[openbeos] Re: openbeos Digest V8 #104

  • From: Ingo Weinhold <ingo_weinhold@xxxxxx>
  • To: openbeos@xxxxxxxxxxxxx
  • Date: Thu, 26 Jun 2008 16:25:41 +0200

On 2008-06-26 at 12:33:58 [+0200], Stephan Assmus <superstippi@xxxxxx> wrote:
> Raymond C. Rodgers wrote:
> > This is really good news if we're on the same page. Just to clarify, the
> > X forwarding that I'm referring to isn't just about running VNC, it would
> > be analogous to having the app_server on machine B managing and executing
> > a program and all of the user interaction occurs on machine A, and for
> > all intents and purpose the program looks as if it is running on machine
> > A. You don't get a desktop in a window like with VNC or Window's Remote
> > Desktop Connection, just the application's windows.
> > 
> >  I sincerely hope that we are on the same page. :)
> The feature is certainly very cool. I can even imaging a solution to the
> BBitmap shared memory problem. It's more a matter of what happens when you
> call DrawBitmap(). The BBitmap memory doesn't *have* to be shared with the
> app_server instance that draws it. On a local machine that is quicker of
> course, but it doesn't have to be that way, so apps would work unmodified.
> It's a matter of the implementation of the link protocol. I don't think
> making a networked version of it is a huge technical challenge. There may
> or may not be invidual problems with running applications remotely, but we
> cross that bridge when we get there I suppose. Most likely not until after
> R1 though.

Instead of letting the application communicate directly with a remote app 
server, it might be easier to let it just talk to its local app server, 
which would transparently forward all necessary communication. Bitmaps would 
work just as usual. The only challenge is to find out what parts of the 
application changed manually in order to minimize the data to be sent to the 
remote app server. The local app server could maintain a copy of the bitmap 
for comparison. Or maybe it is even possible to add some special support to 
the kernel to find out which pages have been modified.

> Anyways, we are certainly on the same page as to whether or not this is
> cool. :-)

Yeah, definitely a must-have, if you ask me. :-)

CU, Ingo

Other related posts: