[CentOS] Monitor Dummy Device

Thu Jan 21 00:52:31 UTC 2016
Warren Young <wyml at etr-usa.com>

On Jan 20, 2016, at 5:08 PM, Jonathan Billings <billings at negate.org> wrote:
> On Jan 20, 2016, at 5:00 PM, Joey <forum at r5d.de> wrote:
>> A hoped its possible to configure a second xorg-server or a
>> fakedevice/-monitor that kde recognize this and you could use this as
>> second desktop in same way as in dualscreen-modus.
> This is why I suggested Xdmx.

The X protocol is a far better choice for this sort of application than video codecs, especially over a wireless LAN.  The local X server inherently knows when changes occur that it needs to send to the remote X server, whereas with video, each frame of video has to be treated like a new world, with a brute-force search over the frame for parts which have remained unchanged, so they can be optimized out.

And even then, a video codec is going to wastefully send frames even when it doesn’t have to, because it’s designed with the assumption that *something* changes in every frame, even if it’s just a text crawl and the talking head’s lips on CNN.

That said, X is fairly laggy, increasingly so as GUIs have become piggier.  (More pixels, more colors per pixel, more use of gradients and such rather than solid fills, antialiasing, etc.)

Then on top of that, by using wifi, you’re going back in time something like 10 years in terms of bandwidth and latency compared to the wired LANs where X grew up.  I just did a test here from my desktop: 36 Mbit/sec and 32 ms average ping time over wifi to a server that gives me 940 Mbit/sec and 0.35 ms ping time over the GigE LAN.

The wifi router is about 20 feet away from me now, through a couple of uninsulated 2x4 and gypsum board walls; the WLAN is quiet at the moment, too.  You speak of education, which suggests much longer distances between WAPs, rebar-reinforced cinder block walls, and dozens of other users on each WLAN channel.  You might be lucky to get even my 36 Mbit/sec result.

I still think wireless HDMI will give a better user experience.