ArduPilot Partner CUAV new hardware

Blog
102

I think you’ve misinterpreted me. I support hardware diversity and derivative designs, I just think it’s important to respect open source licences.

From the CUAV GitHub README.md:

commit c2b5f417750d242d566f14683d3206ac5f6d362b

The PixHack V5 design document cites many design files or reference files and related libraries from these organizations or companies. Including PX4\PIXHAWK\3DR\Proficnc\ST\Altium\TDK\Hirose\Hrs\PANASONIC\molex\TI\ and other organizations or companies.Some design reference copyrights should belong to them.

commit 7a800145a710a3e859218ce74741d4e4a2344335

The PixHack V5 design document cites many design files or reference files and related libraries from these organizations or companies.
Including PX4\PIXHAWK\3DR\Proficnc\ST\Altium\TDK\Hirose\Hrs\PANASONIC\molex\TI\ and other organizations or companies.Some design reference copyrights should belong to them.
Thanks for their contributions to open source hardware. If there is any infringement, please contact us for deletion and modification.

commit 524a2d3aeb6da392418ad3aa368310ae5fb54182

[See edit below, the following statement is incorrect]

Acknowledgement of using reference designs was deleted, and replaced with the following:

note: It may look like a SOLO or PIX2 Cube, because it all looks like a small square��but it is completely different, including the shape, screw hole installation position, interface definition,size,internal structure, are completely different, so V5 CORE and Cube are completely incompatible, if you need to V5_Core For secondary development, please refer to our open reference design.

Edit: My mistake. I see the reference is still included at the bottom in the latest version:

Design files or libraries may reference some references from PX4\PIXHAWK\3DR\Proficnc, thanks for their contributions to open source hardware. If there is any infringement, please contact us for deletion and modification.

I’d suggest this falls short of best practices, and attribution should really be on their product page as well, but I accept I was mistaken when I said there was no attribution given.

1 Like
103

They link directly to their GitHub repo from their product page

Screenshot%20from%202018-06-19%2022-11-31

http://www.cuav.net/?dt_portfolio=pixhack-v5-autopilot&lang=en

These guys have been around for awhile and when all this controversy started, I got my dev board, looked it over, learned about it, looked at the docs, cloned their repo and studied that, and it became apparent they are making every effort to do it right.

I really don’t know what somebody would expect them to do further. I haven’t seen anybody over at PX4 quiverin’ in their boots over it.

104

I’m not sure I understand your motivations. Hopefully I’ve made mine clear, but please let me know if they come off as confusing or contrived. I realise this discussion is off-topic, but I’m hoping it might lead to a better understanding of the Creative Commons licence for all involved (myself included), and resolve some of the misunderstanding regarding CUAV so that we can all move on from the inappropriate and unhelpful ‘Clone’ accusations.

The link to GitHub you mentioned is on every page, and goes to CUAV’s account page, not to the README.md for the v5 that provides attribution. Hopefully you’ll agree that the average consumer is very unlikely to see this, and that it would benefit the open source community to make this attribution more prominent.

To quote a public Facebook discussion I had with Philip that gives some insight into why attribution is important for creators:

yes, though i have days where I want to quit open source… I love the community… and plan on continuing to do open hardware. But, I do appreciate it when the community jumps in and holds people accountable for violations, as I sometimes feel like i just piss people off when i bring it up :slight_smile:

I do think ProfiCNC/Hex has a great standing in the community, but unfortunately, volume is king when it comes to being able to deal with companies like STM etc… if we cant get our numbers up, we cant get cheap parts, and it becomes harder to compete. It is life, so that’s OK… but at least having others play by the same rules is helpful :slight_smile:

Getting attribution by companies like these however, does help the whole project… in a few ways…

  1. makes me feel warm and fuzzy…
  2. it gives us credibility with the whole supply chain.
  3. it encourages others to play by the rules.

and by credibility with the supply chain, what I mean is… I then talk to STM for example, and they can then see my influence on a bigger share of the market, giving them reason to talk to me, and work with me on cool future chips etc

The attribution provided is missing key requirements from the licence, specifically the title of the original work, the name of the original author/organisation, and the copyright notice.

If a representative from CUAV is reading this, properly conforming with the attribution licence requirements is simple. All you have to do is modify the README.md to include the following, and clearly link this file from the product page.

The PixHack V5 is a derivative of The Cube Autopilot by ProfiCNC which is licenced under CC-BY-SA 3.0.

Note that this does not apply to the IMUv3 or PSM boards, which are licensed under CC-BY-NC 3.0, and require you to get permission from the copyright holder (Phiip Rowse) for commercial use.

It might seem pedantic, but proper attribution in this manner includes the following features:

  • It links to the original work so that others can compare designs, and/or create their own derivatives
  • It links to the original author, so that they may benefit from making their designs open source
  • It links to the specific details of the licence, so that future derivatives can ensure they conform

As it stands, it’s unclear what CUAV have referenced, and what licence the referenced designs were released under.

1 Like
105

There ya go.


Now we’ll see if they’ll push it.

1 Like
106

My motivations are that when a Partner introduces a new product as a developer version and asks us to support it, which we gladly want to do, it is not necessary to immediately begin bashing the product and company because somebody doesn’t agree with it. ArduPilot is not in the hardware business. This makes the community look just a little unprofessional to any casual observer. And drags us back into quagmire of the days of when 3DR attempted to lock out what was perceived as “clones” of an open hardware design with code.

How about instead, if this product has licensing or attribution deficiencies, the affected party(s) contact them directly to get it resolved instead of jumping on the soapbox? And then if the deficiencies aren’t resolved, and that is brought to the attention of the dev team, we can make a decision as to whether we should support this product or maybe shouldn’t because it may not be legal or meet some requirements.

Regardless, ArduPilot is going to support the FMUv5 as it is the next gen cpu platform, etc… But it is not necessary to put one of our partners on trial and declare ourselves judge, jury and executioner when we are not in the hardware business.

1 Like
107

Perhaps everything beyond the first post should be moved to the #general:rants-raves section…I know I contributed a bit to the discussion, but the #blog seems like a bad place for it. Is that possible?

1 Like
108

This is an open source project that has excelled where many others have failed by virtue of being open source, so wouldn’t it look ‘unprofessional’ for us not to care about open source license violations?

During the 3DR days, the majority of the community were opposed to efforts to control the use of ‘clone’ boards, myself included, and for similar reasons. Open source licenses are important.

Given that the creator of the original works has asked for community assistance in holding the license violators accountable, don’t you think the opposing viewpoint is worth a bit of critical analysis? It’s not like ProfiCNC is a patent troll, intent on suppressing innovation from others while failing to contribute anything back to the community. They were one of the first ArduPilot partners, and in addition to providing open source hardware, together with their manufacturing partner Hex they support the community by hosting developer conferences. There’s one going on this weekend in Shenzhen.

Objectively, the license deficiencies haven’t been resolved yet. There’s an attempt at attribution (which I was unaware of until you pointed it out), but it fails to comply with the basic license requirements, which can easily be verified by simply reading the license. The NC issue is more complicated. There’s no concrete evidence that CUAV is violating the IMUv3 or PSM license that I’ve seen, but there’s details like a 1:1 copy of the IMU heater, and resister values for various subsystems that are highly suspicious.

I’m more than willing to support CUAV, but I think they should take responsibility for learning about the licenses the design files used to create their product were shared under, and ensuring they comply with the requirements. It’s really that simple.

109

Please bring it to PR discussion on GitHub

1 Like
114

I believe that the arguments presented represent several points of view.

If someone wants to discuss the goodness of the licensing issues there is a PR created for that.

Some links to the Hardware can be found here https://docs.px4.io/en/flight_controller/pixhawk_series.html

and on the ArduPilot docs here: http://ardupilot.org/rover/docs/common-autopilots.html#common-autopilots

2 Likes