Returning to Second Life
arstechnica.com
arstechnica.com
I've set the games to email me when someone achieves a respectable score in one. I still occasionally get one of those emails. That's LSL (Linden Scripting Language) code that was compiled into bitcode years ago and the source hasn't been touched in any way since, and neither has the bitcode (or the script would've been reset and lost its memory - all the highscores are stored only in script memory) - and it still continues to run. This amazes me. That's the kind of stability many webhosts can only dream of.
Oh, the SLURL is on the bottom of these emails if you want to check out some really old games: http://slurl.com/secondlife/Stillman/65/81/24
Teen grid also lacked the ability to get items from the main grid so the average build quality of anything was pretty bad. I wonder how many of those builders went on to work in 3d animation or something similar. Pretty much everyone I know who used to play it either ended up in coding or some artistic career.
Fellow former TGer-
Alpha Zaius/Ming Chen, I think? (Andy Ortman and his brother - featured on the news back then)
I think they post here.
She recently wrote about an experience with LindenLabs that caused her to lose interest in developing within SecondLife: http://kathar.in/post/149846332570/lsl-ui
I was on the teen-grid, too. In my time there, I met many really gifted teen programmers. They partly inspired me to pursue computer science, and, years later, I've even had the unexpected pleasure of meeting some of them in person!
Augh. As someone who, back in the day, put significant effort into building up a semi-sane framework for UI construction in my own LSL scripts... that looks so much nicer than what I had to work with it's painful.
Maybe the code was recompiled (i think they 've switched to mono now) , but the script state is saved separately.
Then, I found my job at AWS through Second Life. [1]
Years later, the project became secondary to me. Unfortunately, there was no way to backup my creation, and I didn't want to give 300$/month to Linden Labs. (unfair artificial constraint, to be honest). I shut it down. They might still have a backup somewhere, who knows?
[0]: https://www.youtube.com/watch?v=TXPVJww9H5c
[1]: http://brunozzi.com/2008/05/22/how-i-got-hired-by-amazon.com...
And that's absolutely the case: users don't care that much, despite the teraflops inside our GPUs.
While things have changed in the almost 30 years since "Lessons from Habitat" [1] was written, one of the conclusions is still valid nowadays: the implementation is relatively unimportant.
The entire paper is a gem; worth checking it out. It has aged wonderfully well, and I never get tired of re-reading it over the years.
Survivor bias. I'm one of the ones that played with it for a few weeks, loved the whole idea of the place, and noped out when I couldn't get used to the incessant lag.
An SL user whose avatar is named Penny Patton has written extensively about such issues in her blog "The Digital Pasture"; her posts are worth reading.
I played in the era before sculpted prims were really a thing and everything was made by just stacking a ton together and using texturing tricks. Some of the stuff I built hit the link limit.
And that's neat, but they'd then be shrunken down to the point where you couldn't see them anyway. But it's what "sold" them, so in many ways it mirrors the bloated web dev we see today.
Avatar attachments such as jewelry aren't physics-simulated.
Waiting for the server to affirm your button press to move your character just makes the game feel instantly wrong and frustrating.
Remember: real people have spent real money on in-game items, you can't just rev the API and say the old items are gone. You can't do refunds either: players paid other players for items, ten years ago, using long-expired credit cards, sometimes from expired people.
www.inworldz.com
Here's how it worked:
1. I'd have avatars come meet me in my "studio" - a floating cube with very basic geometry that was way up in the sky away from other objects.
2. I'd use an app called OGLE (OpenGL Extractor) that was developed at Eyebeam to extract the avatar's geometry and textures.
3. I'd then load up the 3D capture (which would be a mess) and clean it up and make a bunch of manual modifications to make the 3D design print worthy.
4. I'd export the resulting image as VRML and then send it to my printing partner who would print them on "Sandstone" - plaster infused with color inkjet and something like superglue.
Here's a picture I found on Google of them: https://farm2.staticflickr.com/1102/1236026167_49aae2fc0b_b....
Speaking of 2007, remember how hyped motion controls were? Remember how excited we got when we saw people "bowling" and "boxing" on Wii Sports? Remember the thrill of booting up LoZ Twilight Princess for the first time? Remember that first sword swing that was blocked on screen even as your real-life arm kept swinging, undaunted but by a pithy vibration?
Even with VR that didn't cost an arm and a leg and ran Second Life at 90fps without trouble, the simple fact is that without realistic force feedback, every user will be constantly reminded that they are not actually flying a spaceship in zero G, they are not actually being double-teamed in a mass orgy, and they are not actually relaxing in mud bath at a sprawling spa.
True immersion is decades away. Until we have Star Trek holodecks or Metal Gear nanomachines, we are only fooling ourselves into believing that virtual reality is completely immersive.
We really ought to have a kind of Turing Test for VR to restrain the hype.
I remember other people feeling that way yes, but for myself and some others it felt like a big set a gimmicks. Motion control on the Wii wasn’t really motion control, it was wiggle control; I remember complaining years ago motion control would only be really interesting what it was 1:1.
I don’t feel that way about VR this time. Virtual reality is rough and it’s it’s in its early stages in terms of hardware, software, and the user experience, but it has so much potential. If anything VR right now reminds me of games back of the time of Wolfenstein and Doom, and I’m pretty excited. Having said that I’m not buying first generation VR gear, I put in my time with early computers. My generation two or three though I expected things are gonna be pretty damn interesting. I’m not a particularly optimistic person, but I am genuinely optimistic about this. I don’t think full immersion is required for VR and especially AR to change the game.
AR is a completely different beast. I don't need force feedback for apps like Measuring Tape to be astoundingly useful. I am quite hopeful at the near-future utilization of AR.
You can't even compare them to the wii. Playing Robo-recall on the oculus is like being in a john woo movie.
And with realistic force feedback, a bluescreen will tear your arm off at the socket. I think you're right about the timescale, but more because we can't write truly safe software.
I find this pretty surprising. I wonder who they are. I never got the appeal of Second Life but if this is accurate then it clearly appeals to a group of people.
It's a bit dated (2010), but Netflix has a highly engaging documentary[0] on exactly this question.
As for who they are. There is baby boomer after baby boomer when it comes to real users (at least 8 years ago when I played. The youngest of the boomers). It's also worth noting that when it comes to the typical MMO, they are really really bad at providing a social experience. So Second Life as a game that doesn't have the grind or ways to kill other players, does a really good job of drawing in lots of the people who will play a game just to talk (since once you provide ways for The stereotypical Achiever/Grinder or Player killer to control gameplay they tend to dominate the game). Also very few games offer the level of expression and modification.
In other groups I see a bot of creative folk, however janky it may be the primitive and scripting is easy to approach and gets people doing virtaul 3D visualizations pretty readily.
SL doesn't require anyone to dodge virtual bullets, they can just wander, socialize, create. Thats what draws me there.
New SL advances have introduced mesh objects into SL and allow the prim count (now Land Impact) to go down to near non existence so you can have more stuff in whatever virtual land you have. Frame rates vary but I think stability has been improving (not as many catastrophic sim-crossings, etc.)
OTOH a lot of people fed up with SL's limitations and extremely expensive land fees have moved to the open source opensimulator platform.
High Fidelity starts their world out as a boring flat plane. They should have started it out as good-looking natural terrain, with trees, shrubs, and grass. Use SpeedTree to generate a realistic looking world. That would set the standard for how good things should look. If people buy land and build things, the good-looking natural features will push them to do better work.
As people move around, they should wear the grass down a bit, and when there's enough wear, an NPC road-building crew moves in and builds a path. There should be an automated Department of Public Works which gradually installs public services.
And if an area is abandoned, grass should grow through cracks in the roads.
Amazon is coming out with a "Snow Crash" show. Somebody should have the Snow Crash metaverse running before it comes out.
[1] https://www.youtube.com/watch?time_continue=42&v=-ivL1DDwUK4
It's full of "For Lease" signs. There's a vast amount of unused real estate. Overpriced. Many billboards with space available. A huge area of empty sex clubs. Almost no avatars. Saw maybe six people in three hours.
Works fine, though. Joystick navigation is terrible, but that's probably fixable through configuration.
I think that eventually we will move back towards the open-world, user-created and maximally immersive metaverse in a huge way. A few things will make that happen. One is peer to peer distributed services, which I believe will eventually replace most centralized services. It will just become the way we do distribute code, content, and money. Sure, people have been talking about this for many years, but while the adoption and applications haven't reached wide popularity, they continue to grow steadily. This is just as much about social change as it is about technology. And groups of people change and adapt slowly.
Another enabling technology will be real-time path tracing or ray tracing etc. This will probably be enabled by better hardware and/or custom chips designed for global illumination. If we can combine this effectively with procedural generation, that can really reduce the bandwidth requirements and make custom, on-the-fly networked loading of diverse complex realistic environments much more practical.
Another thing that will move us back towards those old-school metaverse visions is better VR hardware. When we have widely available inexpensive high resolution, lightweight, untethered displays and convenient, high-fidelity input devices, VR will become the default computing mode for large numbers of people. Within a few years those things will come together.
One last element of the metaverse stories is AI, which is also advancing quickly. So I think we will see that sci-fi vision fully bloomed within the next decade at the most.
But how does it compare to something like SL where it seems you can create much more interesting scenarios?
The real value in buying Linden Lab used to be, in my opinion, the in-world currency called the Linden (or L$). For a long time, I thought that Linden Lab should've made the L$ available as an Internet-wide currency but now cryptocurrencies appear to be a better alternative.