External opengl applications integration

> Hello again
>
> I was wondering: is it conceptually feasible to cast opengl  
> applications within a croquet warp window? For now, it's only vnc-
> like external app integration, right?
>
> Would be truely amazing for packetgarden or google earth,  
> especially if the transition is seamless, and if the collaborative  
> aspect is kept (ex: take your students with you into a google earth  
> portal, and teach them a history voice lesson using history layers).
>
> Why not, using video capture/streaming...

> On Apr 20, 2007, at 7:19 AM, Florent THIERY wrote:
>
>> Hello again
>>
>> I was wondering: is it conceptually feasible to cast opengl  applications
>> within a croquet warp window? For now, it's only vnc- like external app
>> integration, right?
>>
>> Would be truely amazing for packetgarden or google earth,  especially if
>> the transition is seamless, and if the collaborative  aspect is kept (ex:
>> take your students with you into a google earth  portal, and teach them a
>> history voice lesson using history layers).
>>
>> Why not, using video capture/streaming...
>
> In general, is not feasible to embed an arbitrary 3D application such
> that each client has their own view (i.e. they can look around in
> whatever direction they want).  This is especially true if not every
> client has their own copy of the application.  However, even if  everyone
> has the application, it is not necessarily possible.   Leaving aside the
> OpenGL issues, there are serious issues around  synchronizing the apps.
> Taking the example of an embedded web- browser (non-VNC... each client has
> their own app)...
>
> A user clicks on a link, which loads a new page.  They have a really  fast
> machine with a fast net connection, so the page loads quickly,  and they
> click on another link that appears.  Meanwhile, another user  is still
> waiting for the first page to load when they receive the
> croquet-replicated click on the second link (which has not yet  appeared).
> Uh-oh, the applications are out of sync.  This is one  example of how
> things *will* go wrong without very careful design for  even a "simple" 2D
> application.
>
> I won't even get started on the OpenGL issues that can arise  (assuming
> that Google Earth even uses OpenGL on windows, and not  Direct3D... but I
> digress even after saying that I wasn't going  to :-); there are far too
> many of them.
>
> Video-streaming is the only fool-proof approach to sharing 3D
> applications.  Real 3D sharing may be possible for a *specific*
> application (especially if you have the source code for that
> application), but a general framework that will work with any/most/ many
> out-of-the-box 3D applications is currently out of the question.
>
> Josh
>
>>
>> Florent
>
>

> I think this issue of external integration (be it application, document, web
> service, or device) needs more attention. Many successful new technologies
> succeed because they incorporate what goes before them (color TVs  support
> B&W content, current HDTVs support NTSC/PAL content, cell phones
> interoperate with the landline network, etc.) I struggle to find cohesive
> documentation or solutions for easily binding existing ("legacy") data to
> croquet objects.
>
> For example, I've been thinking about some business simulations. I need the
> ability to present emails, existing documents (Microsoft Office, PDF, HTML,
> etc.) to a viewer, and enable actions that may bind to external
> applications. Some basic tools for presenting read-only content would
> greatly accelerate the development of interesting Croquet applications in
> the business or government sphere. I also need a variety of ways to interact
> with external applications, from Java-based apps to databases to
> proliferating media streams. We have basic ODBC interfaces, but not OLEDB
> and XMLA, much less some interesting 2D and 3D metaphors for interfacing to
> databases. We have basic MPEG/OGG/Vorbis support, but no way to load new
> codecs via generalized plug-in host for codecs.
>
> While Joshua makes extremely valid arguments for not embedding a web
> browser, there are many situations where the unpredictable synchronization
> would be acceptable. Heck, two people on different PCs don't even have
> guarantees that what they see in their respective browsers is identical,
> either in content or layout. I don't think that the scenario that Joshua
> paints is any worse. VNC solutions are problematic. (My company blocks the
> VNC protocol, even between two PCs on my desk.)
>
> -- Peter
>
> ----- Original Message -----
> From: "Joshua Gargus" <[hidden email]>
> To: <[hidden email]>; "Florent THIERY" <[hidden email]>
> Sent: Friday, April 20, 2007 2:15 PM
> Subject: Re: [croquet-user] External opengl applications integration
>
> > On Apr 20, 2007, at 7:19 AM, Florent THIERY wrote:
> >
> >> Hello again
> >>
> >> I was wondering: is it conceptually feasible to cast opengl  applications
> >> within a croquet warp window? For now, it's only vnc- like external app
> >> integration, right?
> >>
> >> Would be truely amazing for packetgarden or google earth,  especially if
> >> the transition is seamless, and if the collaborative  aspect is kept (ex:
> >> take your students with you into a google earth  portal, and teach them a
> >> history voice lesson using history layers).
> >>
> >> Why not, using video capture/streaming...
> >
> > In general, is not feasible to embed an arbitrary 3D application such
> > that each client has their own view (i.e. they can look around in
> > whatever direction they want).  This is especially true if not every
> > client has their own copy of the application.  However, even if  everyone
> > has the application, it is not necessarily possible.   Leaving aside the
> > OpenGL issues, there are serious issues around  synchronizing the apps.
> > Taking the example of an embedded web- browser (non-VNC... each client has
> > their own app)...
> >
> > A user clicks on a link, which loads a new page.  They have a really  fast
> > machine with a fast net connection, so the page loads quickly,  and they
> > click on another link that appears.  Meanwhile, another user  is still
> > waiting for the first page to load when they receive the
> > croquet-replicated click on the second link (which has not yet  appeared).
> > Uh-oh, the applications are out of sync.  This is one  example of how
> > things *will* go wrong without very careful design for  even a "simple" 2D
> > application.
> >
> > I won't even get started on the OpenGL issues that can arise  (assuming
> > that Google Earth even uses OpenGL on windows, and not  Direct3D... but I
> > digress even after saying that I wasn't going  to :-); there are far too
> > many of them.
> >
> > Video-streaming is the only fool-proof approach to sharing 3D
> > applications.  Real 3D sharing may be possible for a *specific*
> > application (especially if you have the source code for that
> > application), but a general framework that will work with any/most/ many
> > out-of-the-box 3D applications is currently out of the question.
> >
> > Josh
> >
> >>
> >> Florent
> >
> >
>