Second Penguins

So Apple is coming out with their couch computing platform called the iSlate... no, iPad. And I'm not going to talk about it.  I have three reasons why.

First, I've just gotten back from a great weekend at the Further Confusion conference, and I'm a bit laggy.

Second, I cannot for the life of me see how a $499 10" iPod Touch would fit.  Seriously, that's the iPad in a nutshell.  A 4:3 format, 10" iPod Touch with a tweaked UI to accomidate the extra screen real-estate as well as the hardware to do 720p video.  I have that with my netbook (which has it's own issues) and my laptop (which gets 1080p). Do I really need to spend $499 US (plus $100 for each doubling of the storage, and $130 for AT&T 3G data connectivity plus a pre-paid $30/mo unlimited data plan) for something I already have? I can't see how much use it would be.

Third, and most importantly, it doesn't run Second Life natively!!!

That's the kicker right there.  Nobody has ported Linden Lab's open sourced version of their viewer, Snowglobe, onto the iPhone or iPod Touch.  Now, more than likely, there will be some demand but the port won't land until the next generation iPad comes out.

A lot of folks are saying "next generation, please" on Twitter, and for good reason. No back-facing camera for one. No GPS, so you can't geocache.  No front facing camera ether.  The mechanism for getting files on it is unclear, and it all needs to be shaken out.

No, I'd rather say that one of my predictions came true, curdosy of the guys behind the Emerald viewer:  You can now have multiple attachments on one attachment point.

It turns out to be an artifical client limitation.  There's an XML file that says how many attachments you can put on each point.  On the offical client, it's set to one.  Replace the file with the one from Emerald, and you get more.  I want to set it to three: One for the hand replacement, one for a wristband, one for a gun or weapon.

This begs the question, why was it set so low in the first place?  My feeling is that Linden Labs is very worried about client performance, so much that they don't want to get burned with introducing new features in the next generation viewer, version 2.0.  There's two of these features that folks really want: Avatar masking and flexible sculpties.

The first, avatar masking, I've covered before.  It's the focus of JIRA bug VWR-822, with a ton of discussion.  Everyone has been clamoring about it because it would solve many problems; allowing users to knock out sections of themselves while not pushing up their prim count.  How much of a rendering cost it would be is up to anyone's guess; I haven't gotten my hands on a developer client yet.

The second feature ranks up there: flexible sculpted prims (or flexible scuplties).  To understand them, I'd have to get a bit technical and a bit sexual.

Remember that sculpties use a texture, assigning the X, Y, and Z coordinates that connect a sculpty to the R, G, and B components of each pixel in each texture.  Each pixel is of course linked to each other in four ways, and they wrap around on the sides.  For simplicity I'll refer to the texture's columns as the U coordinate and its rows as the V coordinate.

In previous columns, I used U to be an angle and V to be a distance along a rotating axis, in turn lathing out a single breast.  This turned out to be a good thing, as I made each breast so it was larger at the bottom of the sculpted prim texture - At V=0, the flexible sculptie is tied.  The rest wobble like a regular prim.

What does that mean?  Well, those with large implants on their chest can now jiggle, and tails that required multiple prims to sway and swirl around can be reduced to one sculpted prim.  According to some developers, the GPU load on the flexible sculpties is little more than a regular sculptie.

We need to start poking Linden Labs about these new features, then.  And poking hard.