animats

This is commonly done for better garage doors in SL. They move upward and the upper edge disappears. It's usually done by building the object maximum length, and then updating PRIM_SLICE and PRIM_POS on every update. Then the effect looks right, with the texture moving, rather than shrinking. If you shorten the object, the texture will compress. There are low-end garage doors in SL which do that, and they look bad.

That's a problem. The Firestorm welcome hub has gateways to other places. The WelcomeHub is a dead end.

Now there's a thought. I saw a flock of new bots appear at Welcome Hub, with names such as "grn00020", "grn00021" ... They immediately teleported out. More could be done to manage bot traffic. I have a little experience demo at my place in Vallone for dealing with avatars that just stand around forever. If an avatar stands around near a park bench and does nothing for a while, it's told "please have a seat" and is seated on the bench. This is just a demo, because you have to accept the land scope experience, which a bot will not do. If LL did that with one of their built-in experiences, with a one-hour timer, it would de-clutter the hubs. It's not ejection or banning, just seating. If someone was AFK, when they come back they'll just find themself seated out of the way. Automation which affects users should have a light touch.

Not sure. Here are some more. WelcomeHub, 3 avis stuck in cloud mode for at least 10 minutes. One had previously replied to messages from a mentor. The other two would not respond to chat. I saw another one stuck earlier today for six hours. The mentors are noticing this, but don't have any answer. I'm not sure if this is a new user creation problem, a bot problem, or a general problem. But seeing three new users stuck like this is not good. These avatars are solid, listed as nearby, have readable profiles, and can be sent IMs without message errors. This isn't the old ghost image left behind after a disconnect problem. Those don't show in the nearby list, and you can walk through them.

Some quotes from the book: Page 298: "Every metaverse platform that has been launched without game mechanics has failed to gain mass adoption". From a business perspective, he's probably right. We need games to get new users through the first week. LL needs to fix the pain points for games within SL. Page 298: "Something magical was lost when our focus on metaverse platforms moved away from live, immersive social creation, and we resigned ourselves to offline editing in 3D software programs." There's something to be said for that. In the early days of New Babbage, I'm told, people stood around and watched others build. Maybe SL should have a building construction system and a clothing design system in world. Ones with lots of templates to work from. Ones where you make walls or shirts, not pull vertices around. Much less general than Blender or Clo/Marvelous Designer and easier to use. Those are hard problems and there's not much open source software to mooch.A good start would be to provide a set of standard blanks (T-shirt, dress, pants, etc.) and a texture editor comparable to those ones for ordering T-shirts on line. (Yes, you can start with Robin Wood's old templates and use Photoshop or GIMP. That's still too hard for at least 80% of the potential user base.) My clothing skills are only good enough to make T-shirts and leggings for SL. Anyone from the SL clothing industry want to comment? Page 301: "We have explored how ultrarealistic human avatars can negatively shape the culture of a metaverse platform, arguably shifting Second Life in its later years from a highly imaginative creative platform to a heavily materialistic, often-toxic consumer driven experience (with some creativity still persisting on the sidelines.). This variety of avatar definitely did not grow Second Life's user base." That's an interesting subject. Roblox has been moving towards more realistic avatars, and they're now about up to pre-mesh SL avatars in look. (There's some parent opposition to this, their average user is age 13.) VRchat uses reasonably realistic avatars, although they keep the complexity down to keep the VR frame rate up. Fortnite is deliberately cartoony, but that's because it's a very violent game that would not be a kids game at the realism levels of, say, Red Dead Redemption or one of the Tom Clancy games. I tend to think that SL's problem is that it's in the uncanny valley - avatars are just good enough that we realize how bad they are. That's where Hollywood was about 20 years ago. Go watch the Polar Express trailer. Hollywood is totally past that, and so is much of gaming. SL is stuck just before subsurface scattering, which means skin is in the dead to plastic range. I suspect this drives the trend towards excessive tattooing.

That reply was a bit of a surprise. I'm still trying to figure out why I'm seeing new users stuck at pink/white cloud mode forever. There's one at WelcomeHub now who's been stuck there for at least six hours. There was one on a street in Charlesville last night. When I've seen avatars in this state, they show up in the nearby avatar list, don't move or reply to chat, and stay there for hours. This may be some failed login/logout/teleport state. Something is broken, and it seems to hit new users.

A new pain point: I've been seeing some new users stuck in pink/white cloud mode forever. Reported this in the JIRA. Got back the reply "If one of that avatar's 4 required wearables (eyes, hair, shape, skin) is missing or corrupt, then all viewers will see the avatar as a cloud. This is expected behavior." That shouldn't be happening to new users. But I've seen it three times in the last two days, for users with age 3 days or less. I wonder if one of the starter avatars is damaged. (Confusing the issue is that it takes way too long to get out of pink cloud mode in busy regions. I think this is the BOM baking server falling behind, but I'm not sure.)

In "A Complaint is a Gift", a book for retailers, it says that for each customer who complains, there are ten that said nothing and will never come back.

I haven't said much about region crossings in a while. Server side is still broken. One of the senior Lindens was annoyed at me for bringing it up at Server User Group so often. He said they were working on it. That was in 2022. It's still broken. I'm tired of the excuses from LL on this. I've heard five years of excuses for failure now. In the context of new users, if driving worked reliably, we could have mainland tours for new users. But vehicles with multiple passengers work very badly. Sometimes I try to get people interested in driving. I can't put them in a vehicle and drive around with me without going through a checklist. Low-complexity avatar? Check. Good ping time? Check. Enough network bandwidth? Check. Enough graphics power? Check. Good vehicle? Check. OK, we're good to drive around mainland. I've started people driving with free cars from one of the shops on Robin Loop. I've seen them hit the region crossing at Neumogen, have the avatar get stuck, and the car kept going, running off the road and into a building. I had someone on the back of my bike recently, and their high-complexity avatar came off at the second region crossing. LL people, this is pathetic. If you don't get this fixed before mobile goes live, you will be laughed at by major reviewers. You cannot launch into the mass market/gamer space like this. You only get one chance to make a bad first impression. You don't think so? Go watch and read game reviews. The tolerance for immersion-breaking failures is zero.

I've been over there twice, and haven't seen problems. Just the usual curse of SL - a group of people standing around, not moving or talking.

Yes. That's very much a theme of the book, repeated at least four times. It's something Rosedale disagreed with, but the book author thinks he's coming around. The author points out that all the virtual worlds with ongoing growth have a sizable game component. LL is trying. "WelcomeHub" (no space, and I'd love to know if that's a typo or a design decision) has a laser tag area. So, at last, new users have something to shoot at. I'd suggest a racetrack in the welcome regions as the next game.

Hm. I'm not sure why all those Chromium/Dullihan processes are so big on my system. It may be a problem with running under Wine emulation. Looks like you get as many as I do, but they're smaller. Or maybe the size listed under Windows doesn't count the DLLs.

I've been running the materials viewer release candidate. Right now, I'm stationary, in London City, just watching the crowd. Memory use keeps climbing, at about 1 GB/minute. I previously reached the point where memory was thrashing on a 32 GB computer. The problem is the Chrome browser for media content. Each media stream launches its very own copy. I was getting eight copies each of CRGpuMain, CRBrowserMain, CRUtilityMain, and CrRendererMain. Each full set is about 0.7GB. There are no visible prims playing media, just audio. Turning off media playing caused about half of those to exit and reduced memory usage to a sane level. Unclear why it keeps four spare copies around. This is the default mode for the LL viewer. Uh oh. (I'm running the LL viewer in the Wine emulator, emulating Windows under Linux. Someone please try this on an authentic Microsoft Windows system. Thanks.)

He discusses why SL hit a ceiling while IMVU, VRchat, and Roblox continue to grow. That's the important takeaway here. What have they got that SL doesn't have? Maybe SL needs more game-like content to overcome the huge loss of new users through onboarding failure. After reading that book, I went over to "WelcomeHub" (no space) yesterday, to see how new users were faring. It's not going well. When WelcomeHub first opened, they had maybe five "mentors" standing around. Yesterday, zero. Yes, it's a US holiday. There were about a dozen new users in the welcome hub. A few were willing and able to talk. One was struggling with lag, and using an under-powered MacBook. I told her how to turn down the performance slider, which got her unstuck. One had clothing questions, and I told her how to save her current outfit with "Save As" so she could quickly recover from wardrobe malfunctions. Couldn't help her with Lelutka vs Legacy, but she was far enough along to ask that question, and will work it out. One new user came in completely nude and teleported out quickly. Two were stuck in "pink cloud" mode for over 10 minutes. (Other users saw the same users stuck in "pink cloud" mode. This may be a server side problem.) One user, new but well-dressed, was "Away". One user had found the laser tag arena and had a gun. He'd had fun with laser tag and was looking for the next thing to do. I suggested some combat regions. Four avatars had lower case name which seemed to be composed of two concatenated random words. Those users never moved or talked. Those may be new bots. Nobody was having fun. That's the new, improved onboarding.

"Making a Metaverse That Matters"

Wagner James Au, who was LL's "embedded journalist" in the early days, has written a new book about metaverses which covers SL's history at length. He covers the onboarding and mass appeal problem at length. He quotes an analyst from 2010: "Second Life is an extreme example of how you can develop a very happy group of customers, and still fail miserably at reaching a wider audience." Worth a read. I assume all the Lindens have read it already.

There are many shops like that in New Babbage, but few people live in the flats above them. Bellessaria, a suburb, is less dense than a town would be. In town, you don't have the big side yards. A town would have mostly 1024 sized lots, with maybe a 2048 or two for the big house on the hill, and some 512s on the other side of the railroad tracks. Open space regions use 1/8 the resources of a full region. You can't put much in them, but a road through farmland, hills, or woods is within the limits. One or two non-revenue buffer regions for a town of 2 to 5 full paid regions isn't too much. It's an operating cost like water regions.

I get messages from SL occasionally. Best one recently is from one of my NPCs: [10:20:23] (Testi v22.0a (Animats Corsica)) Inbound threat [777] Cyclone V2 - Funky Neon Pink at <184.17490, 164.56140, 46.97566> vec to threat <-7.22083, -6.43941, 0.00000> ETA 2.427503s approach rate: 3.985591 [10:20:23] (Testi v22.0a (Animats Corsica)) Avoiding done. Someone is driving around my parking lot, and Testi got out of the way. [10:20:25] (Testi v22.0a (Animats Corsica)) Inbound threat [777] Cyclone V2 - Funky Neon Pink at <184.40810, 168.54190, 47.15474> ETA 3.861451s. [10:20:25] (Testi v22.0a (Animats Corsica)) Avoiding done. Happened again, 2 seconds later. This is OK; the NPCs are supposed to get out of the way, so they can patrol parking lots. [10:20:30] Inbound threat [777] Cyclone V2 - Funky Neon Pink at <189.40550, 175.92800, 46.94649> vec to threat <-5.41638, -1.61600, 0.00000> ETA 1.000000s approach rate: 7.318527 [10:20:30] (Testi v22.0a (Animats Corsica)) No walkable below avoid dest at <189.83510, 179.31030, 46.02652> [10:20:30] (Testi v22.0a (Animats Corsica)) Unable to escape incoming threat. [10:20:30] (Testi v22.0a (Animats Corsica)) Testi v22.0a (Animats Corsica) in trouble at Celchu <191.05030, 175.23730, 46.60492>: bhvavoid: Unable to avoid threat by any path from <191.05030, 175.23730, 46.02652> Griefing alert! Someone is driving a Cyclone V2 at my NPC at 7 meters per second, and the NPC is in a spot where they can't dodge. I get a message to my phone. The NPCs say "Ow!" if hit by a vehicle. They become immovable for a while, and the vehicle bounces off. I'll go see what the problem is if I'm not busy, or if this happens repeatedly. I haven't yet had to ban someone. It's a low-key anti-griefing measure.

Bellessaria started as American suburbia. Mostly, it still is. Here's an idea - small-town America. Suppose SL had small towns. Rome, Georgia. Towns would be 2 to 5 regions. One region is too cramped; more than 5 are too big for everyone to know everyone else. Roleplay regions tend to be best in the 2-5 region range. There would be a "main street" shopping street, of course. Houses would mostly be on side streets. Main street would have a cafe or two and some shopping, to give people a place to hang out and socialize. Towns wouldn't be isolated. Towns would be connected by an open space region of forest, farmland, plain, or hills, so you can drive from town to town, going through each main street. This gives towns the feeling of being part of a bigger world. A buffer region might look like this. You're out of town, you're on the road, and you can't see the next town yet. Towns could be quite different from each other. Some might look like Disney's version of America. Some might have heavy industry. Some might look like Crack Den. Some might be college towns. The buffer regions keep this from looking like mainland's jumble. Existing themed regions could connect with the town grid. This should be encouraged. Comments?

PBR needs to be driven by people with more artistic talent than me. It's easy to get tacky effects, and hard to get subtle ones. Now we can do the tacky goldish escalator from Trump Tower. This is "tarnished copper" base color plus "chrome" material. The real world one is even shinier and tackier, but at some point you have to not turn the sliders up to 11. Haven't been able to get a good brushed stainless steel look yet, though. That's more subtle, and harder.

More shinyness. Chroming one of my bikes at Materials1. This is a nice effect if you don't overdo it. Now I need to add a light layer of road dirt. Silly environment, but it's reflected properly. They're low-rez slow mirrors, but they are mirrors. Most SL vehicles are going to need a refresh. Too bad we don't have the glTF clear-coat layer so we could have modern paint jobs. This chrome material has no textures. Here's the actual glTF file I wrote: { "materials" : [ { "pbrMetallicRoughness": { "baseColorFactor": [ 1.000, 1.0, 1.0, 1.0 ], "metallicFactor": 0.9, "roughnessFactor": 0.05 } } ], "asset" : { "version" : "2.0" } } That's the minimum glTF for an SL material. There's no base color, so the base color from the texture, if any, will be applied. There probably should be some standard no-texture PBR materials in SL's LIbrary. You don't need many, because you can apply a color, diffuse texture, and normals to that PBR material. If the PBR material doesn't have those, the old ones are used. A standard set of simple materials like this - chrome, painted metal, brick, concrete, stone, etc.- would cover many use cases. That would improve performance, too, because each PBR material is an extra HTTP fetch. Legacy materials are loaded in bulk when you enter a region. This is all cute, but the thing we really need is subsurface scattering. That's what makes skin look like skin. It's the effect that when light hits your skin, a little bit of reddish light comes out nearby. It's subtle, but humans seem to be evolved to unconsciously recognize that as "alive". Without that, the choices for skin are along the axis that runs from 'dead" to "plastic". I sometimes wonder if the tendency to overdo tattoos and fur in SL is compensation for the bad skin rendering.

You can turn off terrain rendering, but it's not very useful except for debugging purposes, and for finding objects that somehow became stuck underground. At Port Babbage/0/0/0, under the sea floor, there is a lost liquor bottle.

We have mirrors now! Well, sort of. Someone set this up this demo at Materials1 on Aditi. After a few seconds, the reflection probe system updates the scene and you get a low-rez mirror. If you get too close, you don't get reflected, though. The pink pedestal and shiny model never appear in the mirror. It's not really for mirrors. It's to improve indoor ambience.

I don't think that's what SL prefers, I think it's what users have to do because of input latency to move your avatar, lag when someone enters a sim/server, etc. Right. My main point here is that, once LL sees someone else doing something good in their space, they'll consider doing it too. So the way to get LL to improve the product is to show them that something that needs to be done for SL has already been done successfully by someone else. Over in metaverse land, there really is some modest progress. The NFT clown car has crashed and the wreck is now in the clearing stages. Facebook/Oculus/Meta has now failed three times and given up. Magic Leap went bust. The people who were making all the noise are now out of the picture. Others continue to soldier on - M2, RP1, Roblox, Unreal Engine - and are getting real work done. The serious development efforts are worth following.