Fascinating conversation in response to Luther Weymann's somber analysis of Second Life's prospects under new ownership, featuring many comments by readers who seem to be longtime veterans of the game/tech industry, or at least write with that level of expertise.
Among them is Russell Skyther, who wrote this really sharp breakdown of what it would take to upgrade Second Life. It's partly a response to an earlier comment arguing that "Updating SL isn't really all that hard cause they have done it is how we got mesh and rigged mesh and now animesh. They can continue to keep layering new content and it'll work just fine."
Skyther argues otherwise:
The issue isn't that you can't update Second Life's code. or even rewrite it from scratch. BOTH have been done, see: OpenSimulator, an Open-Source clone of SL functionality that's close enough [that] most 3rd party clients support it out of the box.
But ability isn't the same thing as willingness. And far from being 'not all that hard' as the above comment states, each one of the examples features: Mesh, Rigged, animesh, and now Bakes on Mesh are all hacks onto an engine that wasn't designed for it. Now you can do that, and they did. But that incurs a serious performance penalty unless you redo the entire graphics pipeline (which they didn't). That's what this guy's talking about. It's not that you can't update or rewrite SL, it's that it's not cost effective to do so on a level that would give you acceptable performance.
To better explain SL's codebase issues. This is what I would consider a minimum for a true SL-like game in 2020 (VR Chat isn't really SL-like, but more of a Sansar that worked):
- Graphics pipeline efficient enough to give 50 FPS *minimum* regardless of how much content is on screen. This can be done.
- True 'client side' code in its scripting language: Pathfinding shouldn't be the server's job. Collisions shouldn't be, either.
- Modern procedural animation engine: Because multi-character animations of all sorts show just how bad SL deals with animations. Short guy hugging taller girl aught to show just how bad that is for the curious.
- Official world scale: On the above, SL isn't 1:1 with the real world on most servers. and it's not consistent what scale folks use, usually 1.25:1 I do believe. but it means you've got to mess with half the stuff you buy to get it looking as it should.
- Modern shard-based Servers/64 bit worldspace/way more mainland: The big thing SL has, even over it's Marketplace, is the ability to, at least in theory, virtually travel the virtual world from one end to the other. The Marketplace is great, but actually damaged the viability of this unique trait just by it's existence.
- There's also not enough mainland, the roads and airways aren't contiguous on said mainland anyway. Even when they are, going between servers is incredibly janky at times. You could ditch the overworld for a list of servers like VRChat uses and nobody would notice. That's a shame.
Modern content-streaming pipeline that means 9GB of cache is actually enough. This is where most of the lost frames go to die, and nothing that I've just said above would work unless this whole cache system where overhauled as well.
Is all of this doable? Yes. Will it happen? Probably not, unless it's OpenSim deciding to make it's own from-scratch Viewer.
Someone, at some point, will come along and make an actual SL competitor with Unreal, Unity, or some other recent game engine. In that way, SL's legacy would live on, but at 60 FPS with reasonable lag this time. I really hope that competitor is SL 2.0, but I'm not terribly optimistic.
There's arguably multiple virtual worlds that could be seen as potential competitors to SL -- just to name two, built on the Unreal engine, there's Core, and built on Unity, there's Sinespace. Then again, because they are built on modern next gen graphics engines, it's not quite accurate to say either really competes with SL -- anymore than it'd be accurate to say Elder Scrolls Online is competing with Ultima Online.
Pictured: The Black Dragon viewer, basically a user-driven update to Second Life graphics.
In short, it's what I've been saying the entire time.
SL's engine could never handle mesh because it was never built to handle it in the first place. It was built to handle prim-based tools.
For SL 2.0 to flourish, and it's not coming from LL, people need to start looking at Sinespace, which uses the Unity Engine, or even Core, via the Unreal Engine. Both engines are regularly updated and can handle what SL never could... and that's all the Mesh (rigged or unrigged), Animesh, Bakes on Mesh, and hell... even handle clothing physics once that gets fixed.
I can speak from my experiences in Sinespace thus far, and I can tell you this, once things start to come together from its developers (Adam and crew), mouths will drop. Granted, they're still getting things together, adding what needs to be added, and fixing the things that need to be fixed when nagged to death about them *whistles innocently*. Sinespace even has a weekly Tech Office meet up that's streamed via Twitch and Youtube that you can watch, or attend live at The Assembly Region; and they do talk about what they're up to while fielding questions from everyone at the same time.
Folks, the bottom line is that no matter how many hacks LL may put into SL to freshen it up, it's only a matter of time before a virtual world with an actual game engine comes up and surpasses/exceeds it.
Posted by: Alicia | Tuesday, July 28, 2020 at 04:59 AM
Doubtful anyone could come along and give you what SL actually gives you. I am in SL to create environments from the pieces other people have created. There is no where even close that could provide me the content base to create the things I do.
Posted by: Jesse Sondirra | Tuesday, July 28, 2020 at 12:37 PM
"Doubtful anyone could come along and give you what SL actually gives you."
Stop right there.
What LL has given its users and creators over the years is a convoluted mess that they'll never admit, nor own up to. Second, games with actual engines can easily surpass anything LL can ever do with SL when given that chance to do so.
So you create environments that are approximately what, 2-3 blocks via a 256x256m sim square? Oh, and yes, I actually did the measurements via Google Maps. Yes, you can make some stunning sims on a 256x256m area, but what happens when you run out of that space... or prims for that matter. You have to remember, the scaling with avis, etc., is another convoluted mess as people in SL want big yards and space.
However, you're also forgetting all the mesh in that space that will harm everyone's latency. Not even a high-speed connection, 200, 300, etc., can ever offset this. This, in computer terms, is what's known as a bottleneck. Everyone in SL is bottlenecked because of it. That bottleneck comes from an old, outdated engine and more. Simply placing hacks and band-aids on something only fixes it for so long before everyone starts to notice the problems. Simply moving things to different servers, or even the Cloud, won't fix the problems. It's just another band-aid.
As for the last sentence, "There is no where even close that could provide me the content base to create the things I do." Never, say never.
Posted by: Alicia | Tuesday, July 28, 2020 at 01:13 PM
I just recently returned to Second Life after 3 years of playing The Elder Scrolls Online(fun game). I'll say this about the performance of SL vs that of ESO: I really don't mind SL's issues, lag, whatever.
In ESO I PvPed (player vs player) and it was a highly competitive environment. Death is frowned upon to say the very least. Killing someone involved tbagging, hours of arguing, 1v1's, and feuds that lasted YEARS.
I quit ESO and came back to SL due to PERFORMANCE. ESO is a game that the better you got at it, the worse it played. Tiny often insignificant bugs, unnoticed by newer players completely ruined the experience.
I'm a Grade 2 Grand Overlord in ESO right. It took 21 months to achieve that Alliance War Rank (the highest) on my PvP toon. It takes untold hours of theory-crafting (thinking up ways to create a viable "build" using all the tools available class, race, gear etc), testing, dueling, and actual FIGHTING in Cyrodill (PvP zone) to actually get good at that game. THEN...then, because of constant quarterly changes some bug gets introduced and you literally can not play the game as well as you could the day before an update was issued. I quit.
Occasionally, some bug or lag or SOMETHING will happen in SL and I'll go into an instant white hot rage (cause a bug will get you killed in ESO lol). Yet, I'll immediately calm down because it's just not that serious in SL. I'm not going to end up cussing a guy out for the next 3 hours because of something that happened (I died) because of a bug in SL. It's a bit of a relief. I'm calmer. And...I'm very happy to be back in SL.
If performance is an issue in Second Life, I just don't see it as something I can't live with.
Posted by: Jumpman Lane | Sunday, August 02, 2020 at 02:34 AM