Mesh, Phoenix and the future

Jessica Lyon has issued a statement on the Phoenix website concerning the future of the Viewer. It makes interesting and clear reading.

On Mesh

A release of Phoenix will be forthcoming that can render mesh objects in-world. Jessica makes it absolutely clear that credit for this largely goes to Henri Beauchamp and his work in backporting the Viewer 2.x/3.x rendering code into Cool Viewer. She also makes it clear that the Phoenix implementation pretty much is Henri’s code as is. In the post, Jessica states:

“I don’t want you all thinking we’ve changed our focus back on phoenix, truth is we haven’t. Ansariel handled all the work of pulling Henri’s work into phoenix, LGG has helped. Aside from Tonya and Tech fixing some of the bugs.. That’s it.. essentially we’ve only had two developers working on this, and there are no plans to increase development on phoenix beyond that. However, mesh in phoenix will accomplish two things. It will complete [the] adoption of mesh in SL, which is pretty cool actually. But equally important, it will also fulfill our promise to keep phoenix going until it’s dying day.”

She also gives a stark warning on Phoenix with mesh rendering:

“Speaking of QA, don’t expect phoenix to be just like the last release only now it has mesh support. This work effectively makes Phoenix a Ford Pinto with a diesel engine from a school bus duct taped into it. Not only will it have all the existing mesh related bugs, but it will have plenty of its own bugs specific to having a diesel engine in a Ford Pinto. It will have a negative effect on crash rates no doubt, will be a performance drop for some, an increase for others. It will not be perfect, as it is not designed to support mesh.”

On RLVa

Jessica has also indicated that the next Phoenix release will have an update to RLVa as a result of Kitty Barnett’s hard work as well. Details aren’t clear, but one assumes this will bring it into line with the updated RLVa seen in Firestom.

On the Future

Jessica is unequivocal as to the future however: Viewer 1.x is, in her opinion, on its deathbed where SL is concerned, and as such, trying to maintain the Phoenix code on a par with Viewer 3 is going to be too much of a headache. As she points out:

“Consider this.. it took over 9 months to get mesh to work in a v1 viewer.. it took us just over 2 weeks to merge mesh into Firestorm once we started the merge. This will be the pattern with all new things LL releases, making it work in Firestorm or a v2 based viewer will be far easier to adopt faster than making it work on a v1. Maintaining v1 long-term is just not being realistic.”

No date is given for a release of Phoenix with mesh rendering capabilities, other than it will be released once it has passed QA.

The release is also liable to mark the end of the road for Phoenix where non-SSE2 capable computers are concerned. The release for such machines will not include mesh rendering support.

Overall, this news is liable to be met with approval from Phoenix users not yet ready to make the jump to Firestorm and might, conceivably ease some of the pressure on the Firestorm team to get some of the current bugs and issues with the latest Beta release ironed out.

And on the subject of Firestorm, Jessica did offer a small tease: “Mesh upload capability is also under development and making some promising advancements thanks to Nicky Dasmijn.”

Read the full blog post.