Next entry: Followers
Previous entry: X3D and Ontological Space
X3D Community Blog
June 27, 2006
To Push the Jumbo Flyingly
X3D has come a long way. It is now humming along like a well equipped jumbo, with modularity thrown in. The new user interface and rendering initiatives will spruce it up nicely!
But wait, it’s not flying to full potential! Back when the Web was based on a pull-from paradigm, Web3D (then VRML) had plenty of good pull capability. The Web continued to evolve and became more bidirectional: much is now pushed back onto the Net. People are uploading models, photos, QTVRs and linking them to Google Earth or community portals. Chat, VOIP and videoconferencing are more commonplace and Massively Multiplayer Online Games, which require continual two-way interplay, are gaining popularity.
Web3D did not co-evolve and stuck with the pull model. There is no efficient tcp/ip networking node, no advance in http protocol support. What happened? One can use CreateX3DFromURL and LoadURL to push some URL-encoded parameters onto servers, access web services, and pull stuff back. This basic two-way http trickery is nice, but it’s like pulling the jumbo from a mini minor. You can do some more efficient and powerful tcp/ip networking inside purpose built java or C++ classes but that’s not portable. What we need is specification support for portable, efficient, two-way network streaming content.
So, if you too want to get the Web back into Web3D, help me charter a Consortium Network Working Group. Let’s get the jumbo really flying!
Comments
That is an excellent suggestion. Silently and almost invisibly, X3D is getting traction in private but large scale integrated network apps such as are exemplified by Eric Maranne’s work in Crisis Management. If this ember becomes a flame in the public safety industry, the need for reliable networking connections AND massive content is a reality. To achieve this, building content has to be easier so the world builder can concentrate on the high levels of modeling and not networking or even simple logic for opening and closing doors. I noted the BitManagement announcements for courseware and proto libraries. I see Herbert Stocker’s excellent paper on linear filters. I see the KML convertor from Media Machines so content can be rapidly repurposed. The consortium members are on the right track and advancing. Hoorah for that.
But you and Eric are absolutely right that the networking nodes, something that has been discussed at least as far back at when Late Night with VRML and JAVA was published, are the key missing piece and as Eric says, The Problem On the Table.
This is basically what we’re trying to do with VOS. http://interreality.org. Though it’s not based in X3D or VRML.
I would be very interested in participating in this working group. I would love to see something like this evolve into an open specification for virtual worlds, in the same way that the w3c evolved standards and protocols for the first generation WWW.
Some things I have been thinking about are: a system for standard identity management (for authentication and portable avatars), IM communications and avatar gestures for social interactions, and the distribution of world objects (perhaps using p2p to swarm x3d files).
One design approach would be to use BitTorrent-like trackers to anchor world scenes using URLs, discovering who else is ‘present,’ and sharing bandwidth with other users to swarm the large files inherent in a 3d environment.
-Mike
Thaks very much Len and Reed for your comments. As for VOS, I hope we can use the network proposal to get some interoperability going with it as an object store is an important aspect of virtual worlds.
Post your comments
Note: All comments are moderated by the webmaster. Until then, the comment won't appear on the entry. Thanks for waiting.