 14:10:39 only writing, if both try to write you won't be able to know which write actually stayed, thus the unknown behaviour Mar 26 14:12:40 sidbh_: But we do not want that happen, so can devise any mechanism by not allowing other process to write to that pin which PRU is using ?? Mar 26 14:12:48 sidbh_, alexanderhiam : need some feedback on my proposal please Mar 26 14:13:08 kiran4399, or the inverse Mar 26 14:13:48 sidbh_: I did not understand that.. Mar 26 14:13:59 jkridner : please review my proposal Mar 26 14:14:10 kiran4399, don't let the PRU use the pin which ARM is using Mar 26 14:15:40 sidbh_: My main focus is PRU.. so don't want that inverse.. If PRU is using a pin.. how to make ARM not use that pin ? Mar 26 14:20:58 sidbh_: ARM will write to the GPIO by sysfs writes.. am I correct ?? Mar 26 14:21:42 kiran4399, yes, but there are other ways too Mar 26 14:21:52 like mmap Mar 26 14:23:31 sidbh_: OK.. is there any method by which we can close the access to the GPIO from ARM temporarily ? Mar 26 14:24:07 kiran4399, I don't think so Mar 26 14:24:27 kiran4399, but inverse is Mar 26 14:24:54 just put the PRU in sleep Mar 26 14:24:55 sidbh_: alright... can you please tell me how to do the inverse ? Mar 26 14:43:37 i want to ask that how the project "Library of Arduino-compatible functions for StarterWare" will be used by the community ?? how will it help further ?? anyone plss ?? Mar 26 14:53:02 nodebotanist, julianduque: have you guys started commenting on project proposals on Melange? Mar 26 14:53:34 geekswine: perhaps some people in #beagle or #sparkfun could say if they'd use it. Mar 26 14:54:02 also #i3detroit has some BeagleBone users that are familiar with Arduino. Mar 26 14:55:49 alexanderhiam: ping :D Mar 26 14:55:50 jkridner :Hi, my project is orbital imaging cubesat . Mar 26 14:56:10 altairpearl: has nerdboy already reviewed it? Mar 26 14:56:21 yea he did. Mar 26 14:56:23 calculus? Mar 26 14:56:38 nope . Mar 26 14:57:07 alexanderhiam and nerdboy have review my prposal and asked me to make some changes Mar 26 14:57:11 i have done them Mar 26 14:57:31 i just need to know the projects impact on the community Mar 26 14:57:58 calculus : ping Mar 26 14:58:40 jkridner: ok..i'll ask there also..and what about the upstreams of the project ?? wouldn't the other projects can be made using this as it will be fast ?? Mar 26 14:59:29 maybe, but you want to think about future maintainership and exposure. Mar 26 14:59:37 there should be some obvious upstreams. Mar 26 14:59:42 jkridner: and mine? Mar 26 15:00:35 making a cool open source project is great.... improving an established project with a community already around it is greater. Mar 26 15:04:11 hi jkridner Mar 26 15:04:33 ebadawy: The bonecard concept is to have a fixed-sized layout. Mar 26 15:04:38 okk...so should I continue with this project or should I collect the thoughts from the other groups abt this project and continue with it ?? Mar 26 15:06:42 jkridner: it's fixed ,, the slider is for switching between the code and other notes like a video or wiring hints ,, all of them are on a bonecard Mar 26 15:06:48 ebadawy: I see you are looking to split the window and show JS on the same card. Mar 26 15:07:54 yea ,, there is two suggestions that i came up with the tutorial page Mar 26 15:08:12 ebadawy: thanks to the cards all being fixed sized, it is also possible to create additional views of the same content, including having all the cards visible by scrolling. Mar 26 15:08:50 ebadawy: has DiegoTc seen your proposal yet? Mar 26 15:09:04 jkridner : what if I find a lot of inaccuracy/vagueness in a proposal? Mar 26 15:09:32 what exactly should I write as a comment? a general feedback or something else Mar 26 15:10:04 karki: yeah, sometimes some general feedback is all you can give. Mar 26 15:10:12 yes he is ,, but not show if he saw that part of the slider as i just made it yesterday according to his suggestion about to have a fixed size bonecard Mar 26 15:10:13 sure* Mar 26 15:10:23 karki: but, if you can give some examples of the type of thing that isn't clear, that'd likely be helpful Mar 26 15:10:33 jkridner : will do :) Mar 26 15:11:14 vmayoral: what kind of things you expect in that porting openrov to apm project. Mar 26 15:11:33 in proposal Mar 26 15:11:41 ketul: did you read through the mailing list and wiki? Mar 26 15:12:11 ketul: i'd expect a decomposed list of tasks to address the complexity Mar 26 15:12:25 yep I did Mar 26 15:12:30 ebadawy: I prefer having details in the project plan.... like a series of small 1-2 week projects with enough explanation that it is clear what will be done in that time frame.... Mar 26 15:12:31 ketul: i'm happy looking at a draft of milestones if you share it Mar 26 15:13:30 having this big separation between talking about design details and then really quick notes in the timeline makes me need to scroll back-and-forth a lot wondering if it is clear what needs to be done that week. Mar 26 15:14:10 I'd prefer to have a signed-off design as a deliverable with some hints/examples you know how to create/implement designs. Mar 26 15:14:22 yeah ok Mar 26 15:15:04 jkridner: Can you please look into my proposal? Mar 26 15:15:15 my id is coolsounak Mar 26 15:15:33 vmayoral: So will it be a similar kind of steps like that you did for beaglepilot Mar 26 15:15:37 ? Mar 26 15:15:57 ketul: yes Mar 26 15:16:04 let me fetch my proposal notes for you Mar 26 15:16:06 * jkridner isn't sure if that aligns with Abhishek_, karki, av500, calculus, ds2, julianduque, nodebotanist, vvu, but I don't like having a huge project description and a tiny timeline... I'd prefer the timeline be clearly readable on its own and the project description be a fairly simple read as well. Mar 26 15:16:16 please Mar 26 15:16:35 https://docs.google.com/document/d/1oL0NU4d5ZiujAuXJmI40pziRQlXCEmAf5TgK6NVDap8/edit#heading=h.ukpggdendfac Mar 26 15:17:17 can any of the mentors review my proposal? Mar 26 15:19:02 i can edit m proposal even after submitting it on melange ? Mar 26 15:19:07 *my Mar 26 15:19:42 yes Mar 26 15:19:47 and i have to submit it there to get feedback from community member ? right ? Mar 26 15:19:48 until tomorrow midnight Mar 26 15:20:38 Abhishek_: I ahve submitted my proposal on melange. Can u just review it? my id is coolsounak Mar 26 15:21:10 vmayoral: so this time as you r providing erlebrain so interfacing of firecap and pixhawk wont be there right? Mar 26 15:21:10 coolsounak101: so you know, we see the proposals by title and name. turns out we don't see the id easily on the list. Mar 26 15:21:42 jkridner: so u want the implementation plan and time line be in one section, right? Mar 26 15:21:50 * jkridner wishes some of the students with proposals in would idle here. Mar 26 15:22:07 Abhishek_: I didnt get it. How can i get it reviewed? Mar 26 15:22:17 and to have more details on it Mar 26 15:22:24 ebadawy: for me, the time-line should be the implementation plane, but maybe I need to read the template better. Mar 26 15:22:35 coolsounak101: Didn't get what? Mar 26 15:22:41 I have submitted a proposal but not been able to view it in my Dashboard! Mar 26 15:22:51 there should be some overview, but the timeline should make it clear to me what you are going to be implementing and how each 1-2 weeks. Mar 26 15:22:59 the procedure to get the proposal reviewd... Mar 26 15:23:28 ? Mar 26 15:24:21 coolsounak101: When you submitted that, mentors received an email with the title of your proposal Mar 26 15:24:36 coolsounak101: can you explain in a few words why you think it is appropriate to do magnetometer work inside the PRU? Mar 26 15:25:42 jkridner: yea there are same, i just split it to make it easy to know when things will be done Mar 26 15:25:58 calculus , ds2 : can you guys review my proposal please Mar 26 15:26:22 altairpearl: can you give me the name of the proposal? Mar 26 15:26:23 ds2: PRU is used mostly for real-time processing. So it would be Mar 26 15:26:36 good to collect the data in real time Mar 26 15:27:21 orbital imaging cubesat Mar 26 15:27:21 ds2: mostly magnetometers are used for getting directions or tracking positions. So in space it is very important to get that data in real-time. Mar 26 15:27:30 ds2 : ^ Mar 26 15:27:36 coolsounak101: define "real time" in this context please Mar 26 15:28:10 altairpearl: what is the real name it is under? (reading it using email - quicker then via the melange interface) Mar 26 15:28:26 ?? Mar 26 15:29:04 anand17GSoC: Title of your proposal? Mar 26 15:29:08 jkridner: for BBUI, it's realy not easy for my to make new suggestions as I don't have a BBB to try the current version, DiegoTc told me that he will make a video for that but seems that he was busy these days Mar 26 15:29:12 So in order to track the position or the direction of the cubesat we need to get the current magnetometer readings. By real-time I mean the PRU will capture the magnetometer data in a certain time window process it and then move on to the next window Mar 26 15:29:44 ebadawy: fair enough that the hardware interactions are a limitation.... Mar 26 15:30:03 ebadawy: I've seen some people hack the hardware access code to make it do some things without hardware. Mar 26 15:30:32 ds2: am i clear? Mar 26 15:30:34 ebadawy: just looking at the JavaScript itself should give some thoughts, especially if compared to my (very incomplete) re-write. Mar 26 15:30:37 coolsounak101: ok... do you know what magnetometer you are using? Mar 26 15:30:59 ebadawy: I'd just want to be sure you can immediately jump in the code and be effective with it. Mar 26 15:31:00 jkridner: I want to merge the contents of project goals to the deliverables section as it is becoming too long.. what do you think ? Mar 26 15:31:02 i have submitted my proposal, please review it. my id is zeekhuge. Mar 26 15:31:13 ebadawy: the hardware interface portion turns out to be quite minimal. Mar 26 15:31:15 Utilize the Energia fork of Arduino to push support for BeagleBone and BeagleBone Black Mar 26 15:31:19 alexanderhiam:I have submitted my proposal, please review it. my id is zeekhuge. Mar 26 15:31:27 coolsounak101: there are complexities in dealing with the mag that defeat the point of doing it on the PRU so I am looking for a reason why it should be there Mar 26 15:31:34 I have worked with hmc5883l but in this project i am not sure with what magnetometer i am going to work Mar 26 15:32:04 karki: my proposal is up there, please review it. my id is zeekhuge. Mar 26 15:32:36 kiran4399: yeah, I thought it was becoming too big. Save off your work, but I agree that combining some areas would be good. personally, I'd put the meat into the timeline. Mar 26 15:32:44 ds2: can u elaborate on the complexities? Mar 26 15:32:51 * jkridner just read a proposal that did that really nicely, but doesn't think he can share it. Mar 26 15:32:56 coolsounak101: for something like the HMC5883, I'd say there is no reason to do it on the PRU. I2C is not realtime. And unless you are bitbanging it locally, getting to the I2C block is going to make it non RT Mar 26 15:32:58 jkridner: ok ,, then i think that i should have another deeper look on the code Mar 26 15:33:17 ebadawy: great. Mar 26 15:33:33 coolsounak101: what's more - the HMC5883 wants a reset/recal periodically which makes things really screwy Mar 26 15:34:21 ds2: So what do u suggest? Mar 26 15:34:31 abhishek? Mar 26 15:35:02 coolsounak101: do it in the ARM Mar 26 15:35:11 anand17GSoC: What is the name of your proposal? Mar 26 15:35:26 Utilize the Energia fork of Arduino to push support for BeagleBone and BeagleBone Black Mar 26 15:35:51 I don't see ot Mar 26 15:35:54 *it Mar 26 15:36:03 jkridner: in the timeline and milestones section can I point to the deliverable's by saying...."by the end of the week deliverable-2 will be achieved" or something like that... ?? Mar 26 15:36:04 what can i do to make it sure? Mar 26 15:36:10 resend it? Mar 26 15:36:12 coolsounak101: how non RT can your algs tolerate things? Mar 26 15:36:25 ds2: I have no issues doing it in ARM. I have already done it once, but the ideas page suggests to do the data collection and the processing in PRU Mar 26 15:36:36 anand17GSoC: What is your real name? Mar 26 15:36:52 Anand Suthar Mar 26 15:37:06 anand17GSoC: Resubmit it Mar 26 15:37:11 kiran4399: my preference is to instead move deliverable-2 into that point in the timeline and remove it from the earlier description. Mar 26 15:38:18 when I read these proposals, I want a nice *high-level* overview of what you are going to be doing, how it is useful, what are the challenges, etc. at the top, then.... Mar 26 15:38:48 I want to see week-by-week little tasks that will ultimately achieve that vision. Mar 26 15:39:14 a huge benefit to me is that even if the end-end-goal isn't reached, that every week or two something notable is improved toward that end goal. Mar 26 15:39:30 * jkridner isn't happy with the way some projects tried to bite off too much at a time. Mar 26 15:40:04 if you can't describe with a single paragraph what you'd do in a week or two such that it is clear, you are probably trying to do too much. Mar 26 15:40:19 av500: ^^^ do you agree with these statements? Mar 26 15:40:38 oolsounak101:yes, and my comments are - what's the point. if it is to be done in the PRU, I'd like to see some justification Mar 26 15:40:59 jkridner: would it be good if have some code to support the deliverable in the timeline section ? Mar 26 15:41:18 while we are beagle fans - there is more value to having algs that work on other parts w/o PRU unless the PRU is absolutely needed. Mar 26 15:42:23 * Abhishek_ recalls ds2 saying - the PRU is not a panacea Mar 26 15:42:58 :) Mar 26 15:43:01 well it is saying it exceeded the word limit! Mar 26 15:43:15 but in real it is far more less Mar 26 15:43:23 Abhishek_: i have submitted my proposal, please review it. Id=zeekhuge Mar 26 15:43:34 and also a data saved successfully message was displayed Mar 26 15:43:56 For General Information of all students: Please refer to your proposals by the actual title, not by your user id's please Mar 26 15:44:34 Abhishek_,: proposal title : BeagToPr Mar 26 15:44:48 both if possible Mar 26 15:44:57 alexanderhiam: my proposal title is BeagToPr Mar 26 15:48:12 ds2: i think the cubesat will transmit the magnetometer data continuously to earth in order to get the location of the satellite. In that case we need the data in real-time and free from noise. That may be the reason to implement it in PRU Mar 26 15:48:33 jkridner:I have submitted my proposal on bonescript please review and give me the feed back Mar 26 15:49:57 coolsounak101: HOW real time? mini mags like the HMC5883 are very noisy Mar 26 15:53:06 will review more later. Mar 26 15:58:37 ds2: thats why we are using two mags and some noise reduction algorithms deal with it Mar 26 15:59:44 coolsounak101: So you going to execute these algorithms on the PRU? Mar 26 16:00:43 jkridner: alexanderhiam:I have submitted my proposal on bonescript please review and give me the feed back Mar 26 16:01:12 coolsounak101: how do you sync up the 2 mags? Mar 26 16:02:25 alexanderhiam: ping Mar 26 16:02:29 ds2: I have not yet validated whether the filter can be implemented in PRU. But we can always give it a try right? Mar 26 16:02:33 jkridner I am of the long timeline short description camp for proposals Mar 26 16:02:53 The timeline convinces me whether or not the proposal is too ambitious Mar 26 16:03:44 coolsounak101: that makes it a bit too vague for a GSoC project Mar 26 16:03:50 GSoC is SHORT Mar 26 16:06:14 ds2: i have iplemented filter algorithms on DSP cores. I think with little bit of optimization we can make it work. But for that we have to at least try it. Mar 26 16:06:39 coolsounak101: how many tap are your ags? Mar 26 16:06:41 algs Mar 26 16:08:05 I would normally prefer around 25 taps. But if the memory does not permits we can always reduce the taps and do a trial and error. Mar 26 16:09:11 ds2: If the adaptive filter entirely does not work we can always use an FIR filter which would be very easy if we create a lookup table to store the filter coefficients and just do the convolutions in PRU Mar 26 16:09:58 coolsounak101: I think you need to address these uncertainities Mar 26 16:10:31 ds2: but untill and unless we start working on it how do u propose to do that? Mar 26 16:12:27 coolsounak101: problem is there isn't time in GSoC to do all that Mar 26 16:13:09 ds2: but we got 11 weeks right? I am sure i can come up with something in this 11 weeks Mar 26 16:16:18 whoa, it seems the number of proposals has doubled overnight Mar 26 16:18:41 hello want to ask if I apply in beagleboard org and in other also . then whether would it impact? Mar 26 16:19:56 alia: you mean can you apply to two organizations? Mar 26 16:20:05 `yep Mar 26 16:20:50 I think I can but will it create negative impact ? Mar 26 16:21:35 I don't see anything saying that you can't, but you can only work on one project, so it would require the two organizations communicating to work that out Mar 26 16:22:13 how that can happen i mean the communication? Mar 26 16:22:45 orgs only have a fixed number of slots, so if you got accepted by both you'd then have to drop one, meaning that org could have used that slot for another project Mar 26 16:23:07 I'm inclined to say it's too much hassle Mar 26 16:25:10 alia: yes you may apply to as many orgs as you want but I personally believe quality is better than quantity Mar 26 16:25:19 because there wont be any guarantee of the acceptance because of multiple students applying the same proj. Mar 26 16:25:46 but if you do so whether will it create negative impact on you? Mar 26 16:26:29 alia: one of the things orgs are looking for is students that are very commited to the project, and if you're applying for a bunch of totally different projects chances are you're not all that committed to any one of them Mar 26 16:27:17 alexanderhiam:that i can understand but ultimately student is doing to do a single project Mar 26 16:27:44 going* Mar 26 16:27:51 it will likely show in your proposal if you're applying just for the sake of getting accepted for a project vs. working on that specific project because it's something you really believe in Mar 26 16:28:39 ok Mar 26 16:29:24 so imo it wouldn't actually increase your chances of getting accepted Mar 26 16:30:56 ok thanks. Mar 26 16:31:56 They have a system of conflict resolution if a student is accepted by multiple orgs a day or two before the results Mar 26 16:33:00 alexanderhiam: I am changing my proposal in the area of generating the pruspeak code entirely on the node.js side.. I saw your message last night.. It was excellent !!!! Mar 26 16:33:51 alexanderhiam Abhishek_ : sorry for the delay(have my tests going on)...i have answered the questions asked on melange(PRU-Bridge) :) Mar 26 16:34:41 jkridner: Would love to hear your feedback on my project(PRU-Bridge) Mar 26 16:34:41 alexanderhiam: so basically the node.js interpreter will invoke the respective function.. and the function will generate a pruspeak code and will append to the main script.. This script finally will be executed by the pruspeak interpreter via TCP.. what do you think ?? Mar 26 16:35:22 kiran4399: I think that sounds like a better approach Mar 26 16:36:35 alexanderhiam: I don't understand why I did not get this idea before... now I am making changes in the proposal during my exams... :-( Mar 26 16:37:04 ok, who needs proposal feedback from me? Mar 26 16:37:19 alexanderhiam: I spoke to jkridner .. he also liked the idea.. so basically, there will not be any RPC server running right ?? Mar 26 16:37:29 ZeekHuge and chandy_ are on my list right now... Mar 26 16:37:42 kiran4399: yeah, I don't see a need for another server Mar 26 16:39:17 alexanderhiam: Tomorrow my Mid-term exams will finish.. Is it possible for you to be little earlier on IRC so that I can finalize everything ?? Mar 26 16:39:55 sure Mar 26 16:39:57 Abhishek_: the data passing use case must go go with 4 bytes ? Some help please :) Mar 26 16:40:30 ok Mar 26 16:40:33 ok Mar 26 16:40:48 shubhangi: in reference to? Mar 26 16:41:31 Abhishek_: you told me 8byte overhead wont be able to handle transfer rates beaglelogic achieves. Mar 26 16:41:41 in my proposal Mar 26 16:41:43 alexanderhiam: waiting with all ears Mar 26 16:43:06 can we access the proposal template? Mar 26 16:43:06 karki : can you please have a look at my proposal if you have time Mar 26 16:43:36 shubhangi: how much data are you able to move currently in a syscall? Mar 26 16:44:55 alexanderhiam : i need your feedback on my proposal :) Mar 26 16:45:09 ok, adding it to my list Mar 26 16:45:14 shubhangi : I had a look, one or 2 minor suggestions. I'll comment later Mar 26 16:45:37 waiting for feedback on my proposal :) Mar 26 16:46:10 karki: okay. thanks Mar 26 16:49:36 Abhishek_: I was only sending a single char. What would should be the page_size of sysfs for sending down data which can cope with most applications. Mar 26 16:49:44 should* Mar 26 16:50:11 See the implementation of a syscall in 3.8.13 Mar 26 16:50:20 okay Mar 26 16:58:29 alexanderhiam:any suggestion on my proposal Mar 26 17:01:42 chandy_: haven't got to it yet, I will start reading proposals as soon as I can Mar 26 17:11:07 alexanderhiam:ok Mar 26 17:11:47 vmayoral: can you pull in anyone from OpenROV? Mar 26 17:11:55 can try Mar 26 17:12:35 alexanderhiam: in any case, we've looking into the code for a while and we are familiar with it Mar 26 17:12:41 but having them onboard will be great Mar 26 17:13:20 yeah, some interaction at least. Seems like getting stuff pushed back to their code base would be a good goal Mar 26 17:13:57 jkridner might be able to ping them as well Mar 26 17:14:34 who can I put in contact? they have a nice google group, AFAIR Mar 26 17:15:26 that's probably a good place to start then Mar 26 17:16:30 vmayoral: they may have some good ideas as well Mar 26 17:16:42 vmayoral: Fire Cape is only now being given a social campaign? Mar 26 17:16:56 jkridner: yes Mar 26 17:17:39 jkridner: didn't wanted to launch it until we were 100% sure of all the capabilities. We've cracked all the issues (it took 6 re-designs) and it's ready for a wider community Mar 26 17:18:00 the price has also been set to what the DIY community requested Mar 26 17:18:07 DIYDrones* Mar 26 17:18:17 jkridner: please go through the proposal on integration of OpenCV with bonescript Mar 26 17:18:21 so hopefully many beagle wills start taking the skies Mar 26 17:20:20 vmayoral: I am backer #1. :-D Mar 26 17:20:32 jkridner: :) Mar 26 17:20:39 hopefully you have a big promotional e-mail list already to send out? Mar 26 17:20:53 the count-downs are usually nice, but I didn't see one here. Mar 26 17:22:34 thanks jkridner! ;) Mar 26 17:22:55 I'll make an official announcement in the mailing lists but any help will be more than welcome :) Mar 26 17:26:29 jkridner: hey, I'd like to take your answer for the 'if implemented what will be the impact?' question Mar 26 17:27:03 ^so that I can put it up on my proposal for the nw.js app Mar 26 17:27:05 regarding openrov, https://gitter.im/OpenROV/discuss seems to be the right place for chat... there is an IRC bridge. Mar 26 17:29:04 jkridner: yeap, already asked if someone is interested there Mar 26 17:29:23 just saw that Mar 26 17:30:58 vmayoral: backer #2 :P can't wait to get my drone on Mar 26 17:31:32 alexanderhiam :just saw it! :). Much appreciated. For everyone in the channel wishing to make BBB-based drones, http://igg.me/at/pxf ;) Mar 26 17:33:37 vmayoral: did you note they applied to be a mentoring organization: https://forum.openrov.com/t/google-summer-of-code-2015-application/2172 Mar 26 17:33:43 * jkridner wonders if they were accepted. Mar 26 17:34:02 "The 2015 application did not make the cut this year" Mar 26 17:34:13 jkridner: didn't notice that. If so they probably announced it somewhere Mar 26 17:34:18 now's their chance! Mar 26 17:34:33 * vmayoral reads OpenROV blog Mar 26 17:35:21 jkridner: found https://forum.openrov.com/t/google-summer-of-code-ideas/2322/6 but it does not confirm acceptance Mar 26 17:35:34 I don't find them on the list. Mar 26 17:43:59 jkridner: Please do have a look at my proposal when you get time. Mar 26 17:50:38 * anujdeshpande thinks students should rename the abysmal name he gave the bbb android idea in their proposals Mar 26 18:07:29 OpenROV did apply as a mentoring organisation for GSoC but unfortunately didn't get selected... Mar 26 18:09:35 well, there's now a great chance to merge BeagleBoard, ROS, OpenROV and APM with BeaglePilot 2 now Mar 26 18:09:44 if i were a student i'd fight my way in there ;) Mar 26 18:10:16 vmayoral: I saw the topic on the Beagle GSoC mailing list. I am interested in taking up the same as my GSoC project. Mar 26 18:10:36 following up your e-mails now Mar 26 18:10:44 let's keep the discussion for that in the mailing list Mar 26 18:11:02 i'd suggest hurrying up putting together a good application Mar 26 18:11:36 Sure. I'm having a look at the Erle Robotics forum and documentation for Snappy. Mar 26 18:12:39 I've completed the cross-compiling task and am in the process of putting together my application. Mar 26 18:16:18 jkridner: small favor please, could you give me your answer to 'if implemented how it will impact the community?' question for the nw.js app Mar 26 18:18:21 jkridner: you could drop it as a comment on my proposal or mail it to me or just enter it here, which ever is convinient for you Mar 26 18:21:20 jkridner: just 2-3 lines would do I guess, sorry for repeatedly asking you, really need this Mar 26 18:35:34 jkridner: ping Mar 26 19:45:18 shubhangi: any progress with the usecase code sketch? Mar 26 20:25:03 Abhishek_: still understanding how machinekit works. they are using libprussdrv to write dir and step values. Should this project also include writing functions available in libprussdrv to use the new driver ? Mar 26 20:26:37 a libprussdrv-like python api for the new driver ? because without that the driver would still remain inaccessible to the endusers at large. Or can that be a part of other gsoc project like pru-bridge Mar 26 20:27:13 shubhangi: a python api would be great Mar 26 20:27:42 it could just be wrappers for the sysfs entries Mar 26 20:28:34 alexanderhiam: hmm... no need for mmaping like prussdrv .. Mar 26 20:28:59 i should probably add that Mar 26 20:29:43 wouldn't mmaping kind of bypass the driver? Mar 26 20:30:12 i.e. how would it know when to send signals to the pru? Mar 26 20:31:49 i meant mmaping is now handled by the driver with a python api writing to sysfs unlike libprussdrv which relies on mmaping. Mar 26 20:32:37 alexanderhiam: sorry, i didn't get your last question ? Mar 26 20:32:51 sorry, no question mark Mar 26 20:34:25 shubhangi: I just mean that if your writing directly into the shared memory from python then you're not going through the driver, so you're not really utilizing the new framework Mar 26 20:39:36 shubhangi: it would ideal if you could include some examples of loading and interacting with firmware using the python api Mar 26 20:40:10 alexanderhiam: okay. will do. thanks :) Mar 26 20:42:24 alexanderhiam: i meant that too. the new framework would no longer require mmaping in userspace unlike libprussdrv.So thats a good thing. Sorry couldntput it forward clearly before Mar 26 20:44:09 oh gotcha, I read 'no need for mmapping' as question :P Mar 26 20:44:18 :) Mar 26 20:44:34 should have put a smiley after it Mar 26 20:48:48 I have a doubt. I still can't figure out where the driver project ends and pru-bridge begins. The proposals might be overlapping. If someone could please help where to draw the line Mar 26 20:51:44 yeah, that line is looking pretty gray right right now... Mar 26 20:53:00 might be worth discussing it with those looking to work on pru-bridge Mar 26 20:53:41 apaar has submitted a proposal for it. will try to discuss with him and karki Mar 26 21:09:10 shubhangi: Consider building a sort of libprussdrv-remoteproc hybrid Mar 26 21:10:15 Ideally pru-bridge is something to be built over the framework which will be built here Mar 26 21:14:56 Abhishek_: Thanks. So it does involve exposing sysfs and writing python wrapper for it. Mar 26 21:15:17 need to edit the proposal Mar 26 21:15:31 The wrapper should be language-agnostic Mar 26 21:15:57 It should be as easy as someone to do it in C as in Python or Node.JS Mar 26 21:16:51 sorry but how does one do that. I need to read up on this. Mar 26 21:17:15 well that would be the sysfs entries Mar 26 21:17:21 yep Mar 26 21:17:39 sysfs is being called the wrapper here ? Mar 26 21:17:47 executing simple read/write wrappers on them Mar 26 21:17:53 okay Mar 26 21:17:57 right, the python api would use the sysfs entries Mar 26 21:18:07 that would be easy. just file io operations for each Mar 26 21:18:09 would be straight-forward to do the same in C and node.js Mar 26 21:18:12 and so would C or Node.JS Mar 26 21:18:18 alexanderhiam was faster Mar 26 21:18:23 ;) Mar 26 21:18:30 :) Mar 26 21:18:40 but time should be spent on demos as well, so maybe just one language Mar 26 21:19:09 I'd say python or node.js Mar 26 21:19:11 * shubhangi wishes deadline was 28th Mar 26 21:19:28 APIs could be contributed to Bonescript and/or PyBBIO Mar 26 21:19:30 It's 28th, 12:30 AM Mar 26 21:20:02 i'll link a gist in the proposal. Mar 26 21:20:04 okay :) Mar 26 21:20:27 BTW, proposals can be made editable by mentors if they need to seek further clarification Mar 26 21:20:42 even after tomorrow's deadline Mar 26 21:21:11 i'll complete the machinekit use case. sorry it might take a day or two. Mar 26 21:21:29 finalizing proposal Mar 26 21:21:44 I would say get the userspace API and demos in there first Mar 26 21:22:03 there's a kernel driver API component as well? Mar 26 21:28:52 shubhangi: How do you handle inter PRU communication use cases? Mar 26 21:30:33 or one PRU affecting the state of the other (halting / resuming / sending an interrupt / ... )? Mar 26 21:34:09 right, taht would be good to include in the PRU API Mar 26 21:34:47 yeah, all the time we are focusing on communication channels, there is need to build a good PRU side API as well Mar 26 21:36:58 BeagleLogic PRU-side firmware has a few functions that could be a part of the API Mar 26 21:37:51 [pru_defs.h] Mar 26 21:40:23 sigh Mar 26 21:40:45 all this "standardizing" PRU stuff is rediculous Mar 26 21:48:54 ds2: this isn't really standardizing anything on the PRU, it's more just getting a nice clean driver base Mar 26 21:49:17 like a nice clean doorstop? Mar 26 21:49:40 the PRU has limited resources and one has to be creative to squeeze performance out of it Mar 26 21:49:57 ``clean'' isn't an option unless the PRU is being squandered Mar 26 21:50:13 not clean on the PRU side, clean on the kernel driver side Mar 26 21:50:36 the PRU code will be some very minimal API that you can choose to use or not Mar 26 21:50:52 that doesn't work either cuz the limitations on the PRU end dictate what you can do for a given choice of things Mar 26 21:52:10 put it another way - what I am interperting what the "API" stuff is doing is forcing all the IP blocks to behave in a specific manner Mar 26 21:52:30 it is like dictating the McBSP/McSPI/McASP/etc have the same API Mar 26 21:53:36 alexanderhiam: +1 for the clean base point Mar 26 21:56:33 ds2: That's probably because of the students' understanding of the PRU usage pattern, hence my giving a usecase example to work on so that there is more clarity on how the framework is supposed to be Mar 26 21:57:20 Abhishek_: and I am against the framework. I think it should be a case by case thing Mar 26 21:58:34 Abhishek_: on a tangent - have you thought about the possibility of BeagleLogic using the GPMC interface? Mar 26 21:58:55 something like Timer triggering the DMA to latch in data from the GPMC Mar 26 21:59:35 ds2: I understand where you're coming from, but I do see space for some libraries that would work for a large chunk of use cases Mar 26 22:00:15 alexanderhiam: have you consider the appropriateness of the PRU for those use cases? Mar 26 22:00:59 yeah, that's the real question to ask Mar 26 22:01:32 half the stuff floating around is essentially "I donno how to do it in the kernel. Let's do baremetal on the PRU." Mar 26 22:01:52 ds2: Haven't considered GPMC so far Mar 26 22:04:08 Since the PRUs did a good job in my case so I didn't look at alternative approaches Mar 26 22:07:43 Abhishek_: Yes. PRU-PRU communication will be a part of the PRU library. alexanderhiam told me to add that in his review. few basic functions like halt_other_pru ( ) which manipulate the INTC registers Mar 26 22:08:29 halt / resume / interrupt Mar 26 22:11:59 on the lines of beaglelogic pru_def. Mar 26 22:12:42 I don't really notice the PRU-PRU communication in Melange though Mar 26 22:12:59 *point raised in the review Mar 26 22:13:36 yes. i am working on a google doc right now. the melange text editor is very cumbersome Mar 26 22:13:48 will put it through Mar 26 22:18:59 Also have a look at the points put forth by ds2 (within an hour ago from now) Mar 26 22:22:34 Abhishek_: Yes. will try to include how the framework doesn't limit what you can do with the PRU.Doesn't abstract too much. Rather gives you elementary tools so that one focuses on application. Mar 26 22:24:49 hmm Mar 26 22:26:24 the framework might not work for BeagleLogic which kind of pushes the boundary. But would have projects like pruspeak and machinekit a little bit easier. Mar 26 22:30:03 * shubhangi dozes off .. lots to do tomorrow Mar 26 23:29:05 nodebotanist:Hello,I saw you review,and I did some modifications on my proposal,please can you add your opinion on what will be the impact of my project to the beagleboard community Mar 26 23:29:35 Yeah just give me a minute or two tanero, thanks for the heads up Mar 26 23:30:39 nodebotanist: Ok,thanks Mar 26 23:35:26 julianduque:Hello,please can you have a look at my proposal,I will love to have your opinion,so I can do last modifications,also can you add your on idea of what impact my project will have on the beagel board community,I am also tanero on melange,thanks Mar 27 00:01:44 tanero I tossed another review in there Mar 27 01:36:47 hi **** ENDING LOGGING AT Fri Mar 27 02:59:59 2015