[Next Steps for Mesh Import]

One thing about standard prims that makes them worth learning about is that there is a wide range of scripted geometric operations one can do on them with llSet*PrimitiveParams*() that won't be available with Meshes.  Like sculpties, one will be able to swap in a wholly different Mesh (ignoring how painful it is to watch the new one download, without the kludgey workaround we used with sculpmaps pre-shown as transparent textures).  And of course a script can still can paint them and slide them around and scale them.  But we can't make them hollow, change their cuts, twists, tapers, etc., as we love to do with regular prims.

Early on, I'd fantasized that the fact Meshes could be deformed to the avatar skeleton to which they're rigged might mean they'd be deformable by script, too.  No such luck.

[Next Steps for Mesh Import]

but can we wait  to see how it works before calling the police? As a professional developer with almost twenty years of experience i'm carried to think that they experimented the feature and that the development was planned by professionals that do know what the issues can be. I can't understand why you think that they wake up one morning and say "Hey let's have fun today, let's put the mesh import in the platform!" (btw it could even be true reading some comments here). Sure not always the development of a highly structured feature works out of the box. But ...baby, it's programming

I'm not promoting "doom and gloom" here, I'm trying to see what's ahead and how to adapt to it.  I agree that LL has probably considered the survival of its own network and infrastructure (notwithstanding prior experiences to the contrary--e.g., the overselling of underpriced OpenSpace sims).  There are, however, some clear directions where I think residents should be focusing their attentions.

For example:

• Somebody in the market for human avatar skin or clothing right now should be aware that it will likely have only very temporary utility.  (See the post I linked, if the reason for that isn't obvious.)
• Folks in the human avatar texturing industries (skin and clothing) need to figure out what product and business model they plan to follow, come Mesh.
• Anybody re-upping their ISP contracts might want to reconsider bandwidth overage costs.
• Lindens (were any so foolish as to wade into this thread) might want to quickly hack the Mesh viewer to enable individually-settable avatar complexity limits.  (We know they're pondering what limits to place on avatar complexity overall--and there are good reasons for setting those overall limits far too high for many users on lower-bandwidth connections.)
• Builders may want to find SL-specific niche markets to target.
• People may even want to consider whether building and texturing are where they'll have the most fun in the future.  As opposed to scripting, say, or animation.

Or we can all just go back to oohing and ahhing over the pretty videos, congratulating LL on a nice addition to SL.  That's fine.  As with every technological advance, however, with Mesh we are pawns in a Red Queen's Race.

[Next Steps for Mesh Import]

It's fairly evident that no Lindens are reading any of this (not that they said they would), so we're all on our own here, and need to make the best of it.

I share Vivienne's and Prok's expectations about the flood of Mesh imports from external sources.  I don't quite share their views, however, on the economic impact of that flood.  That's probably because I'm more pessimistic about the current prim-and-sculpty economy.  As I see it, there's already no real money to be made in general interest prims and sculpties.  The flood of Meshes will just change the geometry of failure for anyone trying to make L$s off general interest builds.

By "general interest" I mean the sort of stuff already in great supply in SL, which is the exact same sort of stuff that's all over the net as 3D models.  What's not "general interest" in that sense is SL-specific "cults".  Not necessarily SL-unique; I'm thinking of Steampunk, for example.  Stuff that is not in every post-apocalyptic urban grunge game, nor every bright shiny futuristic setting. 

Steampunk is not my aesthetic, so for all I know there may exist tons of external Steampunk-specific models to munge into SL upload format, but my point is that only some such "cult" built content will have much of a chance to make L$s.  But this is nothing new.

I get it that there are folks eeking out a few L$s by selling some of the easy stuff, based on prims.  I don't know how they manage to do that, but I suspect they'll manage about the same with easy mesh-based stuff, too.  I really don't think building-for-modest-return is where Mesh will have massive economic impact.

I won't belabor it here; I've posted some in an earlier comment here and in more detail in another thread.  The big impact will be on human clothing and skin makers.  They are in for a wild ride.

All that is the SL-internal economics, but there's another RL impact ahead: bandwidth.  Mesh is many times as transport-intensive as a sculpty or a prim, per vertex.  On the plus side, efficient meshes can use many fewer vertices than either of those other representations for equivalent shape complexity, so downloading and rezzing a Mesh can be just as quick or even slightly quicker than a sculpt (and rendering FPS can be very much better, which is nice, but not what I'm talking about).

The problem is that folks are not looking to Mesh in order to make more efficient shapes of current complexity.  None of those YouTube vids are showing how fast they can render a sim full of plywood Mesh cubes.  Rather, they're showing how many layers of wedding cake ornamentation they can plaster on their builds--and everybody goes "ooo" and "ahh" right on cue, so that's what's ahead for the grid.

And because all that ornamentation demands even more bits downloaded, rezzing times are going to get worse, and users are going to see bandwidth caps kick in.

That would be true even for builds; the current prim equivalent costs are more about matching rendering complexity than they are about download requirements.  (Yes, they're based on filesize, but the prim-equivalence is several hundred times larger than the transmission size of the most complex prim.)

But builds are the tip of the iceberg, compared to avatars.  Although LL will be forced to limit the complexity of avatars, unless they give users some way to control the download complexity of the avatars in view, it will be possible to blow a big chunk of the typical ISP's monthly bandwidth cap just rezzing-in at a dance club filled with Mesh-attached avatars.

[Next Steps for Mesh Import]

Oh, thanks.  Hmmm.  I saw that, for mesh clothing.  Sort of a built-in "skirt sitter" or whatever those scripts are called.  Never considered it for hair.  Not immediately obvious to me what bones one would rig to that would have natural-looking results across a variety of poses -- but perhaps it would help if I didn't immediately think of bondage poses.

[Next Steps for Mesh Import]

There are non-viewer in-world editors for sculpts and especially animations--and there will be for mesh, too--although they're a pain to use because LL won't expose a scripting or messaging interface to the internal representations, so everything has to go through an external upload.  That's just dumb, especially for these 3D immersive things.

Just in passing: I use one of those non-viewer in-world animation tools regularly.  It would be hopelessly difficult to use for elaborate continuous motion animations, but I gotta tell you: when I need a vehicle pose, because I'm able to do it in-world, I'll have it uploaded and working correctly long before the requisite prop could even be loaded into Poser.

[Next Steps for Mesh Import]

Imagine hair that moves with your body instead of cutting through it with flexiprims for sweeping movements. 

Huh?   How does mesh matter for that?

[Next Steps for Mesh Import]

In fact, the LSL editor is an interesting analogy.  Those of us who fancy ourselves "serious" scripters don't use the in-world editor for most development.  It lacks the "pro" features of an integrated development environment, so instead we use something like eclipse with our favorite text editor customized for LSL, syntax checkers, preprocessors, and most important, version control.  Nonetheless, the in-world script editor is an essential part of our workflow (and not just for uploading scripts).

I see an in-world Mesh editor in much the same way: adequate for many users, and still essential for power users who do external development.

Really, though, I see building (including Mesh) different from scripting, and I think building benefits much more from an in-world tool.  I guess that's partially because building is such a viscerally rewarding 3D experience, in contrast to the much more abstract reinforcement from scripting.

I've said a lot about "social building" but if I be honest about it, I'm not always very good company when I'm deep into an in-world prim build.  Still, though, I'm connected, I sometimes entertain interruptions, and (not insignificantly) I have some excuse for owning land.

For me, at least, that will change with Mesh.  I can tell right now that I'll spend weeks at a stretch not even logging in to Second Life, while I'm obsessing over some Mesh builds.  Or, rather, I'll do that until I get bored with that solitary obsession, as I did with Sculpties -- and never touched another sculpted prim.  That time I returned to SL, but this time around, I can't be sure I won't drift off to some altogether different pursuit, once I'm over my Mesh masturbation phase.

That'll be fine for me.  Maybe by then an OpenSim will have Mesh and a TPV with in-world mesh building.  Or maybe I'll finally finish the Great American Novel, or take up ballroom dancing.  But if others have the same reaction, it's not the best outcome for Second Life.

[Next Steps for Mesh Import]

Many of the primcount-equivalent questions were answered at Nyx's office hour yesterday.  I took the liberty of posting a transcript under his User page on the wiki, and linking to it from the missing-in-action Mesh article.

(As I understand it, the size is determined by the "filesize" of the maximum complexity model or the physics cost--the formula for which is not yet determined, but yeah, based on the Collision model, which can be automatically generated.  I gather that the "filesize" is that of the internal format, not of the raw .dae, and the viewer does the conversion and very substantial compression.)

[Next Steps for Mesh Import]

The blank assertion that there could never be an in-world mesh editor because it's too difficult for LL to program is simply silly.  Do you really think basic operations on mesh are more complicated to implement than the existing build tool?  There are dozens of free and commercial mesh modelling packages--but LL's developers are too lame to build one, despite having a whole rendering, environment, and manipulation model pre-built.  Uh huh.

If LL exposed a messaging format for manipulating temp mesh assets and applying them to objects, multiple TPV implementations of such an editor would exist within a month.

Even if they were rudimentary at first, even if they had to export the model to get UV maps, even if they were little more than a glorified Blender backend, they would make Second Life a social building experience again.

And it would kill Blue Mars overnight.

(And if Blue Mars does it first -- and it's really not that difficult -- it's curtains for Second Life.)

Sure, there's a small percentage of SL residents who have their favorite external mesh editor and desperately want to use it.  Everyone in the closed beta, for example.  Fine, so there will be mesh import.  And if that's all we get, then we'll have a product that satisfies the closed beta testers and disappoints most everyone else.  (Sound familiar?)

---

And then there's the "mesh modeling is easier than sculpties, so everyone can do it" claim from some of the same folks who dismiss anybody who's not a professional 3D artist as being unable to handle in-world mesh building.  It's not an attractive pose, welcoming the masses with one hand while with the other patting oneself on the back for being such a l33t pro.

---

I'm still waiting for the skin and clothing makers to get what this means to their businesses.  Yes, there's a freely available mesh model corresponding to the existing human avatars that's compatible with the UV maps used for existing content.  Even if that were to remain the norm (it won't; see below), one cannot distribute skin and clothing assets for a Mesh avatar; one must instead either bake the whole outfit, skin and all, onto a fixed-shape avatar (works for that Agent Smith avatar, always much in demand ), or distribute the textures.  How many skin and clothing makers are ready for that?

Assuming creators can get past this hurdle of selling textures instead of specialized assets, there's a big market for sticking with the existing avatar UV map.  But we all know the existing UV map sucks.  It's obvious that there will instantly be a mesh with the two arms textured separately.  And then it's all over: the cat is out of the bag, and there will be dozens of competing human avatar meshes, each having more vertices than the next and all using every UV map to achieve the finest detail.

I'm actually fine with that.  But I don't sell skins.  I might feel very differently about it if I did and suddenly any one product could only be used by at most 10% of the human avatar market.

[Next Steps for Mesh Import]

I think there are a couple of major points being missed here.

Meshes have no scale limits.  Remember all the hand-wringing about letting the Build tool use numbers bigger than 10?  Worries about encroachment, etc.?  Somehow none of those problems matter with Meshes.  May as well have used the excuse that the Build tool only knows how to count with its fingers.  Remove restrictions on the prim editor before Meshes are allowed on the Main Grid!

Meshes will revolutionize clothing and skins more than building. (This is a bit hard to see in the videos, perhaps because they seem to be using the worst animations on the grid.)  The instant that fashionistas can have a human avatar mesh hundreds of times more detailed than the standard avatar, the clock starts ticking, counting the days until anybody with a standard avatar is immediately recognizable as a newbie.  So all that work that went into avatar alpha masks, multi-wearables, etc.?  Wasted.  And all the current texture (non-prim) clothing and skins?  Obsolete.

There will be future markets for clothing textures and skins, but they'll be fractured, painting their wares on competing avatar meshes.

For a time, existing animations will keep working, but the same obsolescence is certain, once the avatar skeleton can be defined.

On a separate note: I happen to agree (strongly) with those such as Vivienne Schell who believe that in-world Mesh building would preserve an important aspect of Second Life that otherwise will be lost with Mesh import.  I think this is a much more serious problem than concerns about needing more skill to make Meshes.  That's backwards anyway: It's horridly tedious work, but it requires very little ingenuity to make a very complex mesh; shaping and assembling prims to achieve a similar effect is much, much harder.

It's really not that the "less skilled" will be frozen out of the market because there's no in-world Mesh building.  Sure, for a brief time, the NDA-signers will enjoy a monopoly, but that will last days, at most.  The real problem is that anybody who wants to build (or clothe the new avatars) will spend all their time and attention outside SL from now on.  Not a good thing for concurrency.  Ultimately not a good thing for Second Life when both the "playing with blocks" and "dressing the dollies" crowds are both invited to stay disconnected from the grid to have their fun.

[Next Steps for Mesh Import]

These YouTube videos are all at least three and a half months old.

They're okay as kind of teasers, but if it was going to take this long for public release, I was expecting a lot more detail at the same time, such as documentation of avatar rigging and the scripting functions for manipulating textures, as well as the formula for counting mesh complexity as prims, and how that limits avatar-attached mesh complexity.

So, those are questions.

There's also the meta question of why these particular vids were under wraps for so long, and what's been happening since.

[Changes to Display Names Based on Your Feedback]

Jack, thanks for the updated information and the adjustment in defaults.  It may not completely remove everyone's concerns, but it makes a big difference.

FWIW, I don't see any significant downside of a nominal charge for changing display names, nor of color-coding display names that differ from usernames.  Probably don't want to charge for setting the first display name, however, because new residents are born $Linden-less.

[Second Life Economy Stable in Q2 2010]

Wait till they fully integrate Market Place into the viewer and fully shut down malls and in-world shopping.

You could see a serious decline in In-world transactions. Malls and shopping places will have a harder time and be unable to circulate money by hiring employees for their clubs/businesses

Agreed.  The continued rise in Xstreet/Marketplace transactions is very worrying, as a leading indicator of decreased land ownership.  The more sales made on the web, the less reason for merchants to pay lots of expensive tier.

If the overall economy grew as fast as the web merchant transactions, there'd be a different interpretation of these data.  But the economy is not growing by any measure -- most telling: the relatively flat amount of L$ in circulation accompanied by a slight decrease in L$ value.

That's okay; times are tough, etc.  But if web-based economic activity is growing and the overall economy isn't, then the land-based economy is shrinking.  Land is much more profitable for Linden Lab than Marketplace can ever be, so expect further belt-tightening on Battery Street if this trend continues.

Could SL survive if there were no in-world sales, no land-based stores, and instead all purchases were made on the web?  One could imagine a (very different) virtual world where nobody went in-world to shop.  A trend in that direction is a trend toward less user time spent in-world.  And, bots notwithstanding, that's another trend we see in the data.

A completely unrelated observation, but a good thing about the new metrics: The always misleading voice-usage statistics are now gone.  I take this as a hopeful sign that LL itself now has a more realistic understanding of that feature's actual usage.  That's not much of an economic indicator (the original plan to add a monthly charge to voice-enable a sim was abandoned long ago, and the penetration of voice morphing is presumably near zero), but it may be important for prioritizing investment in viewer development.