Daily Archives: December 21, 2011

Mesh, maths and the mainstream

Nalates Urriah carries interesting news from Charlar Linden by way of the Mesh User Group, in that:

  • Some 55% of SL users are running Viewers capable of rendering mesh
  • Some 18% of regions on the main grid now have mesh objects rezzed in them.

Nalates speculates the data may have been gathered before Phoenix gained its mesh capability. If this is right, then it would put those using mesh-capable viewers in the majority for somewhat longer than has been thought to be the case (given mesh capability has been growing across both V3 and V1 based viewers since September).

However, it is the second figure that attracted my attention. On the one hand, it suggests mesh is gaining traction, but on the other, it hasn’t been the hit Linden Lab may have been anticipating.

It is certainly true that overall take-up has been slow in terms of mesh objects becoming more and more apparent on the grid (although telling them apart from anything else without detailed inspection is admittedly becoming harder and harder with the rise of mesh-capable Viewers).

One of the reasons for this is likely to be the fact that, like it or not, there are technical issues surrounding mesh as currently implemented. Some of these are doubtless down to the complexities of having to shoehorn mesh rendering and support into SL’s existing architecture without actually breaking bits of it completely. However, the fact that there are issues has tended to lead to proclamations that mesh is “impractical” on a broad range of uses where this is actually far from the case.

Mesh palm: 9 prims

One example of this is with trees and plants. Back when mesh was still in Beta, a content creator posted an article about being unable to create a tree in mesh without the PE (Land Impact as it is now) being in the 500-600 prim range. This was seized upon by some  – and is still cited today – as “proof” that mesh doesn’t work. The fact that there are examples of trees with a PE/LI down in single figures is simply brushed aside; I’ve actually had a post where I pointed this out dismissed as “politically motivated”!

Similarly, people insist that houses cannot be made in mesh without incurring massive Land Impact values. This again isn’t strictly accurate. Sure, there are houses that incorporate mesh  that hit a Land Impact of 350-400 – but then, there are also prim builds of equitable size, and intended for the same parcel size, that also hit the same LI values.

And, like it or not, mesh can compete with builds of more modest prim footprints. Take the example below; it has an overall land Impact of 91 fully furnished. The house structure itself, textured beautifully, has a LI of just 29.

Partial mesh house by Novocaine Islay, as seen through a mesh-capable Viewer (l) and non-mesh Viewer (r) to demonstrate it is mesh. total Land Impact: 91 prims with furnishings

Elsewhere, mesh clearly does bring with it benefits that are obvious to grasp. Furniture is one such area, where prim counts can easily equal or better than prim or prim/sculptie equivalents and give a greater level of detail, even allowing for the complexities of including scripts and animations – as the two examples below pulled from the Marketplace demonstrate.

Tiki Tattoo Rotan sofa (l): seats 5; 59 animations 7 prims with shadow effect; Rustica Gothic chair (r) 11 animations, shadow effect, 2 prims

However, the flipside of this is that, in fairness, there are areas where mesh is an unsuitable option, simply because it does incur higher “prim counts”, or can have unforeseen consequences. Ergo, its use is going to remain limited.

Take the aforementioned tree; it may well be possible to develop a beautifully detailed palm tree for just 9 prims. It’s quite a modest count – but it’s still a poor proposition when you can get three sculpted palms trees with shadows for just two prims; and people really aren’t going to be bothered about the underlying resource cost of the latter, because it’s not something that can be readily seen.

It is also somewhat impossible to stretch and resize mesh objects without ending up with some alarming results on the Land Impact scale. Again, this tends to rule out mesh for applications where resizing is required – such as the supply of in-world building component sets.

So yes, there is a case to answer for mesh at times being “impractical”. But this has perhaps been over-emphasised in some respects, and may well be a contributing factor – albeit a small one – for the slow take-up of mesh.

However, there is a much more direct reason as to why mesh would appear to be struggling to gain a more secure foothold in SL. It is this: creating mesh can be a nagging discomfort in the region of one’s posterior.

It’s Not Easy Being Mesh

For many well-versed in in-world content creation, the making mesh objects feels like something approaching alchemy. There are new tools to learn – Blender, Wings 3D, Maya or whatever you opt to go for. People who are perfectly conversant with gluing prims together can find these pretty mind-boggling to understand (as I know from experience).

Then there is the whole process of optimising a creation for in-world use and getting it uploaded. A creator-friend of mine who is working with mesh described this as akin to a visit to the dentist: it has to be done, but it’s generally far from pleasurable.

So for many, the creation of mesh objects isn’t the same as creating with prims; the fun element is removed, and the whole thing just becomes hard work to the point where it may well be felt that it simply isn’t worth the hassle.

It also can’t be denied that the tool set is lacking, and this doesn’t help. Most of us are by now familiar with the concept of a parametric deformer, which should help with the matter of mesh clothing. However, such a tool could have had far wider implications for mesh had one been implemented as a part of the overall mesh roll-out – such as overcoming resizing issues, as Drongle McMahon points out in the mesh section of the technology forum. 

Obviously, as I mentioned above, some of the issues around mesh creation are most likely imposed upon us by SL’s existing architecture. As such, it’s actually unfair to blame LL where this is the case. However, it’s probably equally fair to say that other issues, such as failing to include deformer tools, are down to management decisions which have potentially damaged the take-up of mesh and for which LL must shoulder the blame.

Perhaps the Real Question is, “What is meant by 18%?”

However, I actually have a more fundamental issue with the 18% figure. It’s simply so nebulous, it’s meaningless.

Great, so “18% of all regions have mesh objects rezzed in them”. But, um, what exactly does this mean? How is it qualified? Were I, for example, to be the first person to rez a single mesh object – say an armchair – in my home, would that tip my home sim into the “18% of all regions”? If so, then the value of this figure is drastically diluted.

Also, how does clothing and avatars meshes and avatar accessories (hair, attachments, etc) factor into this? These would appear to be by far and away the most popular mesh items available on the SL Marketplace, yet Charlar’s figure would suggest they are  excluded from any calculations, as they aren’t technically “rezzed on” a region, but rather “worn” by the avatar in question. So it is possible that in practical terms, mesh take-up may actually be greater than appears to be the case.

Thus, on these figures alone, it’s hard to define mesh as a “success” – or even, allowing for the questions around the 18% figure – “mainstream” – and please note I’m deliberately ignoring the comment on mesh sales Nalates also lists. This is again a generic statement; that “sales have doubled” is meaningless without some figures behind it.

However, it would equally be wrong to dismiss mesh as a “failure” on the basis of these figures, simply because they are so very nebulous and vague – at least where percentages of regions and doubling of sales are concerned.

All that really can be determined is that mesh continues to exist in something of a limbo, together with a speculation that it is possibly doing less well than LL had been hoping. Certainly, it’ll be interesting to see how the figures look and another two or three months’ time – and how they are presented.