As reported earlier this week, things are moving forward with server releases. There isn’t much new to report. See: Second Life News 2013-5 (1). Things have gone pretty much as expected.
The Lindens think the package running on the Le Tigre RC channel will get promoted to the main channel in week 6. This is the package with the fix for regions not showing up in search. Yawn.
The Multi-Threaded Region-Crossing code rolled to the main channel this week. That mostly seems to have worked.
Don’t expect to see much difference when you cross a region. I think my avatar responses to navigation instructions faster after a crossing. But, it is not significant enough a change that I can be sure.
For avatars with lots of attached scripts there should be a more noticeable change. My avatars script weight is almost always below 500k, usually 250k, and often under 50k. The max is considered 3mb. Over that limit and some regions will eject a person.
Simon Linden tells us the big visible change should be for the people in a busy region. They should see less impact on region performance when an avatar LEAVES the region. Time Dilation should perform much better. Those actually crossing out of the region won’t notice much change. Those staying behind should.
This change is unlikely to improve vehicle crossings. Simon thinks the changes only affects avatars.
There has been a problem with bots getting flooded with data. Those that use libOpenMV have the most problems. According to the Lindens some other users are seeing high bandwidth use too. High on the order of a continuous 100k/sec.
Andrew Linden told us that zero filled packets are having a problem and there is a packet size issue too. Fixes are in progress. We should see a fix in RC next week. The current RC’s are apparently still showing the problem.
Magnum does have a fix for the problem running this week. But, apparently it did not completely fix the problem.
Someone commented that diagonal regions are showing again. Diagonal regions are those that are SE, SW, NW, and NE of a region. For a time they were failing to rez in the RC’s.
The CHUI, Chat Hub User Interface, is moving closer to appearing in the Beta Viewer. I expect to see it appear in the Development Viewer in February and maybe even make it to the main viewer by the end of the month.
I thought only the user interface was changing. But, Oz Linden says, “It’s a big project. There are a bunch of new preferences, lots of code refactoring, and new functionality in input.” That makes wonder if the are preparing for more chat system changes in the future.
Those of us that have tried the new CHUI like it. Coming soon. Weeks.
Earlier I warned about the crash on ban and eject. Seems that bug is still with us.
Maestro Linden says that research into the problem has shown it existed even before last week’s updates.
In a region-crossing discussion the idea of using server time to measure crossing events came up. In general that is not going to work for millisecond timing issues. While the sim servers do time sync, they are not in perfect sync. They can apparently vary by several hundred milliseconds. I would guess maybe even 1 or 2 seconds and possibly more, but I have no evidence.
So, using scripts which query the server they are currently running in are not going to work for crossing tests, at least not all that well.
While the Materials System is moving forward and will soon be out, what we are going to be able to do with it has gotten a bit fuzzier. It seems there is some concern that being able to program scripts to change normal and specular maps is some kind of griefing opportunity.
We know that when a mesh changes it troughs a huge load on the server, physics engine, and viewer. So, the Lab does not allow mesh changes once an object is rezzed. There is concern that changing normal maps will do the same as mesh changes. I don’t see it. But, I maybe I’m missing something.
Texture animation does not affect the server. It is something that happens in the viewer. Changes have to be sent to all that can see when a texture changes or the animation changes attributes in some way. But, we use texture change to animate mesh faces now. (Thing anime girls big blinking eyes.) No big deal.
Changing the shape of an object affects the physics engine and requires updates be sent to all the viewers. There are many system impacts for a shape change and some are time expensive. I don’t see where changing a normal map would affect the physics engine. I think that normal maps only change the visual appearance, which happens in the viewer. But, maybe I don’t know something.
Whatever, the Lindens seem to think it is a problem of some kind. So, if Simon and Maestro are right, we won’t see scripting functions for changing normal and specular maps. Bummer.
We don’t have any information on how texture animation is going to work with materials. My current thinking is animation would affect all three maps.