Qarl Linden’s departure from the Lab appears to be more closely linked to an overall pattern of ongoing layoffs than to cancellation of the mesh project.
Last week, I contacted a Linden Lab spokesperson who today responded on the status of the Mesh project:
The mesh upload project has not been cancelled, but we do not yet have a timeline for implementation that we can share. As Philip remarked at the inworld meeting, we’re proceeding carefully to be sure there’s a way to implement mesh upload with a positive impact on the inworld economy and without a detrimental impact on rez time and frame rate.
Now for a quick lesson in media literacy.
“Not cancelled” almost always means “Not cancelled”. It’s true that very occasionally it actually does mean “cancelled”.
“Do not yet have a timeline for implementation” has a couple meanings, ranging from “Surprise! It’ll be tomorrow!” to “We’ve stopped work on this and we’re still undecided as to whether we’re going to allocate resources to it in future.” The most common meanings are “We don’t know” or “We don’t want to tell you”, followed by “I’m not sure I understand what the engineers just said.”
My take:
Mesh isn’t, and never was a “done deal”. In order to get to the point where the Lab can decide if it is fit and appropriate to deploy, it has to be brought to a certain stage (ie: a closed beta). That’s probably the most responsible way of doing things, because you’ve got an opportunity to see how things go before you enhance/ruin (pick one) the grid at large with a new feature.
The alternative would be to dump it out there and let the chips fall where they may. I think we’ve all been through a bit much of that already.
So, Mesh features may or may not go ahead, but firing Qarl doesn’t actually seem to have been related to the status of that project, or be a factor in whether or not it might continue.
That’s my take on it, anyway. You’ve doubtless got your own.
[...] This post was mentioned on Twitter by Tateru Nino, slfeed.net. slfeed.net said: New POST in 20 min 『World of SL』 Tateru Nino: Linden Lab says mesh not cancelled 11:34 am http://bit.ly/8ZeUHf #slfeed [...]
My question has been less about the timeline for implementation of mesh, and more about the ability of the current platform to render meshes without suffering a nervous breakdown leading to catatonia.
I think that’s been the big question from day one. Implementation of freeform meshes into the existing platform has a whole lot of potential ramifications, many of which can’t be properly explored until you’ve got the basic implementation up and running.
You’ll note how the Lab refers to it as “Mesh upload” — the implication is that “Mesh support” is now embedded in the server code, and that “Mesh rendering” has been implemented for the viewer. Whether that actually gets plugged in for general use or not seems to be the basic question.
Mesh stuff IS in closed-beta, and has been for almost six months now. It is a working feature that just needs bug-testing and polish. Whether it will see the light of day is another matter.
Indeed it is. The NDA was lifted, and Linden Lab indicated (when I asked) that the NDA on the mesh beta was not subsequently reinstated.
(Oh, for those of you wondering about a previous feature: Expressive puppeteering. That was officially cancelled in 2008, and unofficially cancelled in 2007)
As the NDA was lifted I can give some impressions of how mesh worked in my testing.
As everyone else was trying to figure out how meshes worked efficiently… I figured I’d try doing it inefficiently and tested on some banged together models: basically I sketched and painted some stuff in 3d-coat in a very unoptimized way to see what it cost me. (Think some kid with a rip of mudbox just sculpting, painting and exporting without thought to optimization.)
Mesh accounting, when I tested it, counted my models about 40 prims or so. In terms of vertexes (and download) this was quite a bit more conservative than sculpties. So beta mesh accounting punishes unoptimized models.
If it punishes them enough I’m not quite sure. You will get more bang for buck from meshes than sculpties if you use them well, but anyone who is advanced enough to optimize sculpties properly will have more skills for optimizing meshes. Complex meshes will still lag, and do so in an ugly manner as the complete mesh needs to load, rather than a sculptie blob.
If anything meshes should impact lag/loading less than sculpts.
To be honest, meshes need to be worked out in the SL economy. Folk need to see how they work, how they can be abused, and the cost ajusted to that abuse, before meshes can happily coexist with prims and sculpts.
This is the same issue that happened with the introduction of sculpts so should be nothing new to the Lab. It just remains to see how they handle it from a policy level.
“The mesh upload project has not been cancelled, but we do not yet have a timeline for implementation that we can share”
My interpretation of that is what I wrote in my post The Calculus of Mesh: “I’d conclude that the calculus of mesh says that over a short term, the changes required to integrate mesh rapidly approach zero. Where, when and even if, the curve shifts to positive is a matter for the market to determine.”
It really doesn’t matter if mesh is done and works flawlessly. I think what matters now is the optics of the situation. There are very real universal crises at the moment – search, viewer 2 and overall lag – to name a few. There has to be some positive movement on at least one of these before anything new to the mix and I don’t think any of those are a slam dunk to fix. Search requires a combination of technical and social change, Viewer 2 .. well, moving on … and thank goodness for Imprudence because otherwise the experience on the other standard viewers is abysmal.
For the Lab to release a shiny new feature like mesh that a few residents can take advantage of would be perilous at best.
And I do remember when Avatar Puppeteering was “not cancelled” as well.
In theory meshes should improve performance roughly 50%. The reason is prims have all their sides whether you see them or not. A mesh model can delete the polygons on sides you cannot see, thus reduce total work for the graphics. For example: A section of wall made out of a box prim has 6 sides. One made of mesh can use 2 sides (inside and outside), and delete the ones on the edges where it joins other sections.
In practice, people will use the more efficient models to make nicer looking things. I am disappointed that the SL mesh will not support normal, specular, or other maps. That means SL will still not look as good as places like Blue Mars, that has more modern graphics support.
What this means is “We aren’t going to implement meshes until/unless Stroker’s Eros3d world really takes off.”
Last time I talked to Stroker, he was in nirvana over “dude, 100,000 vertexes in the LABIA!!!”
Actually that “Do not yet have a timeline for implementation” phrase is a stock phrase that the Lab gives me about every feature or bug-fix that I ask about. In some cases, the implementation appeared about three hours after I was told that the the Lab did “not yet have a timeline for implementation.”
So, trying to read anything positive or negative into that particular turn of phrase is fruitless. It may as well not be said, and is probably best ignored.
Re DanielRavenNest;
Normal, specular, other maps are not part of the mesh project. They’re part of the materials project, which wasn’t going to be put into closed beta until mesh project was available to the public. This makes sense as material support requires larger changes to the render pipeline and client. From what I’ve seen of mesh it should be fairly simple from a UI perspective to allow for multiple maps. They’re just not rolling it all out at once.
Well, I remain skeptic
which is rather unusual for me. But… let’s see.
Well, for those with an itch to play with meshes (along with a shared VNC viewer, a shared SVG whiteboard, media-on-a-thing-that-isn’t-a-prim, shared XWindows applications, and immersive audio that’s essentially voice chat with the addition of dial-in telephony support with the VoIP PBX of your choice) there’s always
http://openwonderland.org
…a 100% open-source Java virtual world toolkit.
Note that I said “toolkit”. There’s potentially quite a bit of do-it-yourself in OWL, the upside being that it’s actually possible *to* do-it-yourself, assuming you know *how* to do-it yourself.
This include making meshes (in Google SketchUp, as well as many other tools) that will kill your graphics performance quite dead….nothing will stop you, but then nobody will charge you for it either.
Grab or make a model in SketchUp, export a .kmz file, and drag-n-drop it directly into your OWL viewer.
Fun stuff…rather high geek factor required to use to its fullest.
[...] otherwise) as one of the “prim movers” of Mesh, was recently fired. His words support statements issued by LL earlier in the month that Mesh has not been cancelled. However, the timeline certainly now seems to [...]
lol, mesh not cancelled and neither are Ener’s plans on winning the lottery – wonder which will happen first =)
A new mesh beta viewer came out a couple days ago. It’d be surprising if they can’t get a public beta ready before the end of the year.
*holds breath* =D