stinkingpig Wrote: > rudholm wrote: > > ... > >> meh. Each has its place. VNC's performance is horrible, but it is > >> easier > >> than X. It's probably a point of interest that the most popular page > on > >> > >> my website is this one: > >> http://www.monkeynoodle.org/comp/remote-x-cygwin-howto > >> > >> -- > >> Jack at Monkeynoodle dot Org: It's a Scientific Venture... > >> Riding the Emergency Third Rail Power Trip Since 1996 > >> > > > > You might want to give TightVNC a try, it sports some performance > > improvements over the original AT&T version. > > > > > I have. > > Also, your howto seems to improperly characterize VNC. Perhaps I'm > > mis-reading but you seem to imply that the VNC viewer and server > trade > > packets continuously regardless of activity. That isn't the case, > they > > only trade packets when there are events to share (such as mouse > > movements or screen updates). An idle VNC session uses no > bandwidth. > > > > > I haven't watched Tight with a sniffer, but AT&T VNC definitely polls. > > Additionally, you indicate that a VNC context is limited in size and > > pixel depth to the specs of the video card attached to the server > > hosting it. This is not the case, VNC contexts can be of arbitrary > > dimesions and quantity regardless of any video hardware that may or > may > > not be attached to the hosting computer (given enough memory, of > > course). > > > > > That might be a new feature, but it's still not as cool in my mind than > > X. Anyway, VNC has its place and I'm sure that Tight offers lots of > benefits that I didn't notice the last few times I tried it. It was > still a lot slower than RDP or X, so I didn't care to hang around and > find out what those benefits were. >
No, null-event alerting/polling is simply not part of how RFB (the protocol VNC uses) works, it's not a question of whose implementation we're considering. New feature? Not at all. VNC runs independent of any video hardware on the server, always has. vncserver requires no X11 server because it *is* an X11 server. vncserver doesn't even know if any local display hardware exists. In Windows this wasn't the case because that OS traditionally only supported one user context at a time so when you shared a session with VNC, you were sharing *the* desktop (which, obviously, was tied to a video card). But under unix-like OSes, this has never been the case. In fact, getting VNC to share the contents of screen 0 (i.e. the local video display) is something people have been trying to work out in recent years. Eventually, RFB will be implemented by hardware-based X11 servers as simply another modular extension. What were the particulars of the VNC setup(s) you experimented with? I'm curious because RFB is far more lightweight than X11 so performance is generally better over less-than-robust links. That was part of the whole point of VNC. So if you're seeing other results, I'm curious what the specifics were. I'm also curious because your use of vncserver seemed to be tied to the physical display somehow. -- rudholm ------------------------------------------------------------------------ rudholm's Profile: http://forums.slimdevices.com/member.php?userid=2980 View this thread: http://forums.slimdevices.com/showthread.php?t=22506 _______________________________________________ Discuss mailing list Discuss@lists.slimdevices.com http://lists.slimdevices.com/lists/listinfo/discuss
