Teleport to, Swimming animation, cache permanent locations/characters, Importing 3d

But then he also goes on to show that the data needed to describe a prim sculpture is far greater than that of a mesh.

So mesh size+mesh texture is not greater than prim data+texture

I don't think a prim-packet actually contains 239 bytes worth of information. Any prim that doesn't have FLEXI is certainly not going to transmit 64 bytes of data that is all 0s.
Not to mention that "phantom" (and others) is not 4 bytes. It's 1 bit. On or Off.

Well I'm not going to beat a dead horse. My opinion is that a simple 10x10 mesh won't impose that much extra bandwidth for the benefit that it would bring.

I dont think this is all that unreasonable.

If meshs are equal or less strain on the grid than prim based objects - then the Lindens should eventually come up with a plan to incorporate.

Since currently prims = Land = Money

they will need some way to convert

mesh with this many parts = Prims = land = money.


Im no 3d modeler
But the logic seems pretty basic.
When you search the 3d sites for objects and they come in X form. Then being able to use X form will vastly improve Second Life over only being able to use Y form (prims)

I visited www.Turbosquid.com and typed 'sofa' into their search bar and was presented with stuff of this quality:

http://www.turbosquid.com/FullPreview/Index.cfm/ID/215814


"bed":

http://www.turbosquid.com/FullPreview/Index.cfm/ID/234631

Have you check the polygon number of the bed? It's 40,900! That's certainly not something to be transmit and show real time with current standard connection speed.

...even if it does look more compy than Sl box-bed ;p

10x10 mesh = 100 verts.
100 verts = almost nothing.
Literally.
A cube my be 8, but a sphere doesn't look good until 16 divisions (even better at 32) or 256 verts.
Now, if we had NURBS we could make a BEAUTIFUL sphere with 12 "verts" (4 control poins per spline, 3 circular splines with the surface mapped across them). Theoreticaly you could do it with 8, but I don't have Maya running at the moment so I can't test the idea.
Disadvantage to NURBS: no hard corners. Takes 24 conrol points to make a decent cube.

We don't want to make spheres or cubes. This is the whole point of wanting mesh imports.

Yes, a mesh sphere takes up more data then a prim sphere. But with more complex objects, the difference in data size between a mesh object and a prim object becomes far less.


But don't take my word for it. Go look here: www.activeworlds.com

Active Worlds were using meshes before SL even existed. God knows how though! Because it's just not technically possible, right? They must use magic!

I'm going to go and make that bed out of 1256 prims and move next door to you.

And the same number of nurbs control points generates even better looking stuff than mesh verts do. *gasp*
As for ActiveWorlds, I couldn't care less that they've been doing it. The problem here is figuring out how many verts or polys is "1 prim" worth of usage.
Also an issue is the texturing of a mesh. Spherical, cylindrical, planar, or UV map?
Being larger or smaller than a texture of X size is pointless, because as we all know textures are taking forever to download (so long in fact, that they are starting to impact how long it takes me to download prims). Also, your point about it being smaller than prim data:
How much of that prim data will beed to be applied the mesh as well? All of it.

We're not asking for nurbs. Nurbs is asking too much of LL.




Well that might be your problem. It's not mine. I'm not interested in making an house out of a cube.

Then what are you planning to make? If it's a permanent structure of anykind "X#verts = Y#prims" is an issue.

And you didn't even address the point about mesh object data required that exists in prim data currently that's going to be "subverted" into verts according to your bandwith saving plan.

An issue for you maybe. Not for me. I don't mind waiting a little longer for 3D data to load. Streaming 3D obviously doesn't work. It's clearly never going to stream as fast as an avatar moves. So let's just forget about all this streaming nonsense.

Baka!
I'm talking about THE LAND USAGE!
NOT whether or not X#verts = Y#prims for SPEED OF TRANSFER.
How many mesh verts can I use on my land is an issue.

I've done some calculations and you personally are allowed just one.

One. One...One what, mesh? (How many verts?) One vert = one prim?

Don't you "one one one" me!!

You shall be allowed one vertex. Everybody else is allowed 1000's.

So there!

It's your feature suggestion.
And it ain't going to fly unless you can describe with exacting detail how its supposed to benifit the grid. Residents know how it will effect building, but the Lindens want to know about bandwith and limiting the size based on owned land.

If you don't want to, then, well *shrugs* guess it'll never come to light.

Okay guys, this argument is senseless. LL has already stated they will eventually allow full mesh building in SL WHEN THE TECHNOLOGY IS READY. Obviously that mean it is not ready yet. And so nothing anyone can do but wait.

Vertex limitation on land WILL be a problem, but that's gonna be LL's problem. Is unlikely anything we argue here is gonna be taken too seriously by Linden.

Well yeah, I knew this!.

But what's puzzling is that technology has been ready for the past twenty years. So people can't understand why it's not been added. This is why we keep seeing the feature suggested over and over and the same arguments over and over.

LL aren't stupid. They know what is needed. They aren't sitting around in the office, short of features to add. They're ultimately going to add whatever features they feel are beneficial to them as a company. Voice Chat shows this. Nobody has really asked for it. But LL feel Voice Chat was necessary to keep up with competition and that it would be simple to implement.

LL seem to have a love of anything that can be simply bolted onto the client that doesn't involve having to change the underlying code too much. I expect to see a spreadsheet added any day now.

This 'Feature Suggestion' forum isn't here so that Linden Lab can come to read your suggestions. This forum was really intended to be used for filtering out the annoying suggestions from the 'Linden Answers' and 'Resident Answers' forum. They really should put a sticky at the top of this forum that notifies people that LL doesn't need their suggestions and that the posters are ultimately just wasting their time. But I suppose that doesn't sound very nice. So it's not gonna happen. Instead, we get the bullshit that's there now.

Although trying to be a little more positive; Some people may enjoy discussing possible features. Regardless of whether they're going to be added or not.

Well, probably the same reason SL is still using Havok 1 even when Havok 3 has been out for ages, I guess

I think Nargus has is about right with the comparison to the physics engine (although, when I checked last year around this time, Havok 4 was available).

Just because the technology exists doesn't mean it's compatible. Yes, there are mesh rendering programs and ways to transmit mesh data over the internet, but that doesn't mean that the physics engine, the client rendering code, or the servers are ready for "mesh prims."