I’ve been looking at what other people have been suggesting as necessary features for getting business done in virtual worlds. Since there’s a lot of commonality between Second Life, OpenSim and realXtend, I’ll use the term “Second Life”-like Worlds (SLLWs) to describe the family henceforth. The following ideas were mined specifically from a thread started eight months ago on LinkedIn by Mel Casey. They could inform the product managers of any of the SLLWs. (Does OpenSim even have anyone guiding the development?).
Below is a selection of comments from the abovementioned thread, with some edits to improve readability. I’ve added the contributors’ names so you can find their comments in context. I’ve assigned my own categories to the ideas to organize future discussions.
- Real world lighting (Allan Sturm), Dynamic lighting (Jens Roland)
Allan’s comment, when parsed in the context of another comment he made about real world/virtual world coordinate transformation, suggests that he wants the lighting in a virtual world location to be linked to the actual lighting conditions in the physical world that it mirrors. Whether that means the solar/lunar illumination, or the state of the lights in the conference room is not stated in his short suggestion. Both are possible and quite interesting. Social cues about the hours when people are working are useful in global collaboration situations. Information about the weather might provide starting points for small talk, a recent topic of conversation in New World Notes.
Group Management / IP protection / Security
- create mass accounts, track avatar actions, change passwords for someone else, move ownership from one person to another, etc. (Barry Joseph).
- asset protection (from intellectual/copyright/tradmark/tradename theft, data loss) (Julia Hernandez)
- group management is needed and then I don’t mean the SL group things, but group/user management in a similar way of that of computer networks. Creating/deleting accounts, what permissions they have to see, etc. I think that the way groups are managed in SL is to cumbersome, to much tied to Land. There needs to be a outside system, where you land is like domains in networks. And you can attach users/groups to them, with there specific attritbutes. etc. (Jeroen Frans)
- simple group & ownership management (Jens Roland)
Given that the thread was created eight months ago, it is perhaps excusable that no one talks about single sign-on to federated grids, integration with OpenID, OpenSocial or SSOcircle, mapping of avatars to corporate identities, integration with LDAP, Active Directory, etc. Lulurun added a proposalto the OpenSim wiki to integrate with OpenID on 6/18/08. Various teams are working on mechanisms for teleporting between Second Life, OpenSim and realXtend. Ultimately these efforts will result in some kind of federated security model.
Creation of mass user accounts is fairly easy in OpenSim and realXtend from the administration console. The server command “command-script [name of scriptfile]” runs a command script containing server commands. The scriptfile can contain a series of commands in the form “create user [first] [last] [passw] [RegionX] [RegionY]”. This file is easily built using any scripting language or office products like Excel and Access.
What’s not easy from the console is managing groups, resetting passwords, populating the user’s profile with useful data such as their first life contact information, moving a user’s home region, or even getting a list of existing users and groups. All of these operations require new server APIs to be created, and from them, scripting interfaces to integrate them with command lines and web-based administrative interfaces. It’s not hard work, but it does require some focused effort. Surprisingly, there seems to be little written recognition of these issues in the Second Life Architecture Working Group uses cases document. (I guess that means I should add them.)
Integration with external systems and information
Desktop and application integration
- virtual workspace. Ability for an avatar to connect with and work in our database from SL. This would allow me to have all my remote offices login to SL and work in one large pit where we could allocate human resources to issues in what ever region was in the most need. Would assist in developing a better national team. Basically, controlling their desktops from SL. So dbase, email, browser etc. (Brian Regan)
- import documents from a popular office format (MS, OO…). Use webcam feed for videoconferencing. In-world web browser with Flash support. Video streams linkable to any suitable surface. (Giulio Prisco)
- more seamless integration to business applications and real world media. (Dave Elchoness)
- integration with the current learning environments like Angel. The sloodle project is interesting, but the commercial LMS’s are far more robust. (Larry Dugan)
No surprises in anything above. The real issues in the integration of office documents are: 1) the user experience; 2) the security of the implementation; and 3) the collaboration model with users who are not in SLLWs.
By user experience, I mean whether the user in the virtual world feels that the interface is both natural and efficient. If I had to edit a richly formatted Microsoft Word document using wiki markup in a notecard, I would regard that as an unnatural interface.
Security relates directly to the earlier problem of mapping avatars to corporate identities, but also involves issues not addressed by current desktop security models nor SLWWs such as being able to specify the visibility of a document to a user in the SLLW who has no document privileges in the corporate container that houses the document (file share, intranet, SharePoint site, CMS system, SLLW asset storage, etc.) There is no mechanism today to control what someone looking over my shoulder can see. realXtend and Wonderland (Sun’s virtual world platform) use VNC or XTERM-based systems to link an application on the desktop to a surface. VNC has real security problems at the enterprise level and across the firewall. Both VNC and XTERM need enterprise management support for security and compliance if IT is going to support sharing a document on your desktop via a SLLW outside the firewall.
Collaboration refers to the interaction with the document and its container and how the document is locked, how other users find out who has the lock, how they’re notified when the document is checked in, how changes are tracked, etc.
- one of the most important aspects needs to be that it is not a closed system but part of the ecology of social tools – information needs to easily flow in and out. Your update in-world can be sent out automatically to your Twitter or Facebook; a screen in-world displays information from an RSS feed, etc. (Barry Joseph)
Whilst there tools appearing for twittering and blogging from SL, there is no general architecture for feed support. Ideally, you’d like general support for generating feeds from sensors, chat, in-world events (teleporting, friending, login/logout, etc.) and general statistics generators (uptime, lag, objects/sim, etc.) so that you can build external event logs, blogs, tracking maps, search indices of text, etc., as well as bridges to myriad social networks and sensor networks. The use of virtual worlds as interfaces to energy and environmental systems requires efficient event-based interfaces rather than the all too common high-frequency polling approaches.
HTML & Media integration
- any media on a surface (web, Word doc etc) (David Burden)
- HTML on a prim, baby. Oh yeah, and prims on HTML. (Christian Westbrook)
Read-only media on a surface is achievable today using HTML representations of office documents (Microsoft & Open Office XML formats). The unresolved problem is how to support the numerous add-ins such as Acrobat, Flash and many others which can’t yet render to an arbitrary surface which will be warped by viewer. It looks as if Linden Lab is using the Mozilla-based ubrowser which should be able to accept most, if not all, Firefox add-ins.
What should the user interface be for a read/write office document? Should an SLWW invoke your favorite word processor in a separate window, in an overlaid window inside the application via VNC, or simulate a PC screen in your virtual world running your application? I prefer the latter approach as it models the single-use behavior we expect. I would envisage a new input mode along with chat/say/shout which connects yoru keyboard to the keyboard of the virtual PC until an escape sequence is entered.
Christian’s comment about prims on HTML is insightful. He suggests a subset of the general “Second Life in a browser” problem. Whether this is achieved by mapping a complex object to some HTML/3D object in VRML/SVG or X3D doesn’t really matter. It’s quite interesting to contemplate someone interacting with a object in a browser while someone else is interacting with it in a virtual world. (3D/2D pong anyone?) A related idea is creating viewports into virtual worlds which are viewable in the web browser. The MPEG interchange for virtual worlds proposal seems to address the requirement.
Asset transfer from other systems
- import/export, especially for Google Sketchup (Ed Lass)
Fundamentally, what everyone wants is a cross-platform bridge with plug-ins for every 3D format. You’d want to run this alongside your client (or server) to import content easily from commercial or open source modeling tools and be assured that content would be rendered faithfully at the correct scale. I’d also expect a bridge like this to import 3D bar charts from a spreadsheet. Export should be just as straightforward, with the proviso that the bridge should respect copy permissions.
There are some additional ideas in the LinkedIn thread which I hope to address in a future post. Please chime in with your comments on the above ideas or add additional ideas for enterprise functionality.