Beyond the dead tree

The video above (from a nice post by Robin Ashford) highlights a more subtle conception of AR than the one I posted previously. It also highlights the belief that conventional media such as books might take on a new life. I'm currently reading Neal Stephenson's Diamond Age in which educational technology (in the form of the Young Lady's Illustrated Primer) shares centre stage with the best and worst of nanotech, not to mention human nature.

Of course, as in the video (and the novel), it is just as likely that it will be the next-generation iPad or Kindle that we reference, hopefully in some open format. We can see harbingers of this already in some of the innovative publication formats that have lately passed through FriendFeed Open Science.

• The Biochemical Journal in semantic format as seen via the Utopia engine, allowing readers to further analyse data (paper also reviews other approaches);
• The PLoS ONE collection of structural biology papers using the Molsoft ICM viewer;
• The entirely PDF-based approach used in this paper from Microscopy Research & Technique (NB 7 MB download; probably requires subscription).

Obviously we are, as ever, not "there" yet (closed source engines requiring an additional install are somewhat reminiscent of issues with our favourite virtual world) but, ho-hum, interesting times...

This cup of tea sponsored by...

The video above is a fascinating, if slightly scarey, take on augmented reality from a Masters-level architecture student, not least in the light of Google's recent patent for overlaying adverts on virtual buildings (via Digital Urban). The context for the video isn't entirely clear but raises the possibility at least that AR might be deployed on the back of (hopefully more subtle) adverts.

In an interesting interview with Daniel Terdiman on Terra Nova Bruce Damer extols the virtues of AR:

Bruce Damer: The 1990s virtual worlds and VR, and the 2000s (Second Life etc) and well into the teens, are kind of like the kinetiscope, you have to be locked into one place (your PC, laptop etc) interacting.

Bruce Damer: And the movement, the production values, are all pretty stiff, think of the keystone cops movie shorts

Daniel Terdiman: yup

Bruce Damer: When you get 3D spatial highrez AR, virtul worlds, data worlds, motion capture, all of that put together in your car, on your retina, on your skin and in your brain, we will all look back at people sitting for hours in front of glowing CRTs (or flat panels) and be amazed (well, it was a phase we had to pass through)

Hmmm. Sometimes simple is not so bad and the interview acknowledges the impact of 2.5D Flash-based virtual worlds, both browser-based (such as Small Worlds) and Facebook embeds such as Farmville and Island Life (what Metaplace did next). None of these, as far as I know, support user-generated content which I still see as sine qua non.

Of course, as the Terra Nova interview opines, we can also look forward to a future in which the technology of the movie Avatar becomes accessible to Joe YouTube, presumably as a kind of high-end xtranormal.com, providing another level of sophistication but not a game-changer (I vaguely recall YouTube has been quietly running 3D vids for a while).

Sadly, the boundless possibilities of user-generated content do not come without consequence and it is with regret that I record the passing of Ordinal Enterprises on Caledon in SL and endow its proprietor every best wish for the future.

Simple map on a prim

By way of a quick blog, there is a nifty bit of code I got via Ajaxian that allows you to render a Google map from text in an HTML element. It isn't always 100% successful and sadly the links don't play nicely with Daden Navigator. However, the page opens well enough to the appropriate place in Google Maps in the client browser and hopefully it should be possible to write some server code so place markers can be created from chat. That would be a nice addition to the map browser and a lot simpler, albeit more limited too, than my previous attempt.

Talking through my hat

As I seem to be wearing my fedora a lot, I thought I would give it the ability to generate particle images and chat text from a touch-initiated dialog. It needs a little refinement (the text should come from a notecard rather than being hard-coded) and suffers from the usual wobbble in the absence of a suitable static pose. The image above shows a display put together by a group of Honours students.

In other news, the Where Next for Virtual Worlds meeting has come and gone. Sadly the Twitter feed did not really convey any great sense of the discussions that took place. However, I got the feeling at a distance that the meeting was as much to mark the end of Eduserv involvement in SL as to plan for the future. Clearly there are people having significant success in SL, while others gravitate towards OpenSim and other virtual worlds, at the same time resolutely hacking at the branch from whence they came. While I don't know where I'll be next year (if anywhere), this phase leaves me cold and disinclined to invest energy in pushing my own ideas further.

To understand at least how we got here, if not necessarily where we're going, the architect Jean-Ricard Broek (whom I met transiently on my brief, laggy sojourn in ScienceSim) has compiled a history of SL that at least puts a few dates to events I vaguely recall.

More map browsing and general edu-thoughts

A while back I built a proof-of-concept metabolic map browser that rezzed molecules into position as you moved between maps held at different resolutions. The molecules were stored in Troy McLuhan's holoemitter. While I plan to go back to that, I wanted to see whether I could reuse some of the code to create a more generic map browser for possible use on the Africa sims. The maps would also come in handy for the maze rezzer which also takes Africa as its theme.

In the first instance I used maps from the Wikimedia Commons and sliced them as previously using ImageMagick. I eventually want to add a third level of resolution but at the moment the mat is still just an enlarged version of the righthand vertical display accessed via the lefthand touch-sensitive one, the actual browser component. However, I did add a second layer of mat prims so that I could superimpose a topographical map on top of the standard one. Sadly the two don't match perfectly yet but I think it is still useful and the idea could be extended. As ever, click the image above for a larger version.

Of course, it would be much more impressive to use a sculpted prim for the map and Darb Dabney and the folk at Nottingham have been doing just that, not to mention one or two SL globes built in the same way.

For the moment though I want to explore dynamic uses (display of geotagged images, for example) and the use of Puppeteer in this context. In the latter case the idea is to use animations or, indeed, touch-sensitive access to single frames to replicate some of the feel of Gapminder in adding the dimension of time to the build.

All this feels like functionality that should eventually be much more accessible. While it is OK for me to lash together a few prims, code and content from elsewhere, it is unreasonable to expect everyone to do the same. I hope that in time some of this functionality might migrate into the mythical edu-viewer, e.g. that the sim be able to flag that this content is geo-enabled in some fashion and that RL geotagged content in the equally mythical edu-inventory can be mapped onto it on a temporary basis. Hopefully it's this sort of approach (not necessarily a real edu-viewer but the idea that we might have one) that will encourage forward thinking on the ways in which virtual worlds can usefully interact with RL and other media in education.

In a nice blog Gráinne Conole observes that despite the availability of Open Educational Resources (OER), open platforms and exemplar courses, widescale adoption has yet to take place. She advocates an approach based on Learning Design with designs populated by OERs, the designs in turn being recycled (I think LAMS has a similar take).

Ignoring for the moment whether an iteration of the map browser could be used as an LD tool inworld (and with a temporal component, why not), the question I would ask is to what extent virtual worlds already provide much of what she seeks (answer: a lot, especially if issues of consolidation, repository integration and scaling can be addressed). I also would ask why in that case virtual worlds are so often seen as an edge case (no mention in her blog, no strands at the upcoming TechDis or Dev8D meetings).

My time for doing development is again going to be limited over the next few months (the maze rezzer is more of a priority) so this blog is somewhat by way of an aide memoire. It's also not clear that the browser is compatible with the aesthetic of its new home on the Second Nature/Elucian Islands Village sim so it may have to go into inventory soon.

JorumOpen: now with added caffeine

I'm on painkillers at the moment, nursing bruised ribs after a fall on an icy path a while back. I was slightly amused to note that paracetamol "extra" these days means "added caffeine", as if I don't get enough of that anyway.

A couple of days ago I downloaded the Sami tent from JorumOpen (JO). Let's be honest, it's not commercial quality (issues with camera bounce, sit poses and a fireplace that floats off the ground) but, if you need a tent, it will do and it has the bonus of coming with some nice teaching guides. I could make some suggestions as to how the tent could be improved (in fact, I just did) but JO doesn't seem to support that yet. I wonder whether they will adopt a full-on e-commerce approach with ratings, reviews, etc or leave it to third-party sites (I could easily tag this post with a unique id).

Having been snarky about the tent (remember, I'm not at my best), it does seem that some kind of inworld show-and-tell forum for developer feedback would be a good idea and maybe this is something we could build into the Second Tuesday sessions we're sort-of running every, um, second Tuesday.

Actually getting the tent wasn't 100% straightforward either. The slurl on JO took me to the sim landingsite and I had to follow the beacon to find the resources nearby, one a complete tent, the other a parts list so you can make your own. The tent itself appears to be no-mod, no-copy and looks as though it would be hard to modify anyway so having the bits is a good idea (you can, of course, take multiple copies from the Leicester sim).

The slurl-based go-and-get approach is OK, except that it can be a pain when it comes to having to manage lots of such items, both for the "shopper" and the "vendor". In the latter case, moving stuff round means you have to re-edit the JO record. Some way of tagging the item so you can co-opt the Google-driven inworld search would be handy. Having said that, it is great to be able to see a rezzed version of the tent on the Leicester sim and any search-and-repository strategy has to work with all the wonderful inworld educator sites (like International Schools), not against them.

Of course, most folk would probably like the option of something like an xstreetsl web-based experience (or better) and I think JO should be prepared both to do some of that and/or to link off to such sites provided the materials are OERs (we'll let metasites merge the educational commercial and OER worlds). Getting inworld delivery via some kind of magic box (as per xstreetsl) should not be beyond JO -- lots of people doing that now, the most recent example being Katharine Berry and her megaprim.sl site. Dedric Mauriac is doing something too.

The problem with both these approaches, however, is ultimately that you don't have an archived copy of the artefact, just a link to someone else's. If they decide to move it or delete it, it's gone: your fault for not getting multiple copies of that tent when I told you!

Which brings me back to caffeine. As I mentioned previously, Meerkat lets you archive simple structures so I used Hiro Sheridan's Orac molecule rezzer to generate a caffeine molecule, hit the Meerkat Backup option on the object rightclick menu to create an XML file and then uploaded that to JO. The next person wanting a caffeine molecule in SL or Reaction Grid doesn't have to search for Hiro's rezzer or the smiles code. I could still delete the JO record but at least the consequences aren't down to the vagaries of memory, coworkers or LL servers.

Could someone write some clever code to generate the XML file on a server? Probably. Then it would just be a link from JO, Wikipedia, Chebi, or Pubchem. Alternatively, the molecule might simply appear within the search space or extended edu-inventory of the equally mythical edu-client.

Rezzing the caffeine is simply a matter of File|Import|Import and selecting the XML file, though recreating the structure is somewhat time-consuming. Moreover, this is a simple example (no textures [OK] or scripts [not OK]) but ultimately the end-product is probably what we should be aiming for -- and what Google has established its 3D warehouse to furnish, albeit not necessarily with SL in mind. The major benefit, of course, is being able to move content between worlds with minimum fuss.

Search & repository

I'm not an advocate of a separate educational grid post-18 but in an ideal world I'd like to be able to swap out the SL search engine and repository system (by which I mean an amalgam of xstreetsl and inworld search) for one that is more education-oriented. The negative attitude of LL to freebie listings (including open educational resources, OER) on xstreetsl, not to mention the advent of significant competition to SL, are further reasons to be thinking in this direction.

The opening today of JorumOpen means that there is at last a central repository for OER and the OTTER project at Leicester has added the first SL OER to the database, a tent: http://open.jorum.ac.uk/xmlui/handle/123456789/562 (see picture above).

I have no idea whether JorumOpen will be more successful than its predecessor though the investment made by the JISC in this area can only help. The fact that its metadata will ultimately be discoverable by web spiders presumably means that it can be part of some kind of global effort (only UK staff can make entries) that would include the likes of CLIVE/MERLOT (http://clive.merlot.org/). It would be great if we could also incorporate suitably tagged inworld content (currently indexed for the SL Google search) and even upload content directly to JorumOpen, something that has become more feasible with the ability of third-party clients like Meerkat to export objects, if not their inventory, to local file systems.

How difficult would it be to achieve integration of the custom search with, say, a third-party client? I don't know but hopefully someone can be persuaded

What might we store? Well, not just objects, object collections or even entire sims (as per OAR files in OpenSim grids) but also documents, lesson designs and evaluations. Some things are possibly borderline for JorumOpen, including listing of current slurls for open university islands. Just because an island is open doesn't necessarily mean that anyone is welcome at anytime to come and teach there or that everything there is available to copy (though some or all of it might be).

Ultimately, of course, these repositories would all be accessible to lesson design software. I'm actually not sure that that would be enough to drive mainstream adoption but it's a thought.

Joining dots: Planning the Maze

One possibility when it comes to planning a Narrative Maze is to use Salhazar Stenvaag's Wiki3DBuilder tool, shown modelling a 4x4 maze in the picture. This is available from Apez for L$1 though the open source code is also in the LSL Library on the SL wiki.

Basically you rez prims by touching the cube, choosing NEWCUBE and then touching the side of the parent prim on which you want the cube to rez. A similar strategy allows you to move the active prim. For each node you can change colour, size, shape, hovertext, etc, and also texture, add a URL and object such as a notecard. The permissions can be set in the description field of the root prim (under the main array here) such that only the owner or the group can make edits.

Textures can actually be added to faces simply by dragging from inventory and would presumably best be added to the intended side rather than the top face as shown here.

The only issues I had were with linking and pruning. If new prims fail to link from the start, rightclick|Edit the main node and chose Reset Scripts in Selection. Pruning, i.e. removing nodes, seems to be an almost-all-or-nothing activity: only the root remains.

While the red particle links show the parent for each node, they can't (as far as I know) be reset or recoloured and thus it is rather hard to use them to trace a path through the maze. A systematic use of textures would seem to be the best bet.

It is quite easy for SL users with good camera skills to generate maps but I am not so sure about newbies. However, as the models are linked, it is at least easy to take them into inventory. Some operations also have an inbuilt delay inherent in LSL that can be a little disconcerting. Using a 2D web drawing tool would be another option. Generating an entire maze from a 2D web application is another interesting otion, albeit beyond the current NMC rezzer.

Joining dots: Extending the Narrative Maze

I've added the facility so that walls in the NMC's Narrative Maze drop when you correctly answer a question posed on a texture. This proved to be just a matter of making some minor changes to the code for the maze walls, the modified module then being incorporated back into the maze rezzer. On Eloise Pasteur's suggestion, I also modified the mechanism for avatar detection (used to chat module-specific text) so that it works from collisions; hopefully that will be less laggy.

The question texture was created in PowerPoint, resized and uploaded. The code uses llDetectedTouchUV so it is important not to move the buttons. The correct answer is currently located in the wall's description field. In the Release Candidate it doesn't appear on mousing over the prim but that may not be true of other clients.

At present the wall drops for 10 sec and then comes up again so that the maze can be used by multiple avatars at the same time. I decided to show the same question on both sides of the wall should the mazerunner decide to backtrack -- I appreciate that could be confusing. If the wrong answer is chosen, the texture changes for 3 seconds to reflect this and the option chosen appears in chat. Loss of time is therefore a potential penalty for wrong answers.

While I was tempted to regard this as a bit of fun to start with, I'm now seriously tempted to use maze construction as a generic activity for small groups of students. Maze use in SL is not at all novel: I recall Literature Alive having one, for example. However, for newbies it is a familiar concept, immersive and, if engaged in design, much better than working from a blank canvas. The design aspect is also challenging.

Locking and unlocking the module seems to be the biggest bugbear during authoring and there are still some minor bugs to squash. The effect of lag during use remains to be determined but having a large number of quiz textures is not going to help much in that regard. One further possibility is to have an additional central node that temporarily rezzes particle images and/or objects on touch.

Quite how practicable the system will be remains to be seen. What it does underline, however, is the wisdom of releasing generic tools like this as open source (apart from the actual rezzer) so that they can be further refined. Thanks to the NMC and Natasha Boskic, whose idea it originally was.

Joining dots: Narrative Maze

I'm looking at some approaches to sequencing activities in SL. This is somewhat contrary to SL's "anything goes" ethos but may be appropriate for newbies in particular.

One simple approach is to use the Narrative Maze. This was a prize winner in a contest run by the NMC. You can get it from the Learning sim (slurl) for L$1 (rightclick the sign and choose Buy) and there are some interesting quizzing objects nearby to check out at the same time.

The maze is conceptually very simple although oddly the NMC example doesn't itself make much use of the maze aspect. Building the maze is simple: you get a rezzer that is dragged from inventory to the ground. You touch the rezzer and then chat the number of modules you require, e.g. 4x4 for a grid of 16 modules, each surrounded by four walls.

You then create a path or paths by clicking on walls to lower them. When you are happy with the result, you can lock the configuration via a menu activated by touching the floor. However, this has to be done on a per module basis which is a little tedious.

By default each module has a notecard in its root prim. This can be edited so that text is chatted chat as the avatar moves into the module. Thus the journey through the maze can tell a story that might involve a degree of branching built into the maze.

However, I do wonder how laggy a sizeable maze might be (you could change the scripting to mitigate this) and how well it would accommodate a large class. Perhaps a large maze could have multiple access points telling the same story from different viewpoints but eventually converging on some kind of denouement at the centre with a teleport to the next destination. With a small group, teleporting to the centre at the start and then having to find an exit might be a possibility too. Getting students to write the notecards would be an interesting activity in design and collaboration as well as writing.

The walls can be textured and, of course, the maze could be populated with all manner of interactive objects. Beyond storytelling, it has significant scope for mapping spaces to concepts or spaces. One possibility would be some kind of taxonomic tree or key with representative images or models on the paths. Imagine also using it as a means to navigate a map or a conceptual representation of the body. The maze might also be a rapid way of generating a rather basic (and sombre) gallery or timeline. If you can code, the walls might be drop if you answer a question correctly. It could even form the basis of a PacMan clone!

The challenge of navigating the maze is tempered by the absence of a roof so that experienced users can easily cam over or through walls. I suppose a roof would be only a maginal disincentive for the adept though you could switch flight off to make cheating more difficult for novices. Of course, you miss out on the text so cheating is not actually very productive. I personally dislike low ceilings anyway so am quite happy to see this omitted.

Whether you choose to stress the storytelling or the branching/decision-making is up to you but provided you have the space this is an amazing tool to explore.

(maps in the demo image courtesy of paniglobe.com and sculpties from Bethi Catteneoon xstreetsl)

I am officially all over the place

Confirming what many have always suspected, I really am now all over the place:

• The University sim in SL (for 6 months)
  
• On one of the Second Nature islands (I hope)
• The non-profit Virtual Africa (I pay for this!)
  
• ScienceSim (for 6 months; my first serious foray into OpenSim)

This is interesting in terms of being able to construct multi-sim trails for both students and, perhaps, the general public. Each of the sims has its own rationale, its own story, its own dream, and I think it's easier to convey the diversity of virtual world applications when you can think beyond a single sim encompassing a just one use or vision.

Botgirl Questi has advocated "tracks" as the solution to one of SL's retention problems, viz "why am I here, what do I do next?". While her ideas are more oriented towards skills and practice, I hope a problem-solving trail that illustrates virtual world applications alongside science may also have merit. Of course, I don't want to limit trails to my own builds, so choice of topic will be key. Geographically, the trails are thematically defined by Virtual Africa: diseases of importance to that continent are a possible starting point.

The biggest drawback is the need to register avatars on two different grids. The biggest plus is having the opportunity to network with others and see what else emerges from that.

Predictions for 2010

I could easily recycle half of last year's (failed) predictions. I've seen elsewhere that merger with the Teen Grid on PG sims is probable (and probably mostly good for education) and that a new client with support for clickable HTML and Flash will arrive (muted yay after all this time). I am slightly sceptical about the TG, even though I would support it. Nevertheless, it may be part of a more variegated grid if predictions that regionalised support (afforded by server farms outside the USA) will allow practices permitted under a particular jurisdiction, e.g. gambling.

Anyway, onto my own predictions. These have a strong personal bias, i.e. UK university in a climate of significant funding cuts for higher education. An awful lot depends on the general financial climate and whether the UK can avoid a double-dip recession and whether universities see it as legitimate to continue support for teaching innovation under these conditions.

1. Hardware on campus will become more of an issue as universities try to economise by keeping hardware going for longer. Although there is a requirement for high-end graphics capability to support certain subjects, most students get on fine with the basics for word-processing, spreadsheets and web browsing. Care has to be taken here or an odd disparity may emerge between what students run at home, where they play games such as World of Warcraft, and on campus. Indeed, the better-off may increasingly choose to bring their own hardware to class, complete with their own mobile broadband. A measure for this will be the return of the presently dormant hardware issue as a topic for discussion on SLED following release of the new LL client (shadows and all).

2. The number of education and non-profit sims in SL will decline slightly as institutions assess their needs more stringently. There will be continued departures for other grids balanced to some extent by new arrivals coming in for very specific reasons. Merging with the TG would encourage a fresh look at the use of SL for marketing to prospective students as well as supporting a greater number of outreach projects with schools (most likely on closed sims) and alumns.

3. Development of a more edu-friendly search facility will be initiated and some third-party developers will allow you to swap this into their client in place of the LL default. The backend database will also act as a repository for cost-free distribution of content deemed of educational value, to include individual items, builds and sims. The database will not be limited to SL and will interoperate with other initiatives such as CLIVE/Merlot and JorumOpen.

4. There will be increasing exploration and use of tools that facilitate development of educational sims, both at the level of the institutional VLE (e.g. SLOODLE) and the inworld class (e.g. Pathways).

5. Web-based lesson design tools will extend support to SL and OpenSim.

6. As part of an overall consolidation and recognition of exemplars, there will be a trend towards subject-specific rather than institution-specific sims on the basis that this is presently a pre-competitive phase and that development and operational costs may usefully be shared.

7. There will be a trend for institutions, and more particularly departments, to share sims or rent them for the duration of specific classes rather than year-round. Sharing might be on a time zone basis and/or include collaboration with non-profits.

8. Takeup by education (as opposed to business) of the Second Life Enterprise "behind the firewall" product will be limited unless the pricing for the sector is highly competitive. The same is likely to be true for the Immersive Workspaces product.

9. Sims will increasingly borrow concepts from the games domain, serious or otherwise. The dominant mode of interaction will still be via mouse and keyboard though the arrival of MS Natal will catalyse some community hacks.

10. There will be increased use of SL by small groups in shared physical spaces, e.g. in libraries, using either projection or large LCD screens.

I could go on but I won't. As ever, there is a degree of ignorance, guessing and wish-fulfilment in the above and I won't be at all surprised if the final score this year is even lower than last. I am slightly concerned by ominous statements from LL of more pain to come but heartened by their desire to work more closely with the community.