Count me as another one of the people who thinks the current PE costs are a problem. Or, to be more specific, the PE costs are a problem when compared to other prim costs. Let me give an example. A few years ago, I made a little grove of three beech trees using megaprims and five sculpted prims. That's, per tree, 1 2/3 prims and - I'll round up - 3414 triangles. Tonight, I made a nearly identical mesh tree, a single copy of which weighs in at 800 triangles and 16 PE. If I decide to be scrupulously fair and link three of them together, the whole thing comes in at 47 PE, saving me the equivalent of one whole prim. I've gone from 2048 triangles per prim with the sculpties to a little over 17 triangles per prim with the mesh trees. Per triangle, the mesh trees cost over 120 times more than the sculpties. Now, I haven't done custom LODs for that tree yet, so maybe that's skewing things. As is, however, I view this as a broken system. In addition to making mesh less usable, with all the problems attendent in that (an uglier SL, frustrated creators, etc.), it sets up perverse incentives. To get back to my tree example, in terms of load on the viewer's computer, the mesh trees are over four times more efficient. The PE doesn't reflect this, however, and in fact heavily favors the sculpted trees. Even if I'm being an especially conscientous builder and decide to take the prim hit in exchange for a less laggy build, most people won't do that, either out of preference, necessity, or ignorance. A properly done mesh build will run much better on an older computer than an identical one made of sculpties, but how many of those are there going to be if mesh is so much more expensive than sculpties?