Darb Dabney’s OpenSim Blog

Yea sure, resident accounts do slide upward. I recall seeing concurrency over 42,000 on Sunday 3 June. What I’m struck by this blog cycle is how sticky the figure of about 1.7 million distinct logins in the past 60 days has become. Also, I find myself facing a tedious day now and then when checking “second life” as a search term on news.google.com and finding certain very persistent themes, and a day or two between what still seems new+exciting stuff. Surely I might be getting adjusted, but I’m left with a bit of a nagging impression, like the persistent 1.7 mil uniques in 60 days, that the active online SL community is more cruising along than growing explosively these past couple of months.

But in my little corner of Gualala, there has been some intense building and changes. To the detriment of RL household serenity, I’ve probably logged something more than 20 hours/week on Berkurodam construction in the past month. What’s made it more intense has been an initial, perhaps misplaced, goal of trying to get material together for the Fifth International Symposium on Digital Earth at the UC Berkeley campus. That hasn’t happened, but the desire to show off some urban literalism, and the Berkurodam build’s current theme of surface features and street furniture has driven a few refinements. The hard deadline will be to prepare posters for this month’s 27th Annual User Conference for Environmental Systems Research Institute (ESRI) in San Diego. That means that I’ve got to consolidate my feces in time for a flight on 17 June, only twelve days away now.

And happily, it is coming together almost proudly. A handful of, shall one say, creative insights has led to some new mural panels surrounding the build in nearly all horizontal directions, and a somewhat extraordinary unification of the parcels westerly and northwesterly of The Street. What’s really drawn in hours of my attention has been gaining dexterity with sculpties.

Thanks to the blessed prose from Amanda Levitsky, and the foundational Wiki Book Blender 3D Noob to Pro, I got my start over a couple of evenings. Then as so often happens with these builds, one success inspires a couple of other attempts, and if one of those works out, the chain reaction of skill memes continues unto my exhaustion or distraction by RL issues. I’ve learned that it is important to start with a cylindrical UV space to get Blender filling the output image square. (I haven’t yet found Blender functions to crop the resulting UV image, nor to stretch a generated UV grid along a single axis) Just as the sim in Gualala seems now to accept 1K by1K texture images, the sculpties seem to max out at a 64 by 64 grid of 24-bit pixels in UV space.

Although the Linden Lab descriptions of sculpted prim applications emphasized organic shapes, my need was strictly inorganic: fire plugs, traffic signals, street light fixtures, and deco sidewalk lights. For a public works / asset management application, it seems crucial to stay quite close to a 1:1 ratio between cataloged asset and prim. Thus, if I can represent an asset such as one particular traffic signal fixture (say a basic red-yellow-greed triad) with a single prim (rather than a linked collection of four, six, or more cylinder and box prims) then it remains a simpler proposition to automatically convert a single GIS-managed 3D point into a SL traffic signal. Vice versa, if the [x,y,z] position of that traffic signal and its [roll, pitch, azimuth] orientation are tightly refined in the metaverse, it seems most natural to extract those attributes from a single prim into a GIS asset point. All philosophy of balance between GIS and metaverse object counts aside, it usefully conserves sim resources to use one sculpted prim rather than six where possible.

The last variable in the past month’s excitement has been the exceedingly useful release of Google Street View for the RL site that Berkurodam is modeling. As carefully as I like to take my own texture shots, having Street View online as a reference for object geometry and relative positions is stellar. I surely remember the MS Virtual Earth street-level imagery from Fall 2006, but by comparison their interface was odious. Thanks to MS for breaking the ground in a large-scale street-level web service, and thanks even more to Google for publishing an interface to similar data with vision in its soul. As an exercise in 3D visioning, I have located a half-dozen “shoot the shooter” images where I grabbed views of the acquisition platform. In Berkeley, it was a green minivan with the cameras seemingly disguised to register (in peripheral vision) as a black streamlined rooftop storage bin. Of course, the 15-cm diameter survey grade GPS antenna mounted behind it gave away its purpose to knowing eyes. But very few people seemed to notice the van on Berkeley streets, which is either a tribute to the van’s low profile, or a measure of pedestrian preoccupation. Who knows?

Trackback URI | Comments RSS