The latest mesh development viewer has support for Collada 1.4.1. Download Page. Most problems are fixed. There are still some reporting problems with Blender rigged mesh.
Auto Return ADITI Mesh Sandboxes
Auto-return got turned off in the ADITI mesh sandboxes. That is going to be corrected. So, the area will start clearing again, probably by the time you read this.
As OpenGL, I want applications to not use the fixed function, because it is deprecated. – SL is using OpenGL Version: 3.3.0. There are problems with the viewer and OpenGL. Fixing the issues is a high Linden priority. It is still in process.
For those getting out of memory errors, try turning off private memory pool management, which I don’t know if it is a compile setting or a Debug Setting (MemoryPrivatePoolEnabled). See: VWR-26864
Mesh Ball Render in V1
People are noticing that mesh attachments render as different size balls in version 1.x Viewers. In some cases they are small and some they are large. Runitai Linden says the size of the ball is dependent on the scale of the mesh. So, a worn mesh scaled up twice its import size will be 2m in diameter.
This is counter intuitive. If you attach a rigged mesh, it snaps to the avatar skeleton. So, scale seem irrelevant. The snap to the skeleton doesn’t affect the scale. If you rez it in-world and scale it down or up, that will change the size of the ball seen in V1 viewers.
So, a 1m scaled mesh will be a 1m ball in old viewers.
I was using the non-mesh Firestorm version to attend this meeting. My SLViewer 3.0.6 had crashed when I tried to login for an earlier meeting. Since they are back to back, I stayed in a working viewer for the Mesh Meeting. It proved handy and interesting. We were able to easily check the scaling problem with mesh and the size of the ball seen in V1 viewers.
UPDATE: There is a good post on the Forum by Ashasekayi Ra showing this problem. See: SL Mesh Clothing/A
It was also interesting to see that most of the people at the meeting are using the SLV2 viewer. I suspect FS cannot tell the difference between V2 and V3 viewers. Not is it important. I think one other person besides myself was in a TPV.
All mesh viewers are using SSE2. You can check if your computer has it, it does unless it is 10 or more years old. Get CPU-Z and test your CPU.
It was recommended that LL include code in the viewer for non-SSE2 and SSE2 CPU’s. It is possible to write code for both. But, someone has to write the additional code. At run time the program can decide which code to use.
The problems with adding non-SSE2 support are bulking up code and time spent supporting a minority holds back progress for the majority. So, the Lab won’t be providing non-SSE2 viewers or upgrades to the older V1’s. That is left to the third party viewer developers. I think that is a pragmatic decision.
In the latest Phoenix Hour Jessica discusses the same problem and that team’s decisions regarding limited resources. They will be ending non-SSE2 support.
Summing It Up
Things are a bit boring development-wise as we move through a quieter maintenance cycle. Fewer things are happening with new features as recent features are tweaked, crash issues are fixed, and performance is tuned.
We know from announcements and leaks that other Linden products are coming to market. The rumor of a 2D SL-like creative product coming is out there, which may explain the disappearance from User Groups of some of the Lindens and the idea that User Groups should be self sustaining by resident support. It is obvious things are changing at Linden Research, Inc.
There is a JIRA issue for getting mesh clothes to fit avatars parametric shape. (SH-2374) It seems that JIRA item’s status has been down graded to: Someday / Maybe. Many are disappointed and somewhat upset over that down grade and lack of explanation.
We don’t know why it was down graded. Usually this means they have decided it is a complicated fix or other issues are more important. Those that think the issue is important usually are affected by it. So, their position is somewhat understandable. But, this is a usability issue, not a crash issue. We have several crash issues that are show stoppers for a large number of residents. Many of us would like to see those fixed first.
If you have read my review of the last Phoenix Hour, you know that release of the next version of the Firestorm Viewer is being held back by some of the GPU crash issues. Whether mesh fits as we want or not is much less of a problem than not being able to login or having your viewer crash.
Agreed – the dissappearance of the Viewer Evolution User Group has not gone unnoticed or unchallenged. What is surprising is that LL refuse to answer any questions on the matter which suggests to me that some other changes are imminent.
Nobody is arguing that crashed would be a priority, certainly. Having SH-2374 downgraded to someday/maybe status is a death knell for it, though. Thats the file 13 of the jira. And its not just about whether clothing fits you or not, for the record. There are about 40 things that would be solved by introduction of a parametric deformer.
Take a look at them here: http://images.plurk.com/2767af6225257672540d14ef94a30a43.jpg
Im really so pissed right now. You should be too.
I am disappointed. But, today we do not know what they are doing or have decided. That means reactions are emotional as they can’t really be rational without facts to base a position on. So, in a few days, after I’ve heard some facts, I may be upset. Or, I may agree with the Lindens. For now, this just does not look good. I’ve miss evaluated things before. So, I’ll wait to learn more.
I’ve seen the graphic and previously included it in post. They are good points and worth considering. However, parametric deformers may not solve all our problems. If you think so, provide the Firestorm team or another TPV team the code to show us. For now I’ll remain disappointed and undecided.