Core devs: OpenSim needs new viewer, new people

Moderator Joyce Bettencourt and OpenSim core developers Robert Adams, Melanie Milland and Ubit Umarov. (Image by Maria Korolov.)

OpenSimulator’s core developers said that they’ve gone pretty much as far as they can with the server code, and what the project needs is new people — and a new viewer.

“My own kids say that it looks ancient,” core developer Melanie Thielker said at the developer panel that kicked off this year’s OpenSimulator Community Conference. “We’re looking at more a ten-year old rendering technology that we just have to ditch.”

Thielker was joined by two other core developers, Robert Adams and Ubit Umarov. Joyce Bettencourt moderated the panel. Bettencourt is the founder and creative director of AvaCon, the non-profit which organized the conference.

Future progress depends on moving away from Second Life-compatible viewers to having a separate viewer just for OpenSim, she said.

There was one attempt to build on using the Unreal Engine, a modern video-game development platform, but funding ran out before it was completed.

“I think any future solution is going to involve the viewer,” agreed Adams. “To change the core simulator we have to change the viewer.”

Umarov agreed with this assessment.

“The viewers are the main problems at the moment,” he said. “We are stretching too much with what we can do with the Second Life viewer.”

However, the OpenSim core developers all work on the server side. The viewer development teams are typically a different set of people, and uses a different programming language, and has a different licensing model.

“We are getting behind in terms of technology,” said Umarov. “When it comes to modern games, we are completely out of date. We cannot improve that, because most of that is viewer-side.”

There will be a separate panel for viewer developers today at noon Pacific time.

Meanwhile, there is some bug fixing work that remains on the existing code base, and it could use a general restructuring to be more modern. For example, it could be adapted to work with cloud technologies, said Adams.

“We need to think about what the next ten years will bring us,” he said.

 

That might need new blood, he said.

“Really, the core group has lots of other interests going on and it would be great to get more people in here,” he said.

 

 

Related Posts

maria@hypergridbusiness.com'

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.

  • Han Held

    My understanding is that the server devs are absolutely abysmal at communications, particularly with the viewer devs. I’ve seen a lot of frustration on Firestorm’s Jira (in particular) over the fact that they have no idea what’s up with opensim in terms of what’s required and what’s supported.

    Also viewer development is expensive and time consuming, so good luck with that.

    • Da Hayward

      quite right Han, jessica her self has commented on here before that she welcomes any feed back from developers and such

  • Arielle

    “My own kids say that it looks ancient,” core developer Melanie Thielker said”
    I’m not surprised considering the avatars in the photo looked like they haven’t changed since 2007. Might sound like a silly quibble but when 3 core devs are not even stretching the rendering engine from a decade ago, of what need is there for changes to the viewer?

    • Yeah, I thought that image was FROM a decade ago. When I use my gaming laptop with 4k resolution the worlds look amazing. Especially when the objects are mapped well or textured well, quality made items appear wonderfully even with the viewers we use today.

      • Han Held

        There’s *got* to be a middle-ground between “I have to have the newest styles even if it means copybotting maitreya” and “I’m too 133+ to care about my avi, as you can tell by the fact I’m wearing play-doh on my head like it’s 2003”. (Sorry, that only worked for Philip Linden and even *he* got a make-over in 2010.) To me it doesn’t look 133t, it looks out of touch.

        tldr …balance is a Good Thing, yo!

        • “I’m too 133+ to care about my avi, as you can tell by the fact I’m wearing play-doh on my head like it’s 2003” — oh boy. I really laughed out loud at that.

  • Douglas Maxwell

    smh

  • Geir Nøklebye

    While I agree that the renderer needs a serious update, most people, including the core devs don’t seem to understand there is a close relationship between the viewer rendering code and content. You cannot just develop a new renderer and hope it works unless you also ensure backward content compatibility. This makes it much more challenging than say plunging a new renderer into a project like Sansar.

    The good news is that LL is having a major go at updating the rendering code which will come to fruition in 2018. We have already seen some of this in the form of patches cleaning up buffer allocation that is one of the major causes if instability of viewers.

    it is also interesting to note that most of OpenSim don’t really use the viewer to its full potential, evident in the rather big gap in creativity and fidelity you see in SecondLife compared to OpenSim. Creations that are used and crafted with the exact same viewer code as used for OpenSim.

    Further – and I have asked for this many times before, what exactly is it that the viewer blocks OpenSim from developing? Where is the list of features not covered? Where are the visions (or even rudimentary roadmap) that require code changes to the viewer? Where is even the dialog between those who work on viewer code and the OpenSim developers? That dialog hardly exist at all, and mostly consist of read the code, after the fact, after someone have figured out a “great” idea.

    Not to belittle the work done on the server code by “core”, as the effort over time has been formidable, they – as a group, need a good look in the mirror if they want change to come to the platform that is OpenSim.

    • JozeeTungsten

      Exactly right on all counts. Furthermore, I had occasion to work on a different type of server project several years ago, and the viewer code was developed along with the server using Google Web Tools one of the many application frameworks available. IMHO for this software to ever advance beyond the working prototype phase we have to address a number of basic issues surrounding the platform. Industry best practice would suggest mono/C# is not a good solution; an application server such as Websphere or JBoss would be much more appropriate and enterprise Java would also be more likely to attract developers to the project.

  • Mike Chase

    Maybe since the core team has such an palling lack of vision they should step aside and let someone else run with things. I and a number of other have tried in the past to contribute but often contributions have been treated with a lack of enthusiasm to outright hostility.

    In other words, the core team is at least half the problem.

    • Beth Reischl

      I think maybe it’s time for some curtains to be pulled to the side about the OpenSim core team and how it has worked over the last decade.

      InWorldz got a lot of flack over the years for not contributing to OpenSim. We’ve taken it in stride because quite honestly, we didn’t have time to do battle over whether we should have forked or not. Much less pointing fingers at why InWorldz forked to begin with. However, if OpenSim wishes to be around in 1 year, much less 10 years, perhaps this is the best time for it to take a good long hard look at it’s core team and how that team has treated developers over the years. And that’s where InWorldz comes in.

      Back before InWorldz even had developers, there was me. I was it. And one of the biggest bugs back then was that you could not add a new folder to your inventory. It added, sure, but if you relogged, it was gone. So was anything you put in it. Upon investigation, I found the function for it, and lo and behold, there was no there there. No code, nothing. Just the function, empty. So I wrote it, tested it, everything worked. I then submitted it to OpenSim. It was removed, completely deleted. 3-4 months later, it was resubmitted under the core devs name, exactly as I had written it.

      At that point, I determined that if OpenSim was going to operate under those conditions, InWorldz would not be part of it. We would fork. Over the years, we paid many white-hats to test our security and find bugs in the security (such as being able to insert oneself into a group as the founder of the group and make many changes to the group). Every security fix we applied, was back-channeled to OpenSim. Every fix went in that we submitted quietly. Every grid out there, has a ton of security fixes from InWorldz, although we’ve never taken credit for them because we felt the high ground was the right path to take as a company. Unfortunately, that was not reversely applied to us by others who as a core dev, decided that they would tout those same security fixes as making their commercial grid more safe than any other grid, including the grid where those fixes came from.

      There’s a reason why many groups and devs have left the OpenSim team, forked off, or just quit all together. Maybe it’s time for the metaverse to start speaking up, because if OpenSim wishes to even continue at this time, it’s going to need to do a massive overhaul of it’s core team and which direction it needs to go in.

      • Steve_S

        Whatever happened to Moses / Halcyon ? That also seems to have gone the way of the dodo. As I recall that was forked from InWorldz and seemed on a forward path with a new viewer / UI.

        With the new generation of VR/AR “glasses”, many of which are self-contained things are ripe for big changes which are needed to keep current.

  • Mattie

    The two things the above mentioned core team absolutely excels at are taking credit for somebody else’s work and passing the buck.

    It’s no wonder gamers and/or developers stay the hell away or get out as soon as they can.

  • Cindy Bolero

    It’s hard to believe that after a decade, nothing has been done about it. Attempting professional use (Education, training, therapies, conferences, etc) of SecondLife/OpenSim, is often failure to launch. When it comes to new users and the interface, the usual response is “I don’t have time for this”. When its company or school time, there’s definitely no time to spend on unneccessary complexity.

    Success Step 1:
    How quickly new users can access, and are comfortable and the least confused.
    Hint: The least amount of directions/steps to get them where you want them, is Key.

    Success Step 2:
    Getting them to come back.
    Hint: Provide more experiences they want to come back for

    Success Step 3:
    New users bringing more new users because they were compelled to share with friends or colleagues.

    While continuing to lack that simpler intuitive interface, OpenSim will continue to mostly serve SecondLife refugees who are usually already proficient with the viewers. I don’t see a growth model. That’s why I’ve been developing with Unity3D since 2012. I’m currently building in SineSpace which is packed with Adam’s experience and supportive talented community users. I wouldn’t call them SL refugees, I would call them Innovators that are Making Headway.

  • Walter Betson

    Maybe if the core Opensim developer team wasn’t ignorant to the people at Firestorm there would be better cooperation. I’ve made more than one attempt to get the Firestorm Development team and the OpenSimulator development team to work together and it always ends up in a bonfire. The last time, Firestorm and the OpenSim dev team interacted, a bunch of name calling ensued after the Firestorm dev team didn’t react to a change that would have broken the viewer in SL and rather than work things out with the Firestorm dev team, the OpenSimulator dev team blasted a post out everywhere that Firestorm isn’t compatible with OpenSimulator and doesn’t support it (which couldn’t be further from the truth). Before anyone screams and balks at this remark, Jessica Lyon is MORE than accessible, and Firestorm is the only one of the viewers that actually had some Opensim default grids LISTED and has a version of the viewer tailored for OpenSimulator. Before pointing the finger at the viewer, how about fixing some of the 4 year old Mantis tickets that are out there that cause the platform to be unstable rather than worrying about the physics engine and adding shiny stuff. Make it stable first, actually address the old Mantis tickets, and don’t blame all of this on the viewer. Say what you will but SL seems to be doing quite well with the viewers that are on hand, they have 10’s of thousands of users online at any given time, so the viewer developers must be doing something right. If the dev team actually wants to make some headway with Firestorm’s Dev team, I’m happy to help assist in making that happen, but OpenSimulator has to realize that all the viewer’s priorities are where the people are… and that’s SL. So we have to be flexible on OUR end to accommodate that.

  • Maybe if the core Opensim developer team wasn’t ignorant to the people at Firestorm there would be better cooperation. I’ve made more than one attempt to get the Firestorm Development team and the OpenSimulator development team to work together and it always ends up in a bonfire. The last time, Firestorm and the OpenSim dev team interacted, a bunch of name calling ensued after the Firestorm dev team didn’t react to a change that would have broken the viewer in SL and rather than work things out with the Firestorm dev team, the OpenSimulator dev team blasted a post out everywhere that Firestorm isn’t compatible with OpenSimulator and doesn’t support it (which couldn’t be further from the truth). Before anyone screams and balks at this remark, Jessica Lyon is MORE than accessible, and Firestorm is the only one of the viewers that actually had some Opensim default grids LISTED and has a version of the viewer tailored for OpenSimulator. Before pointing the finger at the viewer, how about fixing some of the 4 year old Mantis tickets that are out there that cause the platform to be unstable rather than worrying about the physics engine and adding shiny stuff. Make it stable first, actually address the old Mantis tickets, and don’t blame all of this on the viewer. Say what you will but SL seems to be doing quite well with the viewers that are on hand, they have 10’s of thousands of users online at any given time, so the viewer developers must be doing something right. If the dev team actually wants to make some headway with Firestorm’s Dev team, I’m happy to help assist in making that happen, but OpenSimulator has to realize that all the viewer’s priorities are where the people are… and that’s SL. So we have to be flexible on OUR end to accommodate that.

  • Ok, this is just an opinion. I say that because many people who have invested so much time and effort can get really upset if someone questions their view of reality, or jokingly Virtual Reality! SecondLife has become old, OpenSimulator follows. New HUD is driving VR and AR now. Checkout WebVR, A-frame, WebXR. you are not going to get new blood. Young people barely know what SecondLife was. I embraced the Diva Distro when it came out, OSGrid, hosted my own virtual world for a time, but even now that the HyperGrids are gone and even though things linger around for a while, VR has gone elsewhere. I revived an old archive (2013), .oar of my original VW called Viradu the other day and have begun using my assets in WebVR… People, the browser has become the new viewer. Sincerely hope that helps.

  • WhiteStar Magic

    Anyone remember Pixie Viewer a Web based viewer prototype that worked for OpenSim ?

    How about Cloud Party which was very similar to OS/SL bought by Yahoo’s Mz Meyer and shuttered ? Also Web based Viewer which worked amazingly well, would have been HUGE with the new VR/AR hardware coming out.

    What happened to the Halcyon Fork (mentioned below somewhere), did that get chucked out with the foam dinner plates too ? or was it more infighting and territorialism that finally clobbered that too ?

    Consider the new generation of VR/AR glasses with their own OLED displays and not tethered to a PC or other device which are actually targeted to be quite affordable, already starting to appear from Google, Occulus and others… That’s another article for HG Business or someone else to pick up and run with… The future is here, yet many are holding onto the past for no good reason other than ear of change maybe…

    • Arielle

      Hypergrid functionality. That is why technically superior forks of Opensimulator lie in the dust while Opensim itself is still supported by an ever increasing population, in spite of being led by dev’s who at times seem to intentionally hold up progress of it. Even when many don’t use the ability to easily roll out their own grid or hypergrid enabled standalone, or even use the hypergrid itself, it still represents a freedom that no other platform out there offers to the masses. Being that the main architect for decentralized but connected virtual worlds is still in Opensim core, at least some of us still see it as holding on to the future rather than the past.

      • Han Held

        I’m not sure it’s the hypergrid itself, though that’s a part of it!

        With Opensim and SL you have a (reletively) easily accessible platform (at least one where most of the foibles are written about on the net somewhere) where you can log on, easily change your avatar into anything you want (that one is a biggy -most other VRs fall down on avatar customizability) and then begin to start creating inworld content without bothering about Blender or some other byzantine program.

        Also you don’t have to go out and buy specialized equipment …equipment that has very, very limited utility outside of VR …in order to use SL. If you have a desktop (or even a laptop), you can probably use that.

        The barrier to entry might be lowering to get into VR (don’t know, don’t care); but it is still doesn’t give as much “bang for your buck” as SL and Opensim do.

        I’m not sure that it ever will.