Viewer Development
Viewer development process changes a bit. The Lab is changing the release process a touch too. More likely now we will see Third Party Viewer (TPV) Developers releasing betas and development viewers rather than production releases, the later being the stable version intended for general use.
The Lab has gotten much more metric driven about their crash rates. They are putting more effort into holding the lid on the Linden Viewer crash rates. This means code in the Beta and Development viewers are likely to see more testing and changes before new code makes it into the release viewer. That is good thing, but may slow things down.
HTTP Library
The developer’s version of the HTTP library is up for developers to start testing. The Lab is looking to developers for information on:
- Regressions – which are w/bad routers – This iteration of the code is much better at working with Linksys WRT units – Belkin is still not fixed. All the fixes that allow it to work have major performance issues. The Lab hopes to improve things, but is non-committal as to whether it is going to be possible with any acceptable level of performance.
- General experience reports
- Design comments and integration problems
Monty Linden thinks this is a near release version of the library. The testing plan is to get the library running in Preview Grids and development viewers. Then slowly move it to AGNI and the release viewers.
We’ll first see textures using the new HTTP pipeline reach users. Most other services are farther down the road, such as inventory. All of these changes are in early development and weeks from reaching the average user. But, there is progress and some, like textures, may reach us way sooner than I expected.
The HTTP Project viewer rez’s the textures way faster than the current Linden release viewer. It is about 3 times faster. Nice!
Mountain Lion
Mountain Lion is Apple’s new operating system just released. The news is just hitting the Apple blogs today, Tuesday August 7. It comes with 200 new features for Mac users. (Reference) See the reference for more information.
A Linden team is looking at how they can implement the Second Life support for the new Mt. Lion and support for the older OS 10.5. Of all the Apple OS’s 10.5 is the most crashy for those using Second Life. If your SL friends are using it, try to get them to update. Save them some agony. If they upgrade, they will have a better experience.
The powers that be in the Lab want legacy support for 10.5, if it is reasonably feasible. But, the Lab is saying the people responsible for making things work have NOT yet made a decision whether it is feasible. I suppose they have not yet come back with an estimated cost and time frame for management to base a decision.
Havok Licensing
The Lab is closer to having a solution or more accuracy may be: a license for a Linden made Havok library for third party viewers.
Now that Pathfinding is out effort is going into getting all of the Pathfinding tools working in TPV’s. The only thing the library is needed for is rendering the Navmesh. None of the other PF Tools need the library.
If I understand correctly, the library is needed to put an image of the Navmesh on screen and to be able to select parts of the Navmesh to change its properties. Most of us are not going to need to edit the Navmesh. Land owners will be the primary users of the feature. Initially all viewers that include the feature will likely be identical to the Lab’s viewer. Depending on how popular editing the Navmesh is, I suspect Navmesh editing will be a low priority feature for most TPV’s.
You can see the PF Tools in the Development Viewer and I think the Beta Viewer.
Click to Walk Pathfinding
Will the Lab integrate Click-to-Walk with Pathfinding?
The current thinking by the interface people is: bad idea. They feel it violates the rule of least surprise.
Because Pathfinding might return a circuitous path, out around things rather than in a straight line, it would likely be to counter-intuitive. If sidewalks were set as walk paths and grass was set as a slower or more expensive walk path then trying to cut across the lawn would create a path that follows the sidewalk. New users will likely be going WTF!?! I might be going WTF!
So, for now the answer is no planned support for Click-to-Walk Avatar Pathfinding.