OpenSim isn’t the next Web 3D standard

I’ve been talking to a few virtual world operators lately that are trumpeting their particular virtual world as the next “Web 3D standard.” They’re the biggest, or the best, or the prettiest, or the fastest, or the cheapest, or whatever — and everyone is going to leave the other worlds and set up shop in theirs.

OpenSim is no exception, by the way. Sure, OpenSim is open source, and supports hypergrid teleportation, but it’s still just one technology platform among many. A fast-growing, cheap technology platform, sure, but still just one.

It takes more than that to set a standard. To me, a new Web 3D standard would have to match what we currently have with HTML, the communication standard for the World Wide Web.

First, we would need to have multiple servers. To host a website, for example, you could use Apache, the open source web server. Or you can use the proprietary Microsoft product. And there are also some other, lesser-known web servers out there.

Second, we would need to have multiple browsers. Websites, for example, can be accessed by Internet Explorer. Firefox. Google Chrome. Safari. And a number of other, lesser-known browsers.

Third, things would need to look pretty much the same no matter what server was running it, or what browser was being used to see it.  If you were to look at a website with the menu bar covered up, chances are you wouldn’t be able to tell what browser you were using. And you certainly wouldn’t be able to tell what was powering the website behind the scenes.

Finally, there would need to be some kind of common language somewhere. In the case of the World Wide Web, that common language is HTML in its various incarnations.

So is there a standard for the new 3D Web? Yes, there is, and its called LibOpenMetaverse.

Someday, that will roll of your tongue as easily as HTML does. If “HTML” can be said to roll.


Why is LibOpenMetaverse the new standard?

First, there are multiple servers. Today, Second Life and OpenSim produce worlds to the LibOpenMetaverse standard. So does realXtend, to a lesser extent (it adds mesh functionality, allowing for better graphics — but also creating some compatibility problems).

Second, there are multiple browsers. There’s the Second Life browser of course. And my favorite browser, Hippo. There are also Meerkat and the realXtend browser, and about a dozen other lesser-known ones.

Third, worlds look pretty much the same, once you cover up the menu bars and the rest of the browser branding. Some browsers have prettier menus than others, or add nifty tools, but to the casual user the stuff in the middle will look and act the same. (Except, again, if you’re in a realXtend world and there are mesh-based objects around — non-realXtend browsers won’t display those.)

Fourth, these worlds share a common language. The information sent between the browser and the server is in the form of LibOpenMetaverse instructions.

It is true that the worlds aren’t a perfect fit. Some complex scripts, for example, will run in Second Life but not in OpenSim — and vice versa.

If you recall, back in the early days of the Web, there were pages that worked well in one browser but not in another.'

Maria Korolov

Maria Korolov is editor and publisher of Hypergrid Business. She has been a journalist for more than twenty years and has worked for the Chicago Tribune, Reuters, and Computerworld and has reported from over a dozen countries, including Russia and China.

  • not sure about standards really. It could be that XML is really enough. I would say though that the bigger issues are persona and inventory. So something like OpenID that would alllow people to maintain their persona across web/virtual environments. Then the issue about inventory, which is only sorta solved in hypergrid. Avatars need to be able to access and use their inventory across areas. Logically inventory should be in the cloud and then loaded according the need/allowance on a grid.

    So I think grid owners/operators will need a set of protocols to allow them to manage how people access their grids and what roles they can have on them. We actually are using Drupal to manage this as it is more about a person's rights on a system.

    My view (today) is that next step will be a federation of grid owners that have some form of interchange agreements allow movement of avatars and inventory. Perhaps there will be a neutral framework (P2P even ) this is like the "Street" that virtually links them. At least it works in my dreams like that…

  • I haven't seen any of the problems you mention. Today, I regularly do hypergrid teleports between various grids, and am able to bring my persona — and all my inventory — with me. And put things down on other grids, and give things to other people, and get things from them.

    For example, the other day I picked up a new T-shirt from another grid. And I've got a pair of shoes around from the French grid.

    It doesn't always work well — items go into my inventory, but I can't always pull them out again. But sometimes I can, and there doesn't seem to be any rhyme or reason to it. Some OpenSim developers have explained to me that they only show up when they're in my cache — but I've accessed grids on brand-new computers, with brand-new Hippo downloads, and have seen items show up just fine.

    And then, I've had problems wearing my items when I just went over to the next region, inside the same grid. In fact, my problems with items during hypergrid teleports have been about the same as with items during normal, in-grid teleports.

    If there's something in particular I'm looking to bring over from one grid to another, I generally find that if it doesn't work the first time, I can try again a day or two later and it suddenly, mysteriously, works again.

    I just figure its OpenSim growing pains.

    I WOULD, very much, like to see inventories be browser-based, rather than intermediated by regions (where a region can send a command, say, to delete your whole inventory to your inventory server). And I'd like to be able to pick and choose where to keep my inventory — on my computer, on my standalone, on some nice grid — the way I can now pick and choose where to keep my email.

    And I'd like to be able to move my inventory easily between different locations. But then again, I can't do that easily with email yet — and we've had it for what, 20 years?

    The big problem now with inventories, I hear, is that rogue regions or grids can grab your inventory away from you, or even delete it.

    And this is a problem, sure. But if a region like that pops up, then the grid that it's on will probably take it down almost immediately — nobody wants that kind of thing on their own grid.

    And if this is a whole grid that does it, then you can't get to it except by hypergrid teleport. And you can't teleport in unless you know where it is – unless there are gates pointing to it, or people giving you directions for how to get there. You can't find a bad grid by accident, except maybe if you were typing in the name of a good grid, mistyped, and there was a bad grid lurking in wait at just that location. It will take a while, of course, before the size of of the 3D internet makes this is a big problem.

    We have websites that do that now — you land on them, and they open up all sorts of porn windows or download bad software on your computer and wipe out your data. But these websites don't normally come up in search results — no same person would link to them! — and some browsers will warn you against them.

    So while I'd like to see this fixed, if it isn't, no big whoop. We'll deal with it, the same way we've been dealing with the same stuff on the Internet all along.

  • Hi Maria,

    Just a few comments 🙂

    OpenSim isn't trying to be a standard, it's aiming to be a flexible platform in the same sense that the Apache web server is. i.e. People use different modules for Apache to support different scripting languages, authentication modules etc. etc. the same is true of OpenSim.

    Apache itself isn't a standard though it does support multiple standards, the main one being the HTTP protocol.

    OpenSim has the potential to support multiple protocols, one of which is the Second Life protocol which is also what LibOpenMetaverse supports, another is MXP (Metaverse Exchange Protocol). LibOMV in itself is not a standard but is a library for making applications that use the SecondLife protocol.

    In terms of different platforms and which one will be the most successful then time will tell, there will always be companies who release their own closed systems or under their own licensing schemes and do well from them. OpenSim's differs from those systems in that it's open source free and uses a BSD license which means you can pretty much do anything with it you want, it will evolve over time to keep up with technology.

    In terms of viewer technology then it does support multiple viewers though admittedly the SL viewers are really the only usable ones currently. I'd be tempted to argue that SL viewer, Meerkat and Hippo aren't really different viewers at least not in the sense that IE and Firefox are, simply because they originate from the same SL codebase and are all covered by the GPL license which actually prevents OpenSim developers working with the viewer for code contamination reasons.

    I agree with you in terms of features I'd like to see particularly surrounding inventory, this sort of inventory separation is one of the features the OGP proposed standard is looking at. With yourself or someone you trust running an agent service that holds your avatars details and inventory. When moving between grids then the viewer would bypass the region and go direct to your agent service to fetch inventory thus removing the region from the picture.

  • For me the platform today makes not an iota of difference whether , second-life opensim , or the plethora in between, as this is just the weaning period where we as a species take on this added layer of communication, which unlike as you point out above the WWW as we have known it with the exception of chat-rooms did not have the real time human to human interaction we have with VW.

    Those visionaries/dreamers amongst us know the tech will catch up making the arguments about inventory a moot point, no it is the protocols which are the significant learning curve we are on, what is your avatar? Who is your avatar? what is transparency? when to be transparent? are you a credible image of yourself? are the relevant questions today, for as I stated this will be just one more layer of the tools our species uses to interact and do what we do best “grin”

    The winner of the VW war of platform provider will be the platform that allows for answers to these questions as the up and coming users of VW will not give a damn about how it works, but what works.

    In the meantime though the evolution continues and great that so many would be providers are putting their chips in as this as nature intended is the only way to move evolution forward.

    The day in the future when we talk to a hologram beamed to us from the home planet of the Jedi is not too far of, tech willing and we better know what to say then is all i would ask “grin”

    Julius Sowu virtually-linked

  • Pingback: links for 2009-07-23 |

  • C3 — I'm using standard in the generic sense of the word, not the ISO compliant sense.

    And you're right, I didn't make that clear in the article.

    The reason I'm not talking much about VBRL or X3D is because I'm not seeing businesses using them.

    As a business user, I'd hate to invest money and time into a technology that isn't going anywhere. So I look around, and try to figure out what bandwagon everyone is jumping on. Right now, that seems to be the Second Life/OpenSim bandwagon.

  • c3

    Hard to fathom how this article could talk about web3d standards and NOT mention VRML 2.0 and X3D the ONLY ISO web 3d standards.

    Most of your responders seem to be just making stuff up.

    i gave a better review of your copyright article.


  • I actually agree on the Apache part — OpenSim functions to the virtual worlds very much the way Apache does on the World Wide Web.

    I wouldn't compare LSL or OSSL to HTML though — it functions more like Javascript does. In addition, OpenSim allows server-side programming, in the form of modules, but the same modules cannot be currently run by the Second Life server platform, which, as far as I'm aware, currently allows no server-side programming. And, in so far as PHP can be used to create server-side OpenSim applications, I would have to say that PHP is the new PHP.

    Meanwhile, the major browser makers have announced that they will be supporting 3d graphics by sometime in the first half of 2010, without the need for plugins.


    And we might not have to wait that long – Kevin Tweedy, who's working on the Xenki browser, says he'll have a web-based client in about six months, using Microsoft's XAML standard. (

    But I do take your point that I need to be more careful about the use of the word "standard." There is a generic sense of the word, and a technical sense of the word, and I use the two interchangeably, without specifying which one, exactly, I mean. Some people can probably tell by context, but I apologize to those who can't and were needlessly confused.

    The folks who are going around saying that OpenSim (or Second Life, or Blue Mars or whatever) is going to be the next 3D internet are using the word in the sense that Microsoft Word has become the word processing standard in business, or Excel the standard for spreadsheets. I believe they're wrong in both the technical and generic sense.

    OpenLibMetaverse is the language in which the browsers — like Hippo and the Second Life viewer — communicate with the servers, whether OpenSim, RealXtend or the Second Life server. In that sense, OpenLibMetaverse is the closest thing we have to a real, working standard today — a standard in the sense that its widely used by different players, and is gainign the most adoption — and may become a standard from the technical sense as well at some point down the line. HTML is the way that the Web browsers communicate with Web servers, so I would argue that OpenSimMetaverse is the closest thing to HTML in the 3D space as well.

    — Maria

  • Arulin of OSGrid

    OpenLibMetaverse I will agree is a agreed on technology. But my suggestion to you is to look at what is out there, what is accepted, and where it is going. Opensim enviroments are becoming the most adopted. That could change and if it does I'll be happy to change mine, but for now the 400 pound girillia is Opensim.

  • Arulin of OSgrid

    Ok , to the author of this….. First of all dingbat of the belltower, Maria yes you seem dumb as dirt to the fact that Opensim is a server not a technology, LSL, JS, VB, etc and engines like OPenDynamicEngine , ZeroMesher, and Meshmerizer are the standard runners for the gold.

    If Microsoft decided tomorrow to imp a 3d engine into I.E you could connect to SL, not that hard to do. Dose the term xml mean anything to you?

    Opensim is to like Apeche ,and the engines re the browser core and, the langs are like what loosely canbe compared to html (loosily compared) , by the way, you can connect SL client to an opensim server and view stuff the same way as you would in Hippo. You have to use a command line promt to get the redirect from Linden Lab’s.

    Only thing stopping this is a curreny format, we need to start begging Paypal to provide the VW with a standard.

    Opensim is the next apeche, LSL is the next HTML(more like the next PHP and kicking html to the curb), and most likely the client side browsers will handle OpenDynamicEngine and the LSL with or without Mozilla or M$ since these are standards in the virtal worlds that are out there.

    It’s not one piece that will make the Virtal web, it takes many imperfect parts togather making a perfect virtal web. Now take your arictle , find the nearest toliet and put it where crap belongs.

    RobS, I give kudos to your post of comparing Opensim to Apeche, happy someone here knows what they speak of.

  • Arulin —

    I totally agree with you. I also think OpenSim is the future. I was being a little cute with the headline of the article, that’s all, since I’ve been hearing people throw around the term “standard” in connection with OpenSim.

    I think we’re going to have different server platforms — OpenSim, Second Life, realXtend — just as we have different web servers. And we’ll have different browsers (16, at last count, I believe for these guys). The thing that connects all these together — that makes it possible for all these browsers to visit all these worlds — is OpenLibMetaverse. Which nobody ever hears about. Maybe it needs a better name. OLM maybe? But an olm is kind of blind salamander (thank you, Google!). Hmmm… maybe it could work.

    – Maria

  • Maggie Darwin (@Magg

    Any alleged discussion of 3D and the web that somehow fails to mention Google's O3D rendering framework has really missed the boat, in my opinion. No, it's not a full VW solution. Nor is it the 21st century replacement for VRML. But it's a key technology that you ignore at your peril.

  • M1

    A bit late to this post, but an interesting article. I agree with C3 on the lack of mention of the existing ISO standards.

    1.To say, "The reason I’m not talking much about VBRL or X3D is because I’m not seeing businesses using them." No, you have seen businesses use them and VW platforms built on them but not openly mentioned. It is what runs under the hood or a part of the production pipeline because it is that well built and established. Most industries that use these standards are the ones that bypass the gimmicky nature of what VWs have become. Not everything needs an avatar and most real world enhancing web3d applications need something more robust than a casual VW platform.

    2."As a business user, I’d hate to invest money and time into a technology that isn’t going anywhere. So I look around, and try to figure out what bandwagon everyone is jumping on. Right now, that seems to be the Second Life/OpenSim bandwagon."

    I know this is an old post, but to say the ISO standards are not going anywhere is a comment made that shows the little information VWs enthusiasts truly know about web3D beyond the VW industry hype. Even at the time of this post, the hype in VWs was about dead and this confirmed by the low attendance at the Engage Expo in Sept. If you are bandwagon jumping on non-ISO standards, I say drop OS and go for Unity 3D. Much better experience in a browser plug-in, already supports most mesh import, multi-user, multi-device, free, and has a user install growing faster than SL and OS. Not to mention has a higher 'cool factor' among most of the multi-user online audience.

    Agreed that OpenSim will not be a 3D web standard but it is good for what it does and has a great community behind it. I think it is a great tool for those that want to build 'open' virtual worlds that look like SL. For those that want to build the 3D web, I agree with Maggie Darwin to look at technologies like O3D, HTML5, WebGL, and X3D.


  • M1 —

    O3D, HTML5 and so on seem to be standards for displaying 3D – not standards for interacting with immersive virtual worlds. These protocols may be used to create viewers for virtual worlds — I believe there is work being done now with Unity 3D, for example — but not for the core immersive functionality.

    The difference between immersive and non-immersive 3D is that immersive worlds engender a feeling of presence, of community, of social interaction. It's the difference between looking out through a window, and going out into the middle of things.

    Meanwhile, though I've seen plenty of examples of 3D games on the Internet, I haven't seen many successful enterprises uses of 3D outside of the Second Life/OpenSim/realXtend environment. There are a handful of companies using Forterra's Olive, but other than that I haven't heard of a lot of big customer wins for other platforms. There are small, startup alternatives like Altadyn 's 3DXplorer and Utherverse, with their own proprietary platforms, but both of these are based on proprietary technology — and I haven't seen a lot of business adoption of either of these.

    Am I missing something?

    – Maria

    Please do let me know if I'm missing any!