I’m lagging getting this update posted. There have been other interesting things to do. Plus there is the mesh deformer news that ever so slowly trickled out. I splashed the update on the Deformer Project here in Mesh Deformer Update. Plus there is lots shaking in the background. Whatever, the latest on mesh follows.
CTS-694
CTS-694 – Missmatch of LOD meshes in multi object upload (linkset) yield viewer crash. This is supposedly fixed… sort of. I’m unclear.
A symptom of the problem is if you make a chair and table and export them as a single object. The Viewer’s uploader was mixing up the textures. It is… was a problem with how the viewer reads the Collada file and assigns the textures. If I understand correctly there is a fix so we can upload and fix textures after upload. I’m guessing a following fix will get the textures to match up with the model parts correctly.
SH-2500
SH-2500 – Second Life 3.0.6 (241379) “Calculate weights and fee” button in the mesh upload window gets greyed out when choosing custom physics shape from file. A resident asked if the Lab would provide a Hot Fix for this problem. The answer was no.
The item has been fixed. The fix along with others are winding their way through QA and will show up in a viewer soon. If you read my #SL Viewer Update Week 45 you understand what is going on. It is hard to describe the complexity and headaches from stepping outside the normal development channel to create a Hot Fix. Considering how quickly changes move through viewers now it just isn’t worth the trouble, IMO.
Decimator Options
You may have noticed that these options disappeared from the Mesh Upload dialog. In discussion about the removal attendees were asked if they used the feature. My answers was yes, but I gave up. When asked why I responded I thought I could do a better job in Blender. Runitai responded, “Well, there you go — they were removed because they weren’t very useful and cluttered up the UI quite a bit — if you want some power options on creating LoDs, it’s better to use a 3D modeling program.”
LoD Models Not Affecting Weight Count
What some noticed is placing models in the various LoD slots was not affecting the weight of the model. The speculation, since there were no examples provided, is that the differences between models were so small they were identical or were in fact identical models.
As the system is now using the same model in multiple slots forces the same model to be downloaded multiple times. One can get a better weight and reduce streaming/download weight by using the option to use the model above.
Another thing that can cause a weight oddity is large models may never use the medium and low LoD models. So, they do not affect weight.
Block Old Viewers
The question comes up often about whether the Lab will block old non-mesh capable viewers. The answer is always the same, no.
Mesh and OpenGL Fixes
The OpenGL fixes are in the QA pipeline. The official ETA is ‘as soon as they pass QA.’ My guess is made considering the rate things are going and with a surprise factor, so I’ll say 2 or 3 weeks is likely.
As to mesh, Vir Linden says, “SH-2500 is with several other fixes in mesh-development. Will be in either the next beta or the one after that, depending on how things go with QA.”
Viewer Performance
It really isn’t part of the Content Creation and Mesh upload group’s focus, but Runitia Linden is in the meetings so we get some viewer insights. One is that he has recently looked at some of the Flexible User Interface processing. I commented one of my problems with the new viewers is the time the viewer spends processing the user interface. (Ctrl-Shift-9 toggle) Some work has been done and the time consumed by the interface is decreasing, which gives better FPS rates.
Mesh Deformation Project
Lots of people are interested in this project. I certainly am. I’ve been sending IM’s and Feed posts trying to find out what is going on. Lots of people are talking about it. Karl Stiefvater (Qarl Fizz) was on Metareality talking about the project. (Thanks Voodoo Radek) Maxwell Graf, a motivator of the project, posted an update on the project (See: Mesh Deformer Update). Charlar Linden posted on the SL Forum that the Lab is willing to work with open source developers on the project.
Karl in the Metareality show says the Lindens have contacted him and want to help with the project. From that and some other things we can surmise there are some things happening. One is that work is moving forward on the private project. Next the Lindens are willing to work with them and are actually making an effort to do so.
From understanding Karl will be cooperating with the Lab we can surmise some things not immediately obvious. One is the time frame for completion will be extended. It takes time to talk and figure out things and then reach a consensus. The other side of that coin is the project is much more likely to be accepted by the Lab. So, we are going to see some delay, but the odds of acceptance increase.
Considering the odds of acceptance… the word is there are no guarantees, which is rational. The Lab is not going to commit to any code until they can see it and test it. As is pointed out in the Metareality podcast, …the Lab is not going to accept anything that is crap. Blunt but it gets the idea across.
We can speculate a bit on what the Lab’s concerns are in regard to the code. We can make intelligent guesses as to what the Lindens might have problems with. From their refusal to block older none mesh capable viewers we can guess that any code that is not backward compatible is not going to be acceptable. While the Lab will not update the older viewers they are being careful to avoid breaking them, which is good news for the fans of V1 viewers. To others using the latest and greatest, it is just frustrating. That’s life.
We can also guess from the new features coming to SL, like we see in the new Project Linden Realms, that new features we are unaware of are coming. If the Project’s code is incompatible with future features, it is likely to be rejected. This point should make it clear that communication between the developer and the Lab is required.
I can see where it is possible for some of the future feature stuff to lead to development disagreements. I suppose it will depend on how well they can communicate in the process. The Lab has future development plans they are thinking about and developing. They tend to keep them secret, as most businesses would. Secrecy could make it hard to explain some decisions to a developer. So, there could be trust issues and different opinions based on different knowledge. I get the impression that several of the Lindens want this to go forward. So, I expect most obstacles to be overcome.
Still, there is no guarantee this is going to work. There are several issues that could block the Lab’s acceptance of the Project. But, I’m optimistic it will work out.
/me winks.