StoryMachine outside-in

StoryMachine rezzes prims based on the instructions on a notecard deposited in the central Controller. This means that you don't need prim handling skills to create something and you have a version that can be readily archived offgrid, i.e. backup.

The instructions are not, however, very noob-friendly so I have been looking at alternatives for at least doing the basics (let's call writing notecards "by hand" Option 0). What follow are some initial thoughts:

Option 1: Nodes can be rezzed using four spheres with values used to store values displayed in hovertext. The first three represent the target node level (height), angle and name, the fourth being used to rez the node. Thereafter nodes can be manipulated using the node touch menu. This allows you to change node colour and, to a limited extent, link nodes (you basically assign one of three tags to a node using addTag and then execute a tagLink command for the chosen tag on another prim so the first sends a particle stream to the second). The spheres could also in principle be used as a HUD to save on awkward camera manipulations.

Option 2: Given that the actions in Option 1 specify the basic info needed, the operations now also chat to a logger so that the instructions are archived in such a fashion that they can be readily copy/pasted into a notecard. The order of operations may need rejigging but the basics are there.

Option 3: LSL has recently been extended to include so-called HTML-in whereby prims can be addressed from ordinary web pages. The spheres in Option 1 can thus have their values set and chatted to the Controller without an avatar having to be inworld! Moreover, sensors could be deployed such that current nodes and their settings are (with the possible exception of links) relayed to the web page for editing. I haven't progressed much beyond simply confirming that nodes can be rezzed under 2D web control. There are significant issues in terms of the number of URLs available, the fact that they change every time the object is reset (workrounds are becoming available), as well as security.

While Option 3 is much less immersive, it does get round the usability shortcomings of SL when it comes to recording text as per Option 1 (you could use the client web browser as a dialog to limit application-swapping). Of course, the options are not necessarily independent of one another. As we saw, Option 1 generates data for Option 2, as do Options 0 and 3. It is possible that the different approaches might suit different people depending on their familiarity with SL and StoryMachine.

While it would be interesting to be able to rez objects outside-in, as it were, it would be of limited use if people could not see the consequences. While a sensor is one possibility, an alternative would be for a screenshot to be taken at intervals as a kind of inworld webcam relayed to the web. That might be one way in which a larger number of students could interact with a build than the sim could actually hold. Second best, I guess, but perhaps better than nothing.

There are going to be a lot of innovative uses of HTML-in. Skidz Tweak, for example, shows how you can use it to control serving of notecards via a simple Notecard Scripting Language. As Skidz says, you could potentially use it to control inworld games "on the fly" and presumably that could be extended to training simulations as well.

Connecting some dots

Catching up with my TED talks, I found Tom Wujec's short talk very interesting. In summary, he identified three areas of the brain as being crucial to good information design: one to stimulate visual recogntion, a second for spatial perception and another for establishing an emotional connection. Trying to touch on each of these is likely to be a key aspect of good educational design in SL.

This maybe follows on from Erica Driver's blog to the effect that her brain recalls information in 3D, based on her experience with ThinkBalm's very nice Data Garden (a good place to see the benefits and challenges of 3D for business)

Of course, knowledge is also about integration and that leads us to the second coming of the Connectivism & Connective Knowledge course run by George Siemens and Stephen Downes. I am dubious that I will have any more success than last year in keeping up but this year I will try to motivate myself by looking more intently for parallels with what I am doing inworld rather than trying to network with others.

When it comes to making connections, my thoughts turn to inworld mindmaps. Although I remain committed to StoryMachine for the moment, two more have come to my attention, one called Ideaographer (simple in a good way) and one from ace scripter Salazhar Stenvaag while on SLED Esme Qunhua mentions yet another, Conceptualiser.

The saga of getting StoryMachine to output to TiddlyWiki is not quite solved yet but an interesting adjunct has arisen from belatedly finding a scrolling web browser that works inworld. Not perfect (hyperlinks don't work) but a useful advance nonetheless.

Mr Grumpy is on your case

JISC-CETIS just published an informal horizon scan. I commented on the absence of any reference to virtual worlds. Scott Wilson replied that (i) it was an internal/informal document, (ii) nothing much did seem to be happening apart from the continued existence of Second Life, (iii) he would be happy to hear my views.

It is tempting to complain that virtual worlds in education are the forgotten children of the 2007 peak in the hype cycle. To its credit, however, the recent CLEX report did mention virtual worlds as did the even more recent JISC Effective Practice guide. Moreover, almost every edtech journal of note has now had a virtual world special issue and JISC RSCs have organised several meetings. The self-evident truth is that educational applications of virtual worlds are emerging, albeit not spontaneously fully-formed (was that ever a realistic expectation?).

John Kirriemuir, with Eduserv sponsorship, has done sterling service with his snapshot reports. However, as the JISC-CETIS horizon scan makes clear, it is all too easy to forget virtual worlds (wilfully or otherwise) and thereby factor them out from future planning. Following publication of the CLEX report, UK Web Focus Brian Kelly called for an institutional blueprint for Web 2.0 but my request in the comments for something similar for virtual worlds seems likely to be ignored. This is despite the fact that Web 2.0 and, if you like, Web 3D have a surprising amount in common (and, heck, you can view web pages inworld too).

Anyway, by way of a cheap blog post, here is my summary for JISC-CETIS. I hope they find something of value in it. It does have a historical perspective as it isn't clear that the context is appreciated. I leave it to Scott to do the necessary extrapolation.

Well, my guess would be that there is more activity in the UK HE/FE sector under the general heading of virtual worlds than many of the topics mentioned.

Yes, SL continues to exist and there has been a focus in the past year on improving both stability and induction of new users with business and education touted as primary markets (as witnessed, for example, by a new series of case studies). There are commercial developments in terms of web integration (Immersive Workspaces) and behind-the-firewall options (Project Nebraska). A number of third-party tools have been developed that assist with interfacing to Moodle and Blackboard and with authoring procedural simulations. Progress has also been made with accessibility, e.g. virtual guidedog. Linden Lab have been running meetings for registered developers under NDAs about future plans that have attracted rave responses from cognoscenti.

Attention is widening to encompass other VWs, notably the SL-like OpenSim (at least two UK universities with their own grids; ReactionGrid targetting education and providing its own Nebraska-like solution), Project Wonderland, Croquet-based peer-to-peer tools such as Edusim and Qwaq, and the commercial 2.5D Metaplace which now runs in-browser using a Flash plugin. This raises issues of movement of avatars, information and content between grids. The US Federal Consortium for Virtual Worlds appears to be working towards establishing VW standards for use by US government agencies (probably Web3D/COLLADA-oriented).

Although Google pulled its 2.5D VW Lively, it has been taken on by the community (mainly Asian afaik) and Google itself appears to be developing in-browser VWs based on O3D. Microsoft has unveiled Project Natal which seems be aiming at alternative camera-based interaction with VWs (LL has something more down-to-earth under development) and greatly improved bot AI.

It is also important to view VWs as one component of the Metaverse Roadmap which also includes augmented reality, mirror worlds and life-logging. Integration of these activities will become an issue in due course.

In passing I should mention that I also responded to the JISC SIS study which does have a section for virtual worlds.

Taxonomic thoughts

Just a pretty picture grabbed from StoryMachine, being a representation of the taxonomic lineage of some of the mycobacterial species (red nodes) the students may be analysing next year. As ever, click for a more detailed view. The display oscillates between use of b and m nodes to distinguish the different levels. The major limitation is having just 6 levels (which I may rethink).

Lars Juhl Jensen has a more interesting SL widget for plotting dendrograms.

StoryMachine, meet TiddlyWiki

Last weekend's Helen Keller Day activities had the effect of raising StoryMachine from the proverbial grave where it has maundered for some time. StoryMachine chats text and rezzes nodes from a story supplied in a notecard. I modified it so that it chatted node names and links as they were rezzed. The voice applet used by the VHH group (responsible for Max the virtual guidedog) runs in parallel with the SL client and speaks chat as it is added to the log (chat logging must be enabled for it to work).

Discussions on the SLED list suggested that there was still some interest in 3D wikis, building off from Eloise Pasteur's venerable (but still useful) spidergram. One thing I was interested in was "flattening" the 3D wiki of StoryMachine so it could be available on the 2D web. Some time ago I played with TiddlyWiki, a single-user wiki where the software (in Javascript) and content are located in the same HTML file. I have modified StoryMachine so the node menu can chat codes for the TiddlyWiki chunks (so-called tiddlers). At the moment the user has to paste this code into the TiddlyWiki HTML file but it should not be difficult to automate the process.

The result is a wiki that runs on a prim and that can also be browsed and searched in the SL client (though hyperlinks out of TiddlyWiki sadly don't work and, indeed, crashed the client when I attempted to open an external browser). As the closeup below shows, links can be preserved (click the image for a larger version). TiddlyWiki also has the concept of permaviews that enables specific tiddlers to be displayed. This to some extent gets round the problems of scrolling being deficient on prims. Tiddlers are normally fairly small and will often display on a single page without scrolling.


Limitations? At the moment the process isn't automated and images are not supported, just text names and plain text notecards (including traditional wiki-style formatting). Text lines in notecards must be less than 255 characters or the line is truncated. While I'm trying to facilitate dynamic rezzing of nodes (using the blue buttons in the bottom left of the top image), the linking model for those differs from that used by standard nodes and isn't factored in yet. Having said that, export as tiddlers does at least provide a version that might incorporate dynamic node reuse and modification. Finally, at present there is no roundtrip possible, i.e. no way of getting tiddlers back into StoryMachine. While it is feasible, the inability to write to notecards means that a web database solution is preferable to much tedious cut-and pasting.

I think StoryMachine's tortuous development history and general user hostility makes it less than ideal for student use at present but the concept of using TiddlyWiki as an adjunct is, I think, worth bearing in mind for other projects.