renky Posted December 5, 2020 Posted December 5, 2020 https://github.com/open-sdr/openwifi https://www.youtube.com/watch?v=NW6-6N_EtrY The goal is to release open source technical blue prints of an 802.11 wifi card. Ready for manufacturing. Hiring people for the task would be about 400000usd. At the present time such funds are not at hand. Openwifi got a nlnet.nl grant. Next move is to build a fpga wifi card device. Please comment if you think there will be a market for such a device? Which price should it have? Should a crowd funding get started, do you know companies willing to join a double up crowd funding? Thank you.
TRS-80 Posted December 5, 2020 Posted December 5, 2020 Awesome project, I wish them all the best! Although I think many of us (including OP) realize the extreme challenges that are involved. The point (in my view) is not to take that as a sign not to try (as average person probably thinks) and to give up but rather to try anyway, in spite of those odds. At first I wondered if they were related in any way to Libre-SOC but apparently not. This is certainly a nascent area, but there are several different projects going in this direction and all need to be applauded (and even IMO helped, by whatever means we are able). Libre-SOC are also getting a lot of those nlnet.nl and EU grants. So maybe not just a pipe dream after all. Cheers to our friends across the pond! We are kindred spirits, although mostly covered up with our own work here, already.
jxj Posted December 7, 2020 Posted December 7, 2020 Thank you renky and thank you TRS-80 for the very good advice. This is Xianjun Jiao which is the primary author of the openwifi project. I am being encouraged and inspired by people like you guys. And I do agree that no matter how big the difficulty/risk is we should try it out! Now I am pushing the USB based WiFi dongle/chip design (in this way, people use openwifi just like any other COTS WiFi dongle!) in my spare time, and already got some progress. I am drawing the schematic and discussing with some hardware experts on the design, and good news is that seems like there is a hope that we can make an openwifi FPGA SDR dongle with price < 100 dollar, which will be much much cheaper than the COTS FPGA development boards currently we support. I believe the cost and "easy-to-use" are two key factors for an open source hardware/chip project to setup a large community (like Raspberry Pi, etc) -- this was also the main feedback I got in the last FOSDEM2020 where I announced the openwifi project for the first time. Why I see the 100 dollar FPGA dongle as the first step? Because I want the openwifi chip project to be sustainable by itself, not taking external funding as a main supportive source. Yes we do need funding for some phases, like the board manufacturing and testing phase. But after that, I hope it can go to crowdfunding and sale to hackers/hobbyists/community, and finally make the project survive by itself. With this FPGA dongle, I believe we can already boost the community a lot (much much cheaper than current COTS FPGA boards), and make the chip design more mature, which will be very good preparation for real chip tape out. You might notice that I mentioned "in my spare time". This actually reveals the difficulty currently I have. As you might know I work for IDLab which is a joint lab between Gent university (UGent) and imec, Belgium. Although I proposed and started the openwifi project internally at the beginning, the reason why openwifi project can survive for years in imec/UGent is that it is useful for some internal projects. Then I successfully convinced my boss and BD (business developer) to opensource the design from the dissemination point of view, that is why you can see it now. So we should realize that openwifi developers (me and my colleagues) have to work for other internal projects. We can't say we work for openwifi project internally, because there isn't such an openwifi project officially in imec/UGent. Only because openwifi is useful for other internal projects, our job positions are kept and we get salary to raise our families. I tried several times to make the openwifi an official internal project, but so far I am not successful. The imec/UGent business development is quite traditional/old-fashion. They mainly care about using the openwifi for other internal projects and then sell the non-open source license/IP/patent of those internal projects, which is understandable because this mode works successfully for the big company/organization. So actually they don't care much about the hardware price, the low cost chip/dongle design, and setup a big open source community. That is why I have to try to push the chip/dongle project in my spare time (and also with my private computer, not the company computer!) I still need to find some time to discuss with my boss to get permission that I am allowed to push the real open source chip direction in my spare time. In this phase, I believe I can handle most of the technical challenges. But from a project management/advancing point of view, maybe I need very skillful management/operation people's help! Any further advice or spreading/discussion will be appreciated! Above are some progressions and thoughts. A bit long text and mess situation. Hope I can make it more and more clear in the near future. Keep an eye on the upcoming libreplanet2021, rC3 and fosdem2021, I might have presentations there. Thanks again for your caring about the openwifi project! 1
TRS-80 Posted December 7, 2020 Posted December 7, 2020 58 minutes ago, jxj said: A bit long text and mess situation. Hope I can make it more and more clear in the near future. It's perfectly clear to me, but then again I understand the challenges you are balancing[0]. Good for you arguing in favor of open source. Our man on the inside, so to speak. I guess, to them, you can pitch the "business case." Shared maintenance costs, etc. Which was the original reason for the creation of the term "open source" in the first place. A lot of F/LOSS actually gets developed in reality by people in positions like yours; I was just emailing privately with one of Pandoc developers who is employed at some institution and works on Pandoc nearly full time because that institution use it for something. So there are a lot of ways to skin a cat, and we all need to earn a living in the meantime. In my view, the fact you have a full time job in the meantime makes this a bit more sustainable, so thanks for sharing that tid-bit. Hopefully you get to work on it in your free time. Which bring to my mind some licensing concerns I hope you addressed already, but I guess I will PM you about that. Thanks for coming here and introducing yourself, welcome to the forums, and I wish you the best of luck with this project! [0] I also just read the project more thoroughly again, and just noticed the dual license. Well, at least it's AGPL on the one side. A lot better than nothing, I guess. Thanks for coming here and explaining the situation, as otherwise dual license might be a bit of a turn off to staunch Free Software advocate like myself. Also, at least you chose AGPL instead of some "opensource" meme license.
renky Posted December 8, 2020 Author Posted December 8, 2020 Quote challenges It is. Like on other fields, it appears to be a matter of getting it organized. Free software people do not have a trusted entity to evaluate projects. A single placeholder which can evaluate a project technically, promote the project and run a crowd funding. I do not see websites like crowd supply being such an entity. 400000usd sounds like a lot of money. But it is not. It is 40000 people paying 10usd. If you knew that most likely the public would get a blue print of an open source 802.11 wifi card, wouldn't you pay the 10usd? Pay 10usd for a notebook that adheres to the free software foundation's respect your freedom requirements? https://www.powerpc-notebook.org/en/ Quote several different projects going in this direction Can you list them? Quote turn off to staunch Free Software advocate Are you referring to the mit license matter? If so, I do not agree with you. It is like some free software people have not realized, that a license similar to the mit license grants more freedom, not less. If a person wants to use a gpl license, I do not mind it. But a mit license grants more options. And the part which is free stays free. A mit license does not stop any person from adding to the work and making the added part free. But about a mit license people can protect what they have added and that may be preferable for the public. Let us say a manufacturer builds a mainboard originating from open source licensed information. Then adds some non open source parts. Maybe the mainboard can be useful to free software people, because the mainboard still is workable without the non free parts functioning. That is the approach of riscv. riscv.org And the business model of sifive.com The main part stays free. My understanding is that openwifi has made the same decision.
Recommended Posts