Workshop at Stanford - January 29th 2009

This early workshop was for intended for the Sirikata stakeholders at Stanford

Attendees

Henry Lowood, Henrik Bennetsen, Gordon Knox, Susan Rojo, Helena Patricia Nordtorp Joergensen, Daniel Horn, Ewen Cheslack-Potava, Patrick Horn, Chris Platz, Jeffery Schnapp, James Neal

Intro

Henrik Bennetsen

Dev and community mailing lists are up
sirikata.com is started. Efforts will aim at building technology and community.
other universities, ex-linden people, intel showing interest
Delivery platform for the next second life (rather than the next SL), MMOGs, etc...
Sirikata will be published under BSD license (wikipedia). (Second Life (SL) is out under GPL license (wikipedia); hence devs on projects like OpenSim cannot look at SL viewer. Sun's Project Darkstar is also out under GPL.)
Being as open as possible to make Sirikata as useful as possible - the pragmatic approach
“alpha in some months”
We need to be more than just the pitch that doesn't actually use it
By having people interested in actually using it should lead to success.

Sirikata Architecture

Daniel Horn (Stanford dev team, programmers) Architecture of the project. Guiding principles:

scalability - use as thesis? Support lots and lots of users and objects
security - secure exchanges, proper ownership, access control, circumventing security is a sport we will work to prevent
federation - networking networked worlds, multiple entities managing the world vs just a linden, e.g., the internet - not only one company running the internet or web
extensibility - allow users to add content, scripts, plug-ins
flexibility - different types of things can be built on top of it, games, academic, museums, wide range of things
standardize protocols - interoperability, needed for federation, cooperation

Design decision (100K view): (split project into three aspects):

space simulation (physics, collision detection ... on one set of computers)
object simulation (scripts ... objects on another set of computers - "simulated elsewhere" - behavior of objects separate from storage)
storage - split out variable storage from static storage (CDN). Saved with different time resolutions and technologies. Variable storage involves things that change a lot (location of objects, e.g.) and static storage involves the long-lived objects that change infrequently. Static storage is the solved problem, e.g. S3 from Amazon (leveraging existing work).

Geometric segmentation and object segmentation -- how are the entities distributed among the servers?

Scripting language that makes it easy to create content is very difficult to write. Problem for SL, but also a more general problem.

Question: Problem of authentication locking out content. Can this be solved for archiving? No easy answer

Ewen Cheslack-Postava Arch details (10K view): specific problems that need to be solved at a lower level of development work.

Space – multiple servers, geometric segmentation, object segmentation, location, proximity – taking messages from one object host and send it to another object host Object host – accept messages from space and deliver them, utilities, download content and messages, different languages could be used Persistent storage – not much done here yet

No clients – clients aren't that special (often considered first class citizens) just treating as an object for Sirikata.

Will use this slide (link to slide needed) to introduce developers to the project

Interested in better languages Writing in second life isn't easy...not linden's fault

Future areas of concentration: Admin tools - Inter-space services – world of worlds? Naming, ownership Federated space hosting - objects Convenient object hosting – languages, nicer prims, Interoperability – want to link with current and future existing worlds, gateways to sirikata

Q&A Multiple entities to run a world together? Handoffs is an open research question. Solved socially rather than technically – sharing the load...I won't mess with you if you don't mess with me.

Points of access – (SL is gated...you can't come in via a browser) how is the world entered... Well defined protocols will allow anyone to write a way to enter. Want to connect from iphone eventually.

Authentication – for the long term, like copyright...after a certain amount of time the copyright is over. Can we design something that allows for opening the “time-based cryptographic controls.” Code is law issue. Copyright worked better when it was harder to copy. =)

Patrick The Code (1K view) what we have in the system Plugin based, multiple languages,

pluggable -
portable – works on any machine and serve
compatible with other architecture like meru developed at stanford

event system – listen for and send events (security issues) plugin architecture – c++ object host system – some code at stanford could be used, link to a binary to existing system scheduler – internals, events, loading things in the background cdn – lower level stuff is done, basic caching is done, but need name lookup webserver to host objects - using hash – filename to encryption hash for security: sirikata//object// dependency loader – world > room > people …. coming back to say x has loaded static vs dynamic objects space server – physics...any object can do whatever, loading meshes, object host system – not yet started, graphics system – need gui, using ochre

dependencies for the mesh loading

Open Source Virtual World Overview

James Neal - Open Metaverse Foundation

review of different projects. Problem of incompatibilities between OpenSim, Darkstar, Croquet.

Question: Separation of assets and game engine -- historically solved problem of having to hack code to make changes.

Stanford Humanities Lab themes and content-oriented issues

Jeffrey Schnapp - Lab founder and co-director

collaborative authoring and content producers -- struggle with limitations of SL.
how to extend practices in education and scholarly institutions that involve virtual worlds, e.g., museums and universities.
humanities lab as like a technology lab but driven by humanities and arts agendas
archives, the past, records, history -- core to concerns of SHL
animating archives
how can we use virtual worlds to create records of events, either time-based or space-based, that allows them to be recorded and observed for future use. Example: Museums. Virtual worlds could be a method for delivering the installed exhibition experience to the future. Perhaps change the contents. Many obstacles to this -- e.g., content creation tools that are available in a world like SL. Aim is to capture content that is not easy to capture.
conservation objective is to preserve digital materials, e.g., virtual worlds, indefinitely.
also interested in virtual worlds that are different from historical worlds; exploration of fantastic worlds. Animation studio.

Preserving Virtual Worlds

Henry Lowood Curator and SHL cofounder, PVW, HTGG, etc...

Interested in Virtual world 1. History of Vws – when were they first talked about 2. Machinima – archives and early history, VW as a medium, realtime animated movie making, we should engage this community 3. Legal issues around virtual worlds...user created content, machinima created by assets created by others, legal protocol rules for machinima proposed, authentication issues (vs preservation) 4. Problem of preserving what happens in virtual worlds

Virtual worlds die – closed worlds: EA-land, google lively, etc...

Captured the actual end of the world (EA-Land)

had to create a new account but couldn't since the world was closing so had to find someone who had an account
video capture like being on the bridge of the titanic
ends with an error message

People form attachments – how to we capture that so people in the future can understand? How do we handle the software preservation and preserving events in a given world

Everything that happens there is just 0s and 1s....you could have perfect capture of the world if you had every version of the software and database but is this feasible?

History of unusual or defining moments in the world

anti-front national protests in Jan 2007....it was a protest and griefing that just disappeared from the vw within 24 hours (maybe less)
WoW funeral where attack took place in world on disarmed avatars

VW as exhibit space

level of Doom and its geometry in world...bring in bits?
preservation of this (and other) historical artifact
exhibit a speed run at the speed run exhibit in world

Historical repository within sirikata

Lynn Hirshman exhibit from sl as example

Preserving Virtual Worlds - NDIIPP grant

moving beyond the easy articles such as books and papers
scoping and background research
schema development
implementation (archiving of test cases) and wrap up

If we can solve the issues around preserving virtual worlds then all other digital preservation issues will be solved.

Repositories: Stanford, UICU and Internet Archive and maybe LoC

An Artist's perspective

Chris Platz - art lead for Sirikata

Lessons learned:

Even at Stanford: getting quality contributors is difficult
rapid prototyping: allow high-level content import early
real-time design environment is quite a rush. Very appealing aspect of working in these spaces
technology that is not ready is very difficult to work with
goal-oriented approach -- deadlines are needed

Chris brought in about ten students to work on the project.

Snapshot into virtual world design. "scary slide" Captures all the work that had to be done to create assets and graphics.

Most tools available are far too complicated. So what is needed in a simple tool?

education design that can be used for teaching
collaborative runtime environment
an elegant interface that is easy to use

goals:

real-time collaborative importing and manipulation of 3d objects
real-time collaborative vertex editing
real-time collaborative digital painting
sharing and connecting to other virtual spaces and worlds

Funding Opportunities

Gordon Knox

Projects:

RES project (Puglia, Italy) and others
Talked about fund-raising for projects that could be mounted in a virtual world