Canonical's Mir 2018 Plans Include Some Potentially Interesting IoT Features

[Palu Macil]

If I recall correctly, a core reason that Canonical developed mir was surfaceflinger compatibility which would give mir more touch or Android capabilities perhaps than Westen

[bregma]

QNX is used by some big companies like Ford and even some Mercedes models for infotainment displays in cars. Apple Carplay is used in so many new cars as well and Carplay runs primarily on QNX. So don't go telling me it's not popular.

I'm not telling you Neutrino is not popular. I know it's embedded in high-end models from many automakers (in fact I could tell you exactly which ones). I could also tell you exactly which models use Linux instead, and I can tell you they're the ones in my price range (because I do not finance automobile purchases, that's a mug's game). Unfortunately I have signed in blood papers that state my agreement not to reveal specifics.

The point is that the people who make cars want to buy systems that make their lawyers and accountants happy. Using Neutrino does that for them, sure, as does using Linux but using Wayland compositors is unlikely to do that because (a) there is no one standing behind it to take to court when things go wrong and (2) the functional Wayland compositors (KWin and Mutter) use licensing that sends shivers down the icy spine of corporate lawyers and makes their sphincters tighten another notch. Canonical's Mir offers both a corporate entity to sue and an offer of non-GPL license terms (thanks to their CLA), and that makes the lawyers pull their lips back around their teeth forming a rictus that, if they were fully human, could be confused with a smile.

See, the use of Linux in IVI applications is a growth market, and Mir is well placed to leverage that to make a return on investment for Canonical. It's not the only possible exploitable opportunity Alan may have been referring to (and indeed I have a little bit of inside information on that, too) but it's a very real, and very sensible, possibility.

[TheBlackCat]

If I recall correctly, a core reason that Canonical developed mir was surfaceflinger compatibility which would give mir more touch or Android capabilities perhaps than Westen

This gets brought up every Mir thread. Any claimed advantages of Mir in regards to Android or phone compatibility in general turned out to either be due to lack of technical knowledge on the part of Mir developers, or was actually technology developed for Wayland that Canonical modified for Mir. In fact all the supposed technical advantages of Mir turned out to be incorrect and had to be retracted by Mir developers.

[Vistaus]

I'm not telling you Neutrino is not popular. I know it's embedded in high-end models from many automakers (in fact I could tell you exactly which ones). I could also tell you exactly which models use Linux instead, and I can tell you they're the ones in my price range (because I do not finance automobile purchases, that's a mug's game). Unfortunately I have signed in blood papers that state my agreement not to reveal specifics.

The point is that the people who make cars want to buy systems that make their lawyers and accountants happy. Using Neutrino does that for them, sure, as does using Linux but using Wayland compositors is unlikely to do that because (a) there is no one standing behind it to take to court when things go wrong and (2) the functional Wayland compositors (KWin and Mutter) use licensing that sends shivers down the icy spine of corporate lawyers and makes their sphincters tighten another notch. Canonical's Mir offers both a corporate entity to sue and an offer of non-GPL license terms (thanks to their CLA), and that makes the lawyers pull their lips back around their teeth forming a rictus that, if they were fully human, could be confused with a smile.

See, the use of Linux in IVI applications is a growth market, and Mir is well placed to leverage that to make a return on investment for Canonical. It's not the only possible exploitable opportunity Alan may have been referring to (and indeed I have a little bit of inside information on that, too) but it's a very real, and very sensible, possibility.

Maybe I understood wrong, but wasn't Intel the one who initiated Wayland development? If so, wouldn't they step up if a company was taken to court regarding Wayland?

[bregma]

Maybe I understood wrong, but wasn't Intel the one who initiated Wayland development? If so, wouldn't they step up if a company was taken to court regarding Wayland?

Wayland isn't a display server, it's a protocol. You can't sue the folks who develop a protocol, only the folks who write the software that actually implements the protocol. That would be the display server. What's the name of the commercial display server that Intel wrote and indemnifies?

[Vistaus]

Wayland isn't a display server, it's a protocol. You can't sue the folks who develop a protocol, only the folks who write the software that actually implements the protocol. That would be the display server. What's the name of the commercial display server that Intel wrote and indemnifies?

I know very well that Wayland is only a protocol. I just mean that Intel was the one who initiated development of Wayland and Weston. So if someone uses Wayland and Weston, then Intel can step up. Or are you going to tell me that Weston is not a product that can be fought over either?

[bregma]

I know very well that Wayland is only a protocol. I just mean that Intel was the one who initiated development of Wayland and Weston. So if someone uses Wayland and Weston, then Intel can step up. Or are you going to tell me that Weston is not a product that can be fought over either?

Yep. Intel has copyright in the source of both the Wayland protocol and associated tools and the Weston reference implementation of a Wayland-based graphics compositor, but they were developed and are distributed under the "open source" license known as the MIT license, which clearly and unambiguously states "IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE." [sorry for the all-caps, but that's a direct quote and I guess they expect that if they shout it gives greater force to their words or something; lawyers are an odd lot]. Wayland is not a product, it's a project. The terms of using code from that project are effectively "you can use this source for anything you want but you assume all the risk."

The neo-postmodern skyscrapers full of lawyers who evaluate software licenses at companies that make safety-critical embedded devices take one look at those terms and toss the printout in a bin (trash can in the USA). Intel is not foolish: they are not going to "step up" and volunteer to be sued in some sweaty East Texas courtroom over an idea (protocol) or demo software (Weston) that has no chance of bringing them returns to their quarterly bottom line. They would only indemnify something that is likely to generate revenue. In fact, that's pretty much the difference between a product and a project: someone offering indemnification and an expectation of revenue.

So, summarizing my original point: Mir is a commercial product from Canonical, with the possibility of indemnification and an expectation of revenue. That's why it exists, and why Wayland (a protocol, or idea, which is used not only by Mir but by a number of other notable display servers) or Weston (an "open source" project with no offer of indemnification or expectation of revenue) are not substitutable.

[Vistaus]

Yep. Intel has copyright in the source of both the Wayland protocol and associated tools and the Weston reference implementation of a Wayland-based graphics compositor, but they were developed and are distributed under the "open source" license known as the MIT license, which clearly and unambiguously states "IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE." [sorry for the all-caps, but that's a direct quote and I guess they expect that if they shout it gives greater force to their words or something; lawyers are an odd lot]. Wayland is not a product, it's a project. The terms of using code from that project are effectively "you can use this source for anything you want but you assume all the risk."

The neo-postmodern skyscrapers full of lawyers who evaluate software licenses at companies that make safety-critical embedded devices take one look at those terms and toss the printout in a bin (trash can in the USA). Intel is not foolish: they are not going to "step up" and volunteer to be sued in some sweaty East Texas courtroom over an idea (protocol) or demo software (Weston) that has no chance of bringing them returns to their quarterly bottom line. They would only indemnify something that is likely to generate revenue. In fact, that's pretty much the difference between a product and a project: someone offering indemnification and an expectation of revenue.

So, summarizing my original point: Mir is a commercial product from Canonical, with the possibility of indemnification and an expectation of revenue. That's why it exists, and why Wayland (a protocol, or idea, which is used not only by Mir but by a number of other notable display servers) or Weston (an "open source" project with no offer of indemnification or expectation of revenue) are not substitutable.

You're right. Thanks for the explanation!

Also, if they don't want Mir, then can also opt to choose for Arcan, which is even more optimized due to the amazing h/w acceleration.