Right. The big question few in the NFT community seems to ask is, who runs the asset servers? The “blockchain” part just gives you a claim ticket. You have to go somewhere else to get the actual item.

So where is that? Who runs that? Can they take your thing away from you? What's the relationship between virtual world operators and asset server operators?

Decentraland is very centralized. They control the asset servers and the marketplace very tightly. It costs over US$1000 to list an item right now. There's no way provided to export an item from their asset servers for use in some other virtual world. So you don't really own the thing.

Some others are worse. See

https://blog.portion.io/the-pitfalls-of-centralized-nft-platforms/​

(which is really an ad, but whatever).

There's a lot of talk about storing stuff in the “Interplanetary File System”, which is sort of like a cryptocurrency pay version of BitTorrent. That's protected against tampering, but not against loss. And no response time or bandwidth guarantees. It's often just someone reselling AWS at a markup, anyway.

@Nagle, That picture is hilarious and accurate.

Anyone who's read a book on solidity knows that the payload for most NFTs are stored on external servers (controlled by someone…not decentralized at all). So what you “own” is a uint256 in your wallet with just an URL pointing to the actual resource on someone else's server.

I mean if crypto kitties affected Ethereum, imagine if all that NFT asset data was stored on chain. It would crush the blockchain as it's duplicated across all nodes. IPFS is the proposed solution. I think decentraland had it in their whitepaper, but for some reason still haven't implemented it.

Just from an engineering standpoint (not from like a “youtuber Moon” standpoint). It's not an improvement.. Traditionally you would hold those items in cheap fast storage in a database with no fees to transact.

But now people want to “own” those items. So you're like “ok cool I will put them on chain". And you do and that. Now that they're there, what they technically “own” is a token with a link pointing back to your ORIGINAL OLD database. Except they will incur a $100 fee and 10 minute wait time to transact in them now. It's not a good time for engineering logic like this. Mostly people today just want new trading vehicles and NFTs fit the bill. It doesn't matter whether it's “true ownership” or any real technical improvement to the game. To many the game is the trading, not whatever is in the actual game.

For us, the metaverse is the primary focus. NFT/Crypto is secondary. We'll include it but it's only as important as any other means of transaction (banks/cc etc.).

From our experience it seems that projects prioritizing crypto/nft end up building big trading platforms attached to a bad metaverse. (at least so far). That's not the result we want.

It's a real problem.

For game devs, it's worth thinking about standards in three areas:

- Portals. You should be able to move your avatar through a portal from system A to system B, and it should Just Work, assuming B will let you in. That's what the author is talking about here. This already works for WebVR, which runs in a browser. Systems not resident inside a web browser need to do more work to do this, but it should be made to work. This is what makes multiple virtual worlds a metaverse.

- Assets. You should be able to take your purchased asset, if you really own it, from system A to system B, assuming B will let it in. The NFT crowd does not have this. Try to export an NFT from Decentraland. In EU countries, article 20 of the GDPR, “data portability”, may already require this.

- Money. When in system A, you should be able to buy from anyone who wants to sell to you, and not be required to use system A's currency or payment system. This is what Apple and Epic are currently litigating.

Those are the main connections we need between virtual world systems. Content portability at the creator level (GLTF, USD, etc.) is also useful, but the user interface for creators doesn't have to be as seamless. It might involve file dialog boxes, for example.

If it works and some systems offer those features, users will clamor for interconnection. (And third parties will somehow bolt it on whether the system operators like it or not.)

Great input thanks. A really decentralized system would require each party to have all the data which is synchronized all the time by peer to peer data connections. Utopia I think. So we need to rely on centralized servers, databases and authorities in some way. It would just be great, if users could take their digital assets into any game or metaverse. A universal interface / standardization as a industry standard like there are many around the world. “Global Digital Asset Exchange Protocol” or something like that…

doemel said:
A really decentralized system would require each party to have all the data which is synchronized all the time by peer to peer data connections. Utopia I think. So we need to rely on centralized servers, databases and authorities in some way.

If you go from one world to another via portals, you don't need synchronization that tight.

If you really want it, though, Open Simulator is able to do that - have regions under different management share a land border.

That's like saying the entire Web has to be hosted by Amazon for links to work.

Nagle said:
Content portability at the creator level (GLTF, USD, etc.)

Asset portability is fairly straightforward, but the really fun one is “concept portability”. How do I bring my flaming sword along with if the world I've portal'ed into doesn't have the concepts “sword” and “fire”?

Interesting article about fb visions and plans:

https://www.cnbc.com/amp/2021/10/30/facebooks-meta-mission-was-laid-out-in-a-2018-paper-on-the-metaverse.html​

Seems they'll pump a lot of money and seek for collaboration, after realizing they can't build a giant Titanic alone. Opportunity for Meatverse devs… ; )

One problem i see with asset portability is the danger to get stuck at outdated / inefficient standards.
If we settle on current state of the art, we probably get something like skinned meshes with skeletons and discrete LODs, plus some PBS standard materials and texture conventions. That's a problem on all ends:

Triangles may at some point turn out inefficient. Point clouds, voxels, or whatever else may become better if detail levels increase further.

Matrix or quaternion skinning can not model accurate anatomy features such as sliding skin, deforming muscles etc. There is little work on alternatives yet, but current standard of addressing worst cases with extra bones can't fix this and lacks common conventions about procedural bone animations. An exchangeable format for HQ characters would need support for programmable procedural animation.

Discrete LODs is no general solution to LOD at all. Epics Nanite is closest, but still too specific and restricted to define a common standard.

Now, til yet we were not able to solve those problems. Even not as we use custom engines, which makes custom solutions possible.
How will a Metaverse, spanning multiple engines and / or some common web standard, be able to allow still further progress on those essential building sites?
My guess: It simply wont. They (or we) will get stuck in bad standards, leading to low quality content lagging behind.
We already see this happening with current HW raytracing as an example: RT APIs blackbox BVH, and effectively prevent further progress towards any potential general LOD solution. Short sighted decisions to help GPU manufacturers for an easy start turn into innovation blockers, preventing progress for years if not decades. I guess the same will happen on the trial to have a global Metaverse, at a much larger scale.

And that's just about the obvious and easy things. If we start to discuss character animation regarding interactions with an unknown environment, it's clear abstract and simplified solutions on the level of Minecraft or Roblox is all we can expect, while true immersion remains out of reach. We can not even solve those problems well within a single game, not depending on any conventions. So how should we make any progress with that burden of low standards and conventions on top? Impossible. Thus the Metaverse will not give us HQ games, but the opposite.
But well, that's not really what the big companies want to deliver at all. They will offload the problem to content creators, and just hope the content is good enough entertainment to attract people spending some time outside of ‘daily work and life in the Metaverse’.

What may help them is chip crisis. As it looks now, next gen gaming in form of PS5 or RTX is simply too expensive to be mainstream, so the problem may solve itself with the extinction of console and PC gaming, and there wont be any competing higher standard anymore.

However, all of this is ignored. Instead working on the actual problems, we have a hype about blockchain. Which is presented as next big thing and opportunity, although it does not even address or solve any problem at all. We also see EA or UbiSoft investing into blockchain. Game to earn. WTF? Blockchain in form of mining has already destroyed PC platfrom, and they propose what? Ingame mining as a new feature? While i wait for just a single AAA game worth to spend more than 30 mins on for years?
It's crazy. The whole tech industry actually looks like a big bubble. Nipping champaign on Titanic and innovating marketing promises but nothing else to keep the ball rolling.

@JoeJ Sounds like their original plan was to lock out everyone from building in the metaverse and just make everyone consumers of it. Really embraced that Nolan Sorrento role. Shocking (not really).

I have to agree with JoeJ to a big part, but may be an open standard still would bring us on a better track than no standard. Even the car industry managed to agree on standards for data interfaces. The face book white paper about the Metaverse or “Meta” reads like one platform for everyone and that is theirs. A huge data collection unit about everything the users think, do or plan to do. May be the “Roman Empire” of social interaction, entertainment and gaming will not happen after all as long as we have the freedom of choice.

JoeJ said:
danger to get stuck at outdated / inefficient standards.

Yes, that's going to be a problem. But don't overdesign, or it will never work. Probably USD-grade assets are good enough for now, even if some systems can't render them fully.

JoeJ said:
chip crisis

In two years, a chip glut, say Forbes and Bloomberg. Everybody is building fabs. By the time the dev community gets the Metaverse figured out, the hardware should be coming out.

JoeJ said:
Discrete LODs is no general solution to LOD at all.

LODs for metaverse content are tough. Much of “metaverse” work means dealing with huge amounts of un-optimized content that has to be pulled in from the network. How well you do that determines how good the player's experience is. I've been dealing with that for textures in my Second Life / Open Simulator viewer, for which I've posted videos. That problem turned out to be solveable. Meshes next.