[X3D-Public] multiuser mode (newbie question)

Paul Aslin fabricatorgeneral at yahoo.com
Wed May 5 22:54:57 PDT 2010



--- On Thu, 6/5/10, John Carlson <john.carlson3 at sbcglobal.net> wrote:

> From: John Carlson <john.carlson3 at sbcglobal.net>
> Subject: Re: [X3D-Public] multiuser mode (newbie question)
> To: "X3D Graphics public mailing list" <x3d-public at web3d.org>
> Received: Thursday, 6 May, 2010, 1:25 PM
> 
> On May 5, 2010, at 12:23 AM, Chris Thorne wrote:
> 
> > 
> > DIS and XMPP are examples of application protocols
> that the network sensor was designed to support at a
> portable node level - i.e. without requiring a proprietary
> java or C implementation. Perhaps there are alternative ways
> to do this in portable X3D code, I don't know.
> > 
> > "Perhaps more effort needs to be put into an easy to
> use, standard dynamic scenegraph that is implemented across
> all browsers, and the question about MU protocols will go
> away." - only if it also addresses peer to peer and
> client-server communications as well.
> > 
> > Anyway, I agree with your summary statement.
> 
> I think stuff like Ajax Push and BOSH over HTTP to server
> will be the solutions in the short term for multiuser
> stuff.  Later, we will need to uniquely identify
> objects as they flow through the web (UUIDs and GUIDs) 
> Perhaps stuff like CORBA will make a comeback.  Right
> now, HTML has id's that make items unique within a
> page.  They need to be extended to be global (within
> one world) or universal (for cross world identification)
> ids.  Perhaps that was what someone was talking about
> when they were talking about identity.  I see no reason
> not to use URI's or URN's for GUIDs within the WWW (perhaps
> that is why ids in JSF use URN-like syntax). 
> Interfacing the WWW to stuff like XMPP or CORBA or other
> proprietary protocols will likely require UUIDs.  One
> thing that is needed is to make UUIDs and GUIDs hard to
> guess/remember, for security reasons.  See
> webkeys.  No one should have to look at these
> ids--perhaps they should be in an encrypted package that
> goes along with the unencrypted stuff, and filled in with
> XPath by the browser.

Thanks John, this pretty much sums up why I gave up trying to use the NetworkSensor nodes. Too much debate about which programming methods to use, and not enough practical implementation or experiments.

What possible value is gained by using a soup of non web3d consortium based communications protocol over simple gzipped xml or classic encoded x3d mini file, or x3d binary format ? Besides making the traffic a few bytes smaller, the extra programming needed for both client and server is going to add noticeable extra expense.

From my direct communications with Chris Thorne, standard http traffic (over tcp/udp) was intended to be an option. For those of us who want to experiment with server and scenes without heavy programming, it would have been very useful way to explore the possibilities of multi user.

ps: An opensource MU X3D server just isn't a viable idea, it would never attract enough interest from programmers or other developers.
Lots of complex and time consuming code is needed, for avatars/gestures, storage, shared events, editing and admin. I can't see the consortium or other groups paying key developers for what is needed to make anything half decent.


      




More information about the X3D-Public mailing list