Replacing X With Wayland On The Raspberry Pi

[schmidtbag]

The only reason i have not bought a rpi is because they don't have the drivers opensource. It really boggles my mind why a teaching/school board like this is bundled with proprietary graphics drivers which essentially is a big black box denying the students access to much functionality of the stack.

There's a lot wrong with that sentence:
1. Open sourcing the drivers is beyond the RPi team's power.
2. RPi is intended for schools with kids in them, or in other words, not colleges/universities. Neither the kids nor the teachers would know where to start.
3. Suppose it was intended for colleges, only PhD students would likely be pushed to study something so complex, and they'd likely be studying something more generic like graphics for MIPS.
4. Does it seriously matter if the drivers are open source? Yes, they may be missing some features, but who's going to add them? Besides, most features features the Pi is missing are things that it wasn't intended for. It wouldn't surprise me if you're one of the people who bought it for XBMC reasons.

[scionicspectre]

Well, why not? People are using the Pi for all kinds of wacky stuff. I'd hardly say XBMC and performant open graphics drivers would be an outlandish idea. Aside from that, I'd say I agree that it's not really up to the Pi guys to have an open Broadcom graphics driver or not, and these kinds of projects shouldn't be the expectation. However, it seems the Pi guys are promoting this collaboration. It is interesting to see just how flexible Wayland can be, in different scenarios. It looks like it could be running on all kinds of graphics chips soon enough.

[schmidtbag]

Well, why not? People are using the Pi for all kinds of wacky stuff. I'd hardly say XBMC and performant open graphics drivers would be an outlandish idea. Aside from that, I'd say I agree that it's not really up to the Pi guys to have an open Broadcom graphics driver or not, and these kinds of projects shouldn't be the expectation. However, it seems the Pi guys are promoting this collaboration. It is interesting to see just how flexible Wayland can be, in different scenarios. It looks like it could be running on all kinds of graphics chips soon enough.

I'm not against using the Pi for reasons beyond its purpose. What I don't like is people who buy it and complain about stuff it doesn't do, can't do, and was never intended to do. That's like buying a motorcycle and complaining about the fact that it doesn't protect you from rain or doesn't stand up on it's own when you're at a red light.

[scionicspectre]

Yeah, I agree. That might be why I haven't bought it, since I don't want to complain. XD I do, of course, think it would be nice to have a 25 dollar computer that can do useful everyday things- it would be quite fun, especially for giving the kids in your family something to play with. A lot of the children in my family are kind of afraid of getting too excited about tablets and laptops since the adults are always freaking out that they're gonna' break them. I'm sure it will come along, even if it isn't the Pi.

Just to make it clear, I agree that it's weird to have those expectations when the device isn't made for anything close to that usecase. It's strange the kinds of things involved Linux users complain about.

[dh04000]

Well, why not? People are using the Pi for all kinds of wacky stuff. I'd hardly say XBMC and performant open graphics drivers would be an outlandish idea. Aside from that, I'd say I agree that it's not really up to the Pi guys to have an open Broadcom graphics driver or not, and these kinds of projects shouldn't be the expectation. However, it seems the Pi guys are promoting this collaboration. It is interesting to see just how flexible Wayland can be, in different scenarios. It looks like it could be running on all kinds of graphics chips soon enough.

Ok, why don't YOU spend the next months to years reverse engineering the closed drivers and provide US with a feature complete and competiting open driver. Here's zero dollars. I look forwards to your repo'd code of which we'll complain and nick-pick about.

[scionicspectre]

Ok, why don't YOU spend the next months to years reverse engineering the closed drivers and provide US with a feature complete and competiting open driver. Here's zero dollars. I look forwards to your repo'd code of which we'll complain and nick-pick about.

I don't see why I would when I don't have a great deal of expertise in the area. I'm just saying that if someone did this, it wouldn't be that weird since people have done far crazier things with the Pi than using it as a computer with performant graphics. I'm not saying it should happen, necessarily.

[GreatEmerald]

4. Does it seriously matter if the drivers are open source? Yes, they may be missing some features, but who's going to add them? Besides, most features features the Pi is missing are things that it wasn't intended for. It wouldn't surprise me if you're one of the people who bought it for XBMC reasons.

It obviously does. Otherwise you wouldn't have people trying to reverse-engineer the thing.

[schmidtbag]

It obviously does. Otherwise you wouldn't have people trying to reverse-engineer the thing.

The question was meant for the average user. People who actually know a thing or 2 about developing video drivers would obviously care and have a reason to create open source drivers. But the average user would have no interest or ability to mess with an open source driver, so it boggles my mind why the average user would care what kind of driver they're using as long as it is free and functional.

[smitty3268]

Note that recently Weston's implementation has been totally replaced on the Raspberry Pi with a *specific backend* with much improved performance, so Wayland allow great local performance (at a trade-off of potentially worse network usage in remote desktop) but you have to work for it!

A "specific backend" which would be unfeasible to implement for X. Part of the whole reasoning behind wayland is to get rid of all the complicated cruft that has built up over the years, and allow these kinds of things to happen to take advantage of modern hardware better.

[Ericg]

A "specific backend" which would be unfeasible to implement for X. Part of the whole reasoning behind wayland is to get rid of all the complicated cruft that has built up over the years, and allow these kinds of things to happen to take advantage of modern hardware better.

What I find most interesting about this.. is the ABILITY to have hardware specific backends. X is a one size fits all approach, and if it doesnt fit you...youre screwed. With the ability to switch out the backend so easily, it really helps to future-proof Wayland because if we realize we screwed up NOW we can just change the backend in the future.