i do think it would be nice if the website at large had, like, screenshots of the desktop environment in question. it’s not very useful to document a graphical program or environment without images of said graphics
This blog post is concerned with accessibility, I think the idea was for a stripped down post that somewhat reflects the experience a blind person would have.
Other posts of the project feature a lot of videos exposing behavior and graphical features. Also, from my understanding, Arcan isnt in a “marketing” stage, so the website inst focused on getting people to adopt it right now, instead it showcases implementation progress.
There be dragons, but the lash interface has gotten very interesting to use in the last year.
yeah, fair enough. the stripped-down post makes sense, i’m more commenting on the website in general (my first response when pointed to a new project is to poke around the homepage)
but yeah that’s fair to not be in a marketing stage at this time. such a stage dramatically shapes a lot of things and it makes sense to be focusing on other angles at the moment.
I have never been able to make heads or tails out of what it is, but other lobsters seem to be really excited about it, so hopefully eventually they will clarify it for us.
It’s a display server. It’s in the same space as X11 or Wayland, but designed correctly. It also has a bunch of features on to such as a text layer that’s much richer than a TTY and a distributed IPC mechanism. You can build a bunch of great things on top, such as remote display but where video streams are decoded at the endpoint, or where the display server goes away and clients automatically reconnect and reattach (e.g. reboot your laptop while doing remote display from some server app).
Basically, Arcan is what Wayland should have been. It also has an X11 compatibility layer and a (now less maintained) Wayland compatibility layer. These inherit the features of Arcan, so can also do all of the transparent reconnection things, for example.
i would say rather: it subsumes the conceptual models under which those things that are called display servers exist. as a result, it happens to include components with somewhat similar structure (and indeed the details there are right, near as i can tell) and even compatibility, but stopping there is missing the point. (what that is, and what its own conceptual models are, i am not sure there is a concise explanation for past ‘listen real hard to everything that bjorn says’, which is why i did not respond to the parent. same as i don’t have a concise explanation for how to learn compilers better than ‘listen real hard to everything sofia says’, or for certain other aspects of computers better than ‘listen real hard to everything fabian giesen says’)
Basically, Arcan is what Wayland should have been.
What does it improve on from Wayland?
I’ve been pretty happy with Wayland personally and am not aware of any major mistakes. Maybe some kind of reflection to make waypipe easier might have been nice, I guess, but that’s all I could think of.
“I’ve been pretty happy with Wayland personally and am not aware of any major mistakes.”
“I am happy with HTTP, why should I use Chrome’. That’s about the extent of what you stated. There’s not exactly a way to provide a response or breakdown to a sentence like that.
I think waypipe and Wayland’s security model are pretty good, so I don’t mind that too much
Do reference anything actual, factual, formal which lead you to that conclusion - considering the Wayland security model is, charitably put, undefined. The documentation didn’t even bother to finish the sentence on >authentication< more than ‘maybe we should have it’. The last traces of anyone with any kind of tangentially relevant experience or background approaching it was Martin Peres and Steve Dodier-Lazaro about 10 years ago. It fell on dead ears and lead nowhere.
Waypipe, on top of that, doesn’t have a security model (since it’s a GSOC monkey patching wayland objects as they broke) and default to ‘use SSH or something because we don’t do anything’. That’s also not understanding what SSH do and do not. If you’ve actually and seriously used Waypipe over SSH on a network outside of your control and typed in anything secret, assume that to be recorded and recoverable as plaintext seconds after you did so.
“I am happy with HTTP, why should I use Chrome’. That’s about the extent of what you stated. There’s not exactly a way to provide a response or breakdown to a sentence like that.
I’m not sure what you mean by this. I was asking what new things are possible with Arcan that aren’t possible with a Wayland compositor.
Do reference anything actual, factual, formal which lead you to that conclusion - considering the Wayland security model is, charitably put, undefined.
I don’t really want to go finding references, sorry. My memory is that Wayland only gives an application access to its own data (surfaces and input events) unless you request them through a separate protocol. Leaving the authentication method undefined for the compositor to choose seems sensible, no? Or do you mean it’s impossible to do authentication?
If you’ve actually and seriously used Waypipe over SSH on a network outside of your control and typed in anything secret, assume that to be recorded and recoverable as plaintext seconds after you did so.
Oh, what’s wrong with SSH? I thought it was secure over the network?
i do think it would be nice if the website at large had, like, screenshots of the desktop environment in question. it’s not very useful to document a graphical program or environment without images of said graphics
This blog post is concerned with accessibility, I think the idea was for a stripped down post that somewhat reflects the experience a blind person would have.
Other posts of the project feature a lot of videos exposing behavior and graphical features. Also, from my understanding, Arcan isnt in a “marketing” stage, so the website inst focused on getting people to adopt it right now, instead it showcases implementation progress.
There be dragons, but the lash interface has gotten very interesting to use in the last year.
yeah, fair enough. the stripped-down post makes sense, i’m more commenting on the website in general (my first response when pointed to a new project is to poke around the homepage)
but yeah that’s fair to not be in a marketing stage at this time. such a stage dramatically shapes a lot of things and it makes sense to be focusing on other angles at the moment.
I have never been able to make heads or tails out of what it is, but other lobsters seem to be really excited about it, so hopefully eventually they will clarify it for us.
It’s a display server. It’s in the same space as X11 or Wayland, but designed correctly. It also has a bunch of features on to such as a text layer that’s much richer than a TTY and a distributed IPC mechanism. You can build a bunch of great things on top, such as remote display but where video streams are decoded at the endpoint, or where the display server goes away and clients automatically reconnect and reattach (e.g. reboot your laptop while doing remote display from some server app).
Basically, Arcan is what Wayland should have been. It also has an X11 compatibility layer and a (now less maintained) Wayland compatibility layer. These inherit the features of Arcan, so can also do all of the transparent reconnection things, for example.
i would say rather: it subsumes the conceptual models under which those things that are called display servers exist. as a result, it happens to include components with somewhat similar structure (and indeed the details there are right, near as i can tell) and even compatibility, but stopping there is missing the point. (what that is, and what its own conceptual models are, i am not sure there is a concise explanation for past ‘listen real hard to everything that bjorn says’, which is why i did not respond to the parent. same as i don’t have a concise explanation for how to learn compilers better than ‘listen real hard to everything sofia says’, or for certain other aspects of computers better than ‘listen real hard to everything fabian giesen says’)
What does it improve on from Wayland?
I’ve been pretty happy with Wayland personally and am not aware of any major mistakes. Maybe some kind of reflection to make
waypipeeasier might have been nice, I guess, but that’s all I could think of.Actual network transparency
A security model that humans can use.
And that’s just as a casual user.
I think
waypipeand Wayland’s security model are pretty good, so I don’t mind that too muchThe main issue I’ve had with Wayland is slow adoption (still waiting on Krita) which unfortunately Arcan would have to deal with too
(note: edited since my tone was unnecessarily critical of Arcan)
“I’ve been pretty happy with Wayland personally and am not aware of any major mistakes.”
“I am happy with HTTP, why should I use Chrome’. That’s about the extent of what you stated. There’s not exactly a way to provide a response or breakdown to a sentence like that.
Do reference anything actual, factual, formal which lead you to that conclusion - considering the Wayland security model is, charitably put, undefined. The documentation didn’t even bother to finish the sentence on >authentication< more than ‘maybe we should have it’. The last traces of anyone with any kind of tangentially relevant experience or background approaching it was Martin Peres and Steve Dodier-Lazaro about 10 years ago. It fell on dead ears and lead nowhere.
Waypipe, on top of that, doesn’t have a security model (since it’s a GSOC monkey patching wayland objects as they broke) and default to ‘use SSH or something because we don’t do anything’. That’s also not understanding what SSH do and do not. If you’ve actually and seriously used Waypipe over SSH on a network outside of your control and typed in anything secret, assume that to be recorded and recoverable as plaintext seconds after you did so.
I’m not sure what you mean by this. I was asking what new things are possible with Arcan that aren’t possible with a Wayland compositor.
I don’t really want to go finding references, sorry. My memory is that Wayland only gives an application access to its own data (surfaces and input events) unless you request them through a separate protocol. Leaving the authentication method undefined for the compositor to choose seems sensible, no? Or do you mean it’s impossible to do authentication?
Oh, what’s wrong with SSH? I thought it was secure over the network?
I’m a casual user, not a developer of arcan.
Arcan seems to be built to allow applications to move between machines easily, while running.
That’s a really cool feature, but at least personally I can live without it. I’m fine with saving and relaunching the GUI somewhere else.
I don’t want to dismiss Arcan too much, it seems really impressive, I’m just not sure Wayland needs replacing.
P.S. I didn’t mean to imply you were representing Arcan development, sorry
[Comment removed by author]