Eats, shoots, glides

This is just a brief blog on a new build, viz some content related to bacterial differentiation. First up is Myxococcus xanthus and I thought I would try to show how the organism uses its pili to drag the cell along on solid surfaces, a mechanism that underpins a rather mysterious mode of motility called gliding. The bug normally dwells in soil as a swarm and feeds on other bacteria. With the proviso that almost everything is wrong (cell length, rigidity, additional motility systems, for example), I was quite pleased that I managed to get the cell to shoot out and retract pili (represented by flexi-prims) and move (three touch-activated cycles shown here). Quite how well this mirrors the real-life organism is, of course, the whole point so expect further iterations and hopefully some insight into whether pilus extension is more or less synchronous as shown here (yes, further research required). Ideally, I guess one would use physics-based prims to retract and drag the cell but that is for another day.

What's going on?

I'm currently "attending" the annual JISC Online E-Learning Conference. I won't pretend that there's been a lot of interest in what I do but that notwithstanding I have a little space to which I can post on the topic of virtual worlds and hope that people read and reflect. Here is a kind of catch-up post, a bit techie perhaps but I'll try to do the issues and pedagogy in a separate post (assuming time permits). Apologies for the lack of links -- Google is your friend.

One aspect I forgot to include was the Hypergrid. If you're interested in the metaverse more generally, you might like to check out Mal Burns interviewing Fleep Tuque on his Crossworlds programme (and yes, the Hypergrid model gets a mention). Anyway, here is my "pitch":

So what is going on in desktop immersive spaces? Here are a few interesting developments over the past couple of years:

• 2,400 avatars attended the Virtual Worlds Best Practices in Education conference held primarily in Second Life but with visits to many other environments used for education.
• Intel with OpenSim and the independent Aurora-Sim project, derived from OpenSim, both demonstrated regions with hundreds of active avatars.
• The US East Coast hurricane preparedness training done in SL by the Gronstedt Group was put to the test! Imperial College in the UK have been contracted by the UK Health Protection Agency to demonstrate the use of virtual worlds for emergency preparedness.
• UK company Daden demonstrated the Virtual Library of Birmingham to UK and Dutch Prime Ministers two years in advance of the real-life building opening. The aim is for library users to provide feedback to architects and library staff that can be fed into the final implementation. Daden have also been contracted to develop immersive visualization spaces for BAE as well as implementing training and visualization projects for the University of Leicester.
  
• The US military upgraded its OpenSim development facility under Project MOSES so that they could support "hundreds" of regions.
• Tipodean demonstrated a new browser that converts OpenSim regions to Unity3D on-the-fly so that they can be viewed using a web browser with the Unity3D plugin. Linden Lab demonstrated use of the Gaikai video streaming service to deliver SL in the browser.
• Kitely opened the beta of a cloud-based commercial on-demand OpenSim service supporting OAR uploads and Facebook login. This means that spaces can be readily implemented and scaled according to need. OARs are backup copies of entire regions.
• Mesh arrived in SL and OpenSim meaning that some of the less complex content in the Google 3D Warehouse and commercial sites such as TurboSquid (which also has freebies) is now available in virtual worlds. Tools that generate COLLADA output such as the molecular visualization package UCSF Chimera can also be used, as can photo-based services such as Autodesk's 123D.
• Both web-based and inworld content resource sites have emerged and designers such as Linda Kellie and Ina Centaur have made freely available significant amounts of generic content (hair, clothing, etc). The US NIH sponsored development of the Universal Campus OAR for UC Irvine which is freely downloadable and covers four regions with high quality multi-purpose buildings and content in a landscaped setting.
• Release of the Microsoft Kinect has spawned a sub-culture of hacks, including animation of avatars via real-life motion offering the potential of low-cost motion capture-derived animations for educational use.
  

Of course, the scene is not 100% unalloyed joy but neither is it the doom and gloom of yesterday's technology that some would have you believe. I'll cover some of the "issues" in a subsequent post.

Go Hypergrid!

I extended my orientation to include a level with suggested destinations, both on NWG and further afield in OpenSim but with an edu/cultural emphasis. There are only 10 places in this first edition but I did at least go and take snapshots of them. This first one (above) is the Universal Campus at its home campus (University of California at Irvine) where by chance I bumped (literally) into the creator, Nebadon Izumi.

Rutgers (above) and the University of Cincinnati (below) on the other hand have incorporated replicas of real-life buildings into their regions. Rutgers has some interesting trails to follow as well as the first football field I've seen inworld.


Other institutions include Edinburgh (Scotland), Sabanci (Turkey) and Masaryk (Czech Republic).
I've put the details in a notecard that drives the destination board. You just touch the left and right sides of the image to page through the sites and touch the middle to jump! The board uses osTeleportAgent but also chats the destination so you can use the link in chat history if your viewer supports it (e.g. Imprudence). If the destination is more than 4096 regions away, it warns the avatar that this is too far. Ultimately it should be possible to send it to an intermediate region while chatting the address of both that and the final destination so the second jump can be made with minimum fuss. Wouldn't it be great if a few more TPVs supported chatted Hypergrid addresses?

For those who want the detail, the paper by Crista Lopes on all things Hypergrid is on open access (pdf) this month only.

Hello, 'moto

It's been a while since I made an Animoto video. This is just a very basic one pulled together for the JISC E-Learning Conference where I'm offering a session on OpenSim (well, the bits I know about). I don't know whether it will attract any interest but it seemed like an opportune moment to grab a few snaps and let Animoto do its "worst". I used Imprudence to grab the images, shadows, WindLight and all...

SL Viewer 3.2 in OpenSim

I finally got a chance to try SL Viewer 3.2, albeit briefly. It basically works as well as any other viewer 2-alike on OpenSim but lacks Dolphin 3's grid chooser. As a consequence you have to use the --loginuri commandline fix to access an OpenSim grid. However, your most frequently used grid may persist in the Linden chooser which is handy at a personal level but no use for classes. As with all viewer 2+ series clients tested to date, it fails to load my main avatar on my admittedly very busy region.

Shared media worked fine on New World Grid but I was unable to test mesh as we are still on OpenSim 0.7.1.1. Accordingly, I used a scripted hypergrid teleport to visit a sandbox on FleepGrid and was able to upload, rez and rerez a simple mesh bison. FleepGrid was a little slower than usual to rez, especially the (sculpted?) trees. The ground texture was also mainly white but apparently that has been an ongoing viewer 2 issue on 0.7.2 (fixed this week).

In terms of the interface, the major change is removal of the widely reviled righthand toolbar. To be honest, I never had major issues with this once it could be popped out without displacing the main viewing area. That said, I do like the new layout whereby you can customise the button arrangement around the left, right and bottom of the viewing area. This means that you can hide irrelevant SL-specific buttons, e.g. to the Destination Guide and the new avatar swap menu.

The file named commands in the app_settings folder seems to determine what is available in the "toybox", the drag-and-drop button repository. There is also a file called toolbars in the same folder which appears to designate the default appearance rather than the set currently displayed (which logically would be in avatar-specific settings). I have yet to play with this extensively but it suggests it should be possible to configure initial button settings for a class, albeit that those displayed subsequently are user-specified unless you use the Restore button in the "toybox".

Anyway, the net effect is that you feel you are getting additional control and potentially extra viewing space. The icons used are good though you also have informative tooltips and optional captions.

The (rather small) IM chiclets have moved to the top right but do the job well enough. Likewise, the chat bar is now summoned by button and manifests as a (too?) small window which can be expanded to show chat history.

While Map search works, the map tiles themselves are not displayed although avatar dots are visible. Again, this appears to be the norm for viewer 2+ on OpenSim 0.7.1.1.

Click on the ground and the avatar does an average job of getting to the spot assuming there are no obstacles in the way. Once in the vicinity, however, he does not settle which is oddly discomfiting.

I haven't played around with appearance but I get the feeling that most of the changes are at the top level anyway. Going deeper, and much as I may have grown used to it, one thing I still dislike is the green arrow that backs you out of lower level menus. All too often you do not know where you are backing out to, choose unwisely and waste time. As a minimum there should be a tooltip on the arrow indicating the next level up. Likewise many important functions are encapsulated in obscure buttons (with gear and plus icons) at the bottom of dialogs. Also, we still have Show, Discard and Block on dialogs where notecards etc are given. Please bring back Keep; this is a much better description of what is going on as notecards, for example, are generally shown by default but only kept if Show is pressed.

On balance I think this viewer represents progress and I would very much like to see an OpenSim-specific version of it. I assume that Search, Profiles and Help can be redirected to their grid-specific OpenSim equivalents or switched off though do not know whether they adapt during hypergridding. All that is required is to remove (or make optional) the currency and shop icons, add a grid chooser and enable chatted SLURLs (or the OpenSim equivalent) for hypergridding. My main avatar loading issues are annoying but not not peculiar to this viewer. More importantly, it remains to be seen quite how the new viewer will be regarded by the key constituency of Third Party Viewer developers.

Primtings on OSgrid

I was sad when the Primtings art gallery closed in SL, both for its owner Ina Centaur and for the avatars that would be denied the opportunity to visit. I have long regarded it as an iconic build that I could unreservedly recommend to others so was delighted to see a tweet saying that it was being relocated to OSgrid. There is clearly still a lot to do but the essence is there already. It's easily found using map search; its coordinates are 10055, 9960.

Playing games in Second Life

Linden Lab's LR project in Second Life has introduced some new features which will shortly be available more generally, to whit NPCs (non-playing characters), the ability to automate addition of a HUD and the capacity to force teleport avatars to specific locations (in this case if they are killed by rock monsters or other hazards). A demo game has been launched for premium account holders. Further details are available from New World News and the Lab. Some uninformed reportage from yours truly follows.

[CORRECTION: It transpires that the rock monsters are probably not NPC but just scripted prims. 2+2=5]

The HUD feature was requested many moons ago by, among others, Rezzable so it is good to see it surface finally. In the demonstration game the HUD addition is triggered by colliding with a hypergate-type device which at the same time teleports your avatar to the twelve game sims. Once there you visit Tyrah's workshop. As far as I can make out (and remember, I am not a gamer) Tyrah is trying to arrange her evacuation by sending signals via flare cannons and your role is to track down crystals that can be used to make flares.


You then head out to collect the requisite number of differently coloured crystals, all the while attempting to avoid the NPC rock monsters. The HUD keeps track of your score as well as reminding you of the task at hand. Rock monsters are fairly quick but not all that clever so generally you need to be running and, if necessary, get an object between you and them as per the image at the top of the page (note also the crystal under the toadstool). If you get unlucky, you are force-teleported to Resurrection Circles, suggesting that this is not mediated by the standard damage system which would otherwise send the avatar home.

The score is maintained on an out-of-world server so you can take a break and resume play at a later date. It's not clear whether that service is part of the project but crystals of colours other than those required for flares can be traded for credits at the collectors at basecamp (the SL Marketplace is opened but I didn't notice whether any L$ were actually transferred to my account -- remember, this is just a demo).


The sample images were taken without shadows enabled so they are probably not doing the graphics justice. The (not unreasonable) assumption by the Lab is that residents will take these developments forward in some fashion and create experiences that others will want to play. Of course, there are complex combat systems already available for SL (and OpenSim) but, not being an expert, I'm unable to say whether they will have anything significant to gain from these developments. I can, however, imagine that the LR project will be useful in providing a social gaming context for n00bs. My guess, however, is that there will be some significant issues in terms of permissions being handled behind the scenes as being teleported in this fashion would otherwise allow griefing.

OpenSim already has NPCs though I suspect the code will not be interoperable with the SL version at present. The capacity to teleport on collision (and other events such as touch) is, of course, widely implemented via devices such as Hypergates. For me therefore the main feature missing in OpenSim is the ability to automate the addition of a HUD. This presumably involves both server and viewer work but nonetheless would be most welcome in allowing tailored experiences for students to be accessed with minimum friction.