Your long-term OpenSim goals

In a recent comment thread, some readers have asked questions about OpenSim’s long-term goals and road map.

I’m going to offer some suggestions now, and ask you, my readers, to contribute others in the comments. Then I will do a survey of our readership to find out which long-term features are most important to them.

There are no stupid suggestions, so be nice!

My ideas for OpenSim for the next three to five years

Browser-based viewer: I understand that there are some technical difficulties preventing OpenSim from running easily in the browser, that need to be fixed on the server side. Now that OpenSim viewers are becoming independent of Second Life viewers, maybe this is something we can pay attention to. Especially with Firefox’s recent announcements about native browser support for virtual reality. A browser-based viewer is a must for mass adoption. It’s not sufficient in and of itself, as Cloud Party unfortunately demonstrated, but I believe that it is a critical step.

Content protection: I, personally, believe that licenses are the way to go instead of trying to use technology to lock down content. The Web grew fine without content protection, and both iTunes and Amazon have dropped DRM protections on music downloads. But many creators are hesitant about coming to OpenSim without protections in place. The technology is there — the “export” permission setting is in place. It needs to be tested and rolled out to the commercial grids, so that everyone can be on the hypergrid without fear of losing content. Or, at least, no more fear of losing content than they currently have on the closed grids.

Gestures: This is supposed to be the big differentiator for High Fidelity, the fact that you can move your arms and your avatar will follow suit. I would like to see something like this. Maybe based on cameras. I think it’s a bit too early to start working on this, though, since the input devices are still up in flux. Once something shakes out, then OpenSim should begin working on supporting it.

Move functionality from menus to objects: This, for me, is the critical one. We all know how to interact with objects. We sit on chairs, we open doors, we bang hammers. By comparison, using menus does not come naturally. The more functionality we can move from menus to objects, the easier it will be for OpenSim to be used in virtual reality mode. GameFace, one of the new crop of headset vendors, is planning to do this with their virtual living room concept. This means virtual closets instead of inventory folders, for example — click on an outfit to put it on. Actual working mirrors instead of the “appearance” view. Magic wands that you can use to point to objects and move them around.

This picture was taken in OpenSim, using water to fake a mirror.

This picture was taken in OpenSim, using water to fake a mirror.

OpenSim’s role in the metaverse

With Second Life about to enter its end-of-life phase and development shifting to a next-generation platform, the current goal of keeping up with Second Life will become less and less relevant.

Plus, OpenSim pretty much has feature parity with Second Life now. The only major Second Life feature set that OpenSim doesn’t support is the new experience keys, now in beta, but it could be argued that OpenSim’s “threat levels,” which have been in place for a long time, offer the same functionality with a great deal more control in a way that’s much easier for grid owners and application developers to use.

OpenSim has long prided itself on being a user-driven, community-led project. Some open source projects — including WordPress, Ubuntu, OpenOffice and Drupal — have a single commercial entity at the helm to provide some forward momentum and long-term planning.

OpenSim has a loose group of volunteers — core developers, big corporate supporters like IBM and Intel, vendors and entrepreneurs, academics, non-profits, and government agencies.

In fact, according to a 2012 study, OpenSim stood out from other open source projects in the diversity among its contributors, the level of collaboration between different types of contributors, and particularly the contributions made by entrepreneurs.

A typical open source project is more homogenous in its contributors.

This diversity is both a strength and a curse for OpenSim. A strength because it results in a broadly-based, multi-functional platform that makes it well suited to becoming a core platform forΒ the metaverse.

That includes not just the “threat levels” functionality, hypergrid, NPCs, various backup options, but also user management features and grid management tools and fine-grained access controls. OpenSim can scale down to a single region on someone’s home computer, to giant grids with thousands of users run on multiple servers.

And, of course, due to the hypergrid and the complete absence of any centralized services, OpenSim can scale to infinite size, as outlined in this 2010 paper by Justin Clark-Casey.

No other current projects with pretentions of becoming a metaverse platform can offer this.

But I can see why people are anxious. Two or three years from now, other platforms are coming online with shiny new features. And we’re worried that, as virtual reality technology improves we’re not going to see the expected influx of new users coming to OpenSim to build their virtual realities. Instead, our existing users will jump ship to the new platforms. Our investments of time and money in OpenSim will have been wasted.

But it doesn’t have to be that way. We’re further ahead than anyone else when it comes to actually having a working metaverse platform. And we can stay ahead.

Add your suggestions for new features in the comments below.

Related Posts'

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. Follow me on Twitter @MariaKorolov.

22 Responses

  1.' Lani Global says:

    Six suggestions/comments: (1) Browser viewing of simulators is essential to the future of OpenSim. Maria appropriately starts the suggestions off with “browser-based viewer”, and I second the motion. HTML5 is a good starting point. (2) But, there is a bigger issue at hand: The trend for players/users is away from specially built “gamer computers”, high power desktops and laptops. (3) For OpenSim to remain viable, system development could move in the same direction browsers are moving, toward “client-recognition” with flexible scaling for offloading some of the rendering power (more like streaming video) to the server side whenever needed for users who are on less powerful mobile clients. [Even if this results in a separation into simplified viewers for mobile, and the full-blown creation-viewers we use now.] (4) Freed from the chains of SL-interlocking of Third Party Viewers, OpenSim development could expand toward some experiments in voxel-based interoperability, voxel conversion sim utilities, or pseudo-voxel features. (5) Simulator environmental changes: Uploadable mesh land that has all the properties of present land (but with greater flexibility in design, such as caves and tunnels), and the easy elimination of void water and void sky entirely whenever desired by the sim creator, especially with Z axis (altitude) parcels or height zones.

    •' Lani Global says:

      (6) The capability to embed a simulator view into an element (window) on a website. This could also be limited to a flying cam, or a choice of some pre-set avatars, with some simple Local Chat and Instant Messaging.

  2.' Minethere says:

    Okie dokie, I will weigh in.

    I want to see the hypergating aspects further refined so that it becomes as seamless as linking to a website. I also see no reason to believe this will not happen, eventually, just as I watched website issues go away.

    I do not like this “forced” permissions thing at all. I like the free spirit of the HG’ed Meta and the sharing people involved in it. I would prefer the licensing idea rather than someone, usually with commercial agendas, force their views upon us. I already experienced that, and the drama it causes is very unpleasant, as I know others similar to myself feel.

    I would really love to see a simple browser based way to enter OpenSim.

    The caveats are that I am not young anymore, which leads to the fact I will likely not be involved all that much longer [maybe a couple of years or so I guess] and that the tech is going past my ability to keep up. If anything, it would be nice, if I were to be able to do so, come back in some future time and find it a simple to use system. I don’t especially have that option anymore, but for others, certainly it would be nice.

    So if it changes to much or becomes too restricted, and does not simplify in some regards, I will just stop using it…and it won’t be a big deal, those who continue to be involved will keep moving along doing their things and I won’t even notice.

    Overall, I want to see some non-forced systems that allow some people to make money if they need it, allow free thinkers to continue having an unrestricted ability to do as they imagine. and not go backwards but forward in some kind of system similar to moving along webpages.

    I like how Kitely does their perms things as it forces nobody to do anything they do not wish to, I do not like the commercial mindthink that forces the many to conform to the few.

    I like the idea of some kind of increasing hostings of land rather than closed systems, which to my eyes are antithetical to how the net has always been. If I wanted to play a game by other peoples rules, I would be doing that, instead of this.

    A blend of some kind would be nice if that is possible.

  3.' Lani Global says:

    We need a much better HyperGrid and Grid search function. Perhaps a way for search engine bots to crawl sims? See my old parody called “Goggle Sim View”.

    •' Lani Global says:

      Parody: “Goggle Sim View, the grid crawler.” A search engine NPC bot that drives a cam vehicle. Although it was intended to be a joke, taking a step toward Search Engine compatibility would put OpenSim on the map.

    •' Gaga says:

      A number of grid search engines already exist including one I coded myself @ with its own web crawler (for stats) to explore what I thought might be useful. I maintain it adding and removing grids and though I haven’t done any further work on the code, I still find it really useful for quickly finding a grid in V2/3 type viewer.

      As for what I would like to see in Opensim I agree with the need for a browser-based webGL service but without the need for building tools since it needs to exist as a first look viewer with minimum learning curve (a full viewer can be used for building, etc). I like the idea of gestures too where, in particular, one click to don clothes would be great. I also agree with a lot of what @Lani Global suggested too especially on the implementation of voxels and “(5) Simulator environmental changes: Uploadable mesh land that has all
      the properties of present land (but with greater flexibility in design,
      such as caves and tunnels), and the easy elimination of void water and
      void sky entirely whenever desired by the sim creator, especially with Z
      axis (altitude) parcels or height zones.” I think stuff like this will help ensure Opensim keeps up with aspects of High Fidelity and SL2 and a more transitional way rather than closing it down to start something new like Linden Lab plans.

      I’m sure we will continue these discussion on our popular Opensim Virtual G+ community @

  4.' Alex Zed says:

    (1) Two level hierarchy linked sets (example: root -> child1 -> c1…cN) or more…
    (2) The function llSit(key avatar, key target) whit “permission_sit”
    (3) all configuration parameters in a table on the db (except for the database config… of course)
    (4) nat-loopback substitution directly in the 3D browser
    (5) a pre-configured server (ex: diva) directly packed into the 3D browser installation.
    (6) why 2D browser have to embed 3D browsers… when 3D browsers can embed 2D browsers???
    (7) after the (6) hypergrid adresses directly in the “address bar” not in the “map”
    (8) a warm greeting to all my friends

  5.' Han Held says:

    As SL enters it’s end of life phase, I imagine that businesses will begin to peel away -either in favor of the next-generation technology, or because it’s no longer possible to serve this space from a business viewpoint.

    I posted the source code to opensim up on to future-proof the hypergrid. Business’s may go, but I believe that the network of desktop grid owners has a solid future, though our numbers will take a hit as time goes on. My personal goals, perhaps pipe-dreams, are as follows:

    1)create a viewer archive, to guard against projects going dark. This most likely will feature singularity as it’s the easiest viewer to compile at the moment.
    2)create an archive of working mono sources (probably 2.8 and 3.2.8 …I need one earlier version but I’ve been lazy about learning which) and get experience with building them on linux.
    3)flesh out documentation on opensim that I started, eventually provide beginner-friendly guides.
    4)maintain my group “all the worlds” ( ) in hopes that it will be useful as a hub for desktop admins and end users to network from
    5) End user conference …desktopCON! A week series of causual events on the hypergrid intended to encourage regular desktop “admins” (sim and mini-grid owners) to meet and greet and share tips and experience and just generally shoot the s***, possibly with a class or two (depends on $$$ so probably a pipedream).
    6) Maintain an archive viewers, mono and opensim sources with a mind towards creating a fork or a “frozen release” in the event that the programmers become too zealous in their content protections at the expense of the “four freedoms of free software” ( )

    I remember the CB crazy of my youth and how it died out and became a niche hobby …I anticipate opensim going the same way. I believe this can be positive (from an end-user impowerment angle) and I’m settling in to use the software on my desktop as long as I have a desktop. I am hoping to build up and spread out the tools to allow others to do the same, and for new people to join me if they wish to.

  6.' Minethere says:


    Justin commented here about some of this.

    As to the for-profit aspect, from what I can see, those are the minority. Even Justin has that donation thing he added but last I looked there was not much in the way of donations [a month or so ago].

    Oren of Kitely, as most of us know, was recently added to core, and of course they stand to profit. Whether or not they do, or they make what they expect, or would like, is, of course, their own roadmap. He did add some very good patches to hypergating aspects as I can tell from a layman’s perspective.

    Even the sl wannabees are not actually making all that much. Of course this is all relative to what each person considers “all that much”. I do expect more of them to join the hypergated Meta as the benefits become more clear to them. TangleGrid did [and maybe still does] have 2 [I think it is] HG enabled regions as introductions to get inside their grid. I used one recently at their Steampunk Exhibition.

    So I think that while some people have done the simple thing and tried to ride on the SL skirtstrings business model, this will all become a combined footnote in the history of VR where likely only SL will be mentioned.

    Some type of combination of freedom which includes hypergating, and some web-based content sells model, with a commission derived for the hostings of it, and a land based hosting scenario likewise open to the HG is what I expect this will all boil down to.

    Once this is all settled properly then we can continue to move ahead with more focused goals….imho

  7. Nice article Maria, you pretty much took the words and thoughts right out of my mind. I think this is a crucial time for OpenSim, we are at a cross roads.

    I pretty much agree with all the suggestions made so far. However, just like in the desktop Linux opensource world. If a particular distribution goes in a direction that a lot of users don’t like – for example when Ubuntu decided to ditch Gnome2 in favour of its own Unity interface, many users just forked the distribution and took it in a direction that the majority of users preferred.

    I am not saying that this should be done, but there is always the option that if it isn’t possible – or the core developers on OpenSim are unwilling – it would be one option to fork the OpenSim project, and create a new one under the direction of a more tightly knit, and forward looking organisation that could take the project on into the future.

    Let me be honest, I realise that what I am about to say will probably be deeply unpopular, but hey ho I have thick skin. I personally would have no issue at all if there was a USD$50 charge to download each new version of OpenSim, ie, a requirement to purchase the code. Once you have paid and downloaded you can use for your purposes, and when the next version comes out, another purchase must be made, possibly with a discount. This could be one way to help support the development of OpenSim, and also to ensure that the developers listen more to its user base.

    I am all for opensource and free as in ‘free beer’, but projects that lack funding in some way, rarely succeed in the long term. I am watching what happens over the coming months and next few years with great interest.

    •' Han Held says:

      A better idea would be to go the Redhat route; charge for the binaries, charge for guarenteed support contracts. The only way to charge for the source would be to make it closed source …and five minutes after you do that, we’re going to make a truly open source fork of it, from a previous version if need be -being open and libre is intregal to the adoption of the metaverse.

      Redhat has given opensimulator an example; it’s up to them to follow it.

        • I would be willing to pay for a nice, packaged, pre-configured, shiny version of OpenSim that does everything I want it to with no work on my part. (New World Studio — where is that update??) And I think that many grids out there are willing to pay for a supported packaged version of OpenSim, even if they do their own hosting. (And I know of some grids that already do that.)

          This doesn’t require any work from developers, though. Any vendor could step up, create a little branded distribution (I’d be MORE than happy to help them do this!) and sell it.

          ReactionGrid was doing this for a while (though I think their product was overpriced) before they fully switched to Jibe. IBM had one, too, also a big overpriced for my tastes, but they were going after big corporate customers, so I wasn’t really the target market there! LOL

      •' Alex Zed says:

        Very good Han

  8.' Serene Jewell says:

    I think OpenSim has a lot more potential than many people realize. The creation of a network of open source, decentralized, autonomous, virtual environments with the ability to hypergrid between them is an amazing accomplishment. I would like to see something akin to the Linux Foundation formed to raise money to pay developers and create collaborative environments for problem solving, training, and promotion. My guess is that we are about to see a huge influx of closed-grid, corporate virtual worlds where people will be given a fixed corporate-created experience. (Kardashian World, NFL World, Harry Potter World.) Nothing wrong with that, commerce is good. But it would be nice if OpenSim could stand as a strong alternative to that where people are empowered to build their own experiences. To be strong, developers need to be paid and the OpenSim ecosystem needs to be promoted and protected.

    I’ll also echo the calls for browser-based viewing, smooth hypergridding, optional packed distributions, and a grid search engine.

  9. Facilitate user content generation and control. Lots of possibilities there, there must be a sweet spot between the minecraft type of building ie block by block, to traditional mesh vertex pulling. We opted for something in between and continue to explore this space. We think that users also want to change their environments in ways similar to how how they do it in the real world. This points to direct manipulation of objects e.g. moving a plate of fruits should allow one to move a plate intuitively with the fruits moving as expected etc.. simple little incremental improvements like that will go a long when scaled up to other objects. You can get more information about our approach from this thread on Kitely We believe simple trivial little increments can have significant impact.

    The next thing we are doing is trying to push all content creation away from the desktop interface to inworld. This migration will have to happen slowly. There are many reasons why it is helpful to have this option. Firstly building process is naturally collaborative, then it becomes easier to get proportions right rather than when designing in a 2D screen.

    We can keep on pushing fidelity and exploring the deep end of physics simulation and lose the plot regarding the importance of good content design. Why is it that 2D games or even isometric games on mobile devices generate hundred times more revenue than virtual world content? that’s where our questions start.

  10. Well, obviously everybody has a different set of priorities πŸ™‚ Because of my own work with OpenSim, I would like to see a much better support for autonomous agents β€” and this means mostly better and more reliable sensors (collisions, for example, are not passed from NPC avatars to attachments where they could be processed; sensors still have the same limitations as LSL; ray-casting solves a lot of problems, but, due to its impact on the physical engine, is seriously limited). In particular, objects in OpenSim are mostly detected at the centre of mass; while you can get a bounding box to determine overall size, this usually means that autonomous agents will avoid huge objects, even if they have doors and windows (unless several tricks are used). The ‘pathfinding’ solution used by LL is not really something I’d like to see in OpenSim β€” it requires far more effort to implement than the benefits it actually confers (better sensors would go a much longer way to solve the issues).

    Hierarchical linksets would also help a lot.

    I would also like to see the ‘gestures’ being better implemented. In particular, moving away from the idea that you need to do your animations in Poser/DAZ/Blender and then import them; while this could be optional (to allow import β€” and eventually export! β€” from other tools), animation/gestures should be done in-world by manipulating avatars directly on the viewer. Animations currently also have priorities, something few people use effectively (everything is set to priority 3 or 4), and priorities ought to be controlled at the user level (i.e. just a common property on the asset). This would allow things like hierarchical animations (which was supposed to be LL’s original plan in 2004 when user-created animations were introduced), and make things like face morphs or small gestures (to follow a camera, for instance) far easier to use. Again, this is something that is ‘almost’ possible today, it requires just a small change at the database level, and some support on a TPV.

    My biggest issue is that as SL1 fades out, TPV support will also fade out as well, so OpenSim needs to address that β€” by having its ‘own’ viewer (i.e. part of the OpenSim code, not as an independent project) β€” pretty much like High Fidelity. This might mean either to co-opt Radegast (because it’s independent from LL’s own code) or to develop something from scratch β€” perhaps as a web-based viewer.

    A redesign of the core services is also something worth addressing. At some stage, the core devs moved several key elements out of the core β€” cache, IMs, groups, profiles, etc. β€” and this meant that grid operators could shuffle these services to different servers (and technologies). These were poorly documented. But they work. However, it also meant that standalone servers would be way harder to deal with. The current trend is to move everything back into the core code again β€” making ROBUST ever so more robust (bad pun, I know) β€” instead of a constellation of independent services. This is not really something I have given a lot of thought about, but I rather like the idea of those attempts to create a PHP-only application to deal with the core services, and just use OpenSim for simulating regions. The main point is that PHP-based applications, being stateless, are very easy to massively parallelize, by distributing it among several web servers, and much more easily cached using out-of-the-shelf web optimization solutions (from proxy servers to CDNs to cloud services…). LL is allegedly doing that for all their core services as well.

    It would also be nice to move towards an ‘OpenSim 2’ which would be simply a web application β€” not a ‘service’ running separately. Web applications are much easier to manage. Ideally, such a drastic change should be way ahead on the roadmap. I favour PHP, because pretty much every Web hosting provider supports PHP, but PHP is simply not fast enough for a quasi-realtime environment. We’d have to wait for PHP 6.0 and a working, well-tested JIT compiler. In the mean time, we have Python or Java. While Java is hard to manage, Python starts to become more accepted by Web hosting providers. Of course, it’s much easier to port code from C# to Java than to port it to Python and/or PHP…

    Well, my second biggest issue is that OpenSim, not being ‘driven’ by a commercial organisation, or a strong foundation with a crystal-clear purpose β€” ‘building the Metaverse’ is a very vague objective, meaning different things to different people β€” it means that its roadmap will not be something well-designed to which there will be a ‘commitment’ by a core team of developers. Instead, each will work on whatever bits they fancy, with occasional ‘spurts’ of development as some organisation or group suddenly drops a lot of code, used for their own projects, and releases it to the community at large. We saw that happening with IBM, Intel, and more recently with Kitely (not to mention many independent groups who tackled a specific area β€” like, say, the Flotsam team β€” and released all their massive code from one version of OpenSim to the next).

    What this means is that all my above suggestions might get a frown and a question: ‘Why don’t you implement them all on your own and contribute the code to us?’ πŸ™‚ My own answer is that I’m not familiar enough with the code to add all those changes. And I’m not a professional programmer, merely an amateur. On top of that, and taking into account how High Fidelity is currently evolving, I would seriously question my decision of investing so much time in OpenSim to try to get it to do what High Fidelity is doing, or to invest the time to change High Fidelity to do what OpenSim already does. What I mean is that, until we are sure that the High Fidelity project really goes ahead, new volunteers wishing to contribute code to a virtual world project which will clearly start to diverge from what SL currently is, have now two choices. They can work with the current OpenSim codebase and try to tweak it to do things that SL never planned to have. Or they can contribute code to a brand new, shiny project, which already tackles all those issues β€” but isn’t compatible with either SL or OpenSim. The choice is tough. Ironically, LL is pointing the way: just forget about the old code, let’s move on to shiny and new. What argument can potential OpenSim contributors be given for them to keep working on OpenSim instead on ‘shiny new’? It’s not as if OpenSim’s user base is massive. I can imagine that people having staked their professional life in OpenSim β€” researchers, grid operators… β€” will be more than willing to join the effort of further developing OpenSim, but what about the rest of the world?

    We have tough choices ahead of us.

  11.' KeithSelmes says:

    I initially saw SL and OpenSim as being useful for education, training and research.
    From that perspective, I would see OpenSim as specialist software, which has already been very usable for several years. The main problems I’ve encountered are the need for good 3D graphics capability on the desktop, and the need to get Opensim through the corporate or institutional firewall.

    On the more basic corporate and ex corporate desktops, some older viewers still run quite well, and seem to handle everything I need, up until varregions. Newer viewers are very difficult or impossible. I wonder if this is because viewer development follows SL requirements, but I do wish we could have a viewer with less hardware requirement but able to work with the new OpenSim release.

    A browser-based viewer might solve the firewall problem, but I’d like a viewer that’s easy to use and runs on a mainboard chipset, and it wouldn’t necessarily solve that problem.

    A standard good quality voice module would also be highly desirable with respect to security, I believe Vivox services route through their own servers in the USA, and that’s why MOSES don’t use Vivox. There’s also a potential saving in running costs.

    Beyond my original interest, the advent of Hypergrid, easy low cost hosting, and improved content availability, should all make OpenSim more accessible to more people, many of whom could benefit from virtual worlds, or simply enjoy using them for relaxation. However, end user hardware is still a problem. Many people who would benefit from this technology have limited resources and old equipment, whilst many now have android or ios devices instead, which are usually less capable.

    From recent news items on VR developments, it seems likely we’ll see a marked improvement in mobile devices in the forseeable future, and I wonder if that will persuade mainboard manufacturers into adding a decent 3D chipset as standard. But, even if improved hardware becomes available, it will take time to filter through from early adopters (with money) to the population at large.

    It seems I’ve persuaded myself, for now, the biggest issue is the viewer, but simply having it work in a web browser mightn’t be a whole solution. I wasn’t able to use Cloud Party on a tablet, even with a WebGL browser, and the real need is for something more people can use, whether it’s in the browser, or it’s an app or client download.

  12.' Herkimer Highmist says:

    For the gestures idea, I’ve got just two words: Wii Remote!

  13.' Samantha Atkins says:

    I think and offering of Grid Services that was not tied to any particular grid would greatly ease creating a viable metaverse with all of your content accessible to you in any grid available in every grid with the permissions it was first acquired with. Much of the hassle of hypergrid is how to restrict content so that other grids with different policies and standalones can’t rip off creators that care about protecting their content.

    In hand with this the one CPU or one server process per region model needs to be removed. Instead the work necessary to run a region should be spread among as many service provider process on as many hosts, scalable on demand, as required. This is the more modern way for distributed software. The efforts to do some of this to support 1000+ users on a region that have been done and open sourced are a step in the right direction.

    Both of these are “think out of the box” type suggestions. Don’t put everything in one “box” for a grid or a region. Don’t put users in a box, especially not for their inventory and avatars. This imho is the direction most critical to making the metaverse real and vibrant.