HELP regarding colors
dls2 "at" Lehigh.EDU
Tue, 24 Nov 1998 12:15:20 +0000
>> I can't do otherwise. :p Starcraft, like many games, resizes the screen,
>> changes its color depth. When such happens, existing VNC connections
>> are killed, and so must be reinitiated. An extension for quick protocol
>> renegotiation might be nice. I don't, however, see a need for adding such
>> to VNC in any immediate timeframe, unless resolution change induced
>> disconnects are more common than I believe them to be. (very uncommon)
>Okay, I've been wondering why this question hasn't been posed before --
>*why* are you trying to play a directx accelerated game on a remote
I was not. Lag is certainly prohibitive of that. I was trying to see what
is capable of. Evidently it does not handle DirectX well.
>You know what directx is for, right ? direct access to the hardware...
Direct access to hardware, yes, for the sake of hardware acceleration only
on the Windows platform. Other than that, it is just another graphics
interface like OpenGL. Remote display of OpenGL graphics has previously
been mentioned on the VNC mailing list. Nobody disparages its usage.
>So go you and give it *the* slowest interface possible... you try to get
>it to draw on another computer over a network ??? what were you thinking?
Over ethernet, bandwidth is less of an issue than latency, and
I believe that ms (not MS ;p ) delays are quite acceptable. Also,
why do you believe that games are the only practical use for
>> Your thinking turns out to be right with regard to DirectX and its lack
>> reliance upon GDI in full screen mode. As such, no amount of enabling
>> or disabling DirectX hardware acceleration has any effect upon VNC's
>> use of GDI hooks. I might as well leave DirectX fully accelerated. :p
>> Ach well, unless VNC expands to include support for sound, then I see
>> no reason why DirectX support should be addressed, or otherwise
>Both of these things I'm sure are possible, but anything that requires
>directx support is probably going to work very poorly, as it is, by
>definition, going to want the fastest possible connection to the
>display hardware... as opposed to the slowest possible access to the
>display by putting the displayo on another computer over a network. And
>sound is going to require streaming audio over your network, and a whole
>nother set of messy protocols.
Audio and video are NOT that different,
and DirectX is not inheirantly fast or slow.
>> VNC would definitely be better with the following:
>> *Multi-user, multi-sessioning, with consideration given to the
>> *between shared and independently viewed client/server relationships.
>> *Furthermore, multiple logon types entailing multiple permission sets.
>As has been said before, this can be done with one of the following:
>1) expensive licensing from microsoft
>2) rewriting half the operating system
>#1 will not work, ever, as this is GNU stuff.
>#2, hey, feel free to implement it yourself....
1. Linux is GNU, and VNC runs atop that.
2. Likewise. :p
>> *Effective multi-casting to large audiences.
>I say again: *WHY* ??
Due to the fact that in certain, possibly academic or
corporate, settings, such functionality can be useful.
If its not useful to you, then its not useful to you.
>No, wait, lemme guess... you wanna broadcast your starcraft games across
>the internet ? <cringe>
The internet is not limited to just 33.6 and slower, and DirectX allows for
more than just games. Also, MS seems quite committed to the standard,
so its not disappearing anytime soon.
>The last time I heard anybody talk about repeaters & multicasters of
>anything like this, they were discussing broadcasting online quake games.
>I think I was around to witness the 1st spark of this thought, and I think
>they've actually pulled it off to some extent (go find MPOG, or visit
>#clanring). This, I would say, is only anywhere near practicle on a
>*much* lower level, hacking the game's protocol itself, and rebroadcasting
>its data packets, and having every viewer run a client that renders the
>data locally. *Not* transferring bitmaps of every entire frame all across
Why would you say such?
>> *"Drag 'n Drop" file transfers with NO command line or menu capability.
>> *Modular file transfer protocol.
>Could be useful, but would probably require too much coding to be
Maybe, maybe not. A modular interface allows certain aspects to be
entirely ommitted, while still encouraging functionality. Consider Winamp.
>> *Modular modes for display. (Hextile, CoRRE, RRE, Raw, and others)
Um, because those are currently the defaults, each has situations
for which one is better than another. Furthermore, just because those
are all that are currently implemented does not mean that those will
always be enough, and other algorithms may work better still.
>> *Encryption; modularly integrated.
>> *Compression; modularly integrated.
>Compression, encryption, sure.... modular ?
Why not? Compression is yet another form of encryption.
MP3 is a compression format. Winamp incorporates that
modularly. Simple piping is all that is required for such data
>Yes, it's good to look at all possibilities, and consider the best way to
>implement all things. I encourage you to try implementing these things
>yourself. You've got the source code, that's what open-source is all
>> Would VNC be more aptly described as VND (Virtual Network Display),
>> allowing for VNA (Virtual Network Audio), or VNS (Virtual Network
>> Sound), with VND and VNA|VNS being components of VNC?
The VNC mailing list - see http://www.orl.co.uk/vnc/intouch.html