Jump to content

animats

Resident
  • Content Count

    2,225
  • Joined

  • Last visited

Posts posted by animats


  1. (A new user recently asked me for advice on this, and I realized there's no good introduction to driving around Second Life. So, here it is.)

    Driving in Second Life frustrates some users. Often for the wrong reasons. It's not difficult, but some advice on getting started helps.

    First, you need a car.

    (I sell motorcycles, so I'm going to concentrate on cars for this, so as not to advertise.) A good place to get a demo car is Burns sim, in Heterocera. There are three competing car builders in that area, and they all have demo car rezzers and access to Linden roads. So you can try different vehicles from different builders. They're all modern SL road vehicles and drive well. Any of the demo cars will work for as long as you stay in the car, so you can tour all of Heterocera for free if you want. This will give you a sense of how a good vehicle in SL behaves.

    robinloop.jpg.d60f0c175964b944c41889be8586b0cf.jpg

    Robin Loop in Heterocera. Car builders are concentrated near the southeast, in Burns and Neumogen. To drive Robin Loop, head east and make left turns when you have a choice.

    This area is on a road which is part of Robin Loop, a loop about 3km around. It's a good place to practice driving.

    Most SL vehicles steer with arrow keys. Steering control is thus rather limited. Many vehicles in SL are tuned for racing on racetracks. These go fast, turn tightly, and are good on racetrack sims with wide tracks. They're hard to keep on two-lane Linden roads, let alone in lane. Many don't handle road bumps or region crossings well. Road cars don't turn as tightly, are expected to deal with moderate road problems, and should deal with region crossings properly. If you try to drive a racetrack-tuned car on Linden roads, you'll soon become frustrated. If you want to try a race car, go to a racetrack sim. Most have demo rezzers.

    There are hundreds, maybe thousands, of different vehicles available in SL. Some are great; some are awful. Some that look really good are totally undriveable. Try before you buy.

    How to drive.

    Left and right arrow to steer, up arrow to accelerate, down arrow to brake. That's the only thing that's standard. Most cars have "gears", to set the speed range. Shifting is usually page up/page down or shift-right arrow/shift left arrow. Some vehicles have to be shifted to reverse to back up; some don't. There's no standard on this.

    Get a demo vehicle and drive slowly around Robin Loop. It's a nice drive. Things to watch for:

    • There's a place where traffic cones block part of the road. Avoid those. (Not enough prim capacity on a tiny corner of the road that's in another sim.)
    • There's a place where a flat road transitions abruptly to a steep hill. Approach that transition slowly or your car may bounce sideways.
    • There's a T intersection at the bottom of a hill. Approach slowly or you will end up in the lake. Turn left to stay on the loop.

    One trip around Robin Loop will introduce you to most of the problems of SL driving in only a few minutes. So it's a good place to practice. As you get better, try higher speeds. You can drive in SL at moderate highway speeds; you don't have to creep along.

    In general, try to understeer. Quick key presses on left and right arrows; rarely hold them down. If you oversteer and have to correct back, you're overcontrolling; back off a bit.

    Lag is a problem. You press an arrow key, and, after a visible delay, something happens. That's why you need to understeer and only make about one steering correction per second. This is weird at first, then tolerable, then instinctive.

    Have fun driving around and sightseeing. Heterocera has enough roads that you can drive for hours.

    Region crossings

    With a modern SL vehicle, the vehicle handles most region crossing problems. You'll have a brief stop at each region crossing, and there may be a moment when things look wrong visually, but you should come out of the region crossing at the angle and speed you went in. The vehicles for sale in the Burns area all seem to handle this properly.  Many older vehicles on Marketplace don't. There's folklore about crossing region boundaries at a right angle, and avoiding region corners. That's the vehicle's problem today, not the driver's. If you're losing control at region crossings, find a better vehicle.

    Most modern vehicles will slow you way down if you are getting close to the corner of four regions. Hitting a region corner fast will often cause a failure in SL and you will usually have to log out. Better vehicles handle this by slowing you down automatically.

    Expect to have about one total region crossing failure per hour of driving. These usually require logging out and logging back in. Known SL server bug. Biggest frustration with SL vehicles.

    Rough terrain

    Some cars can't go over a curb without getting stuck. Some can drive up stairs. How much off-road capability a vehicle has varies widely. Usually, this doesn't matter much for road driving. However, due to an error in road construction in Bellesaria, most roads there are impassible by cars that can't climb a vertical curb.

    Places to go

    • Heterocera has a good road network.
    • Sansara has a road network, but it's not well connected and the older roads, in the western section, are bumpy and have abrupt turns.
    • Corsica has a very nice "Circuit de Corse", with frequent, well-marked rez zones.
    • Bellesaria  is very driveable, with good rez zones. Watch for bicycles and horses.

    Things to do

    • Drivers of SL group. Weekly rallies, guided by a clever HUD. There's usually one every Saturday at noon and midnight SLT.
    • Get The Freight Out (GTFO) Grid-wide driving game, with tasks, deliveries, and scoring.

    openroad2-for-marketplace.png.79c416d5a3cf39b4a9bb8a992131b6d5.png

    There's a whole world out there.

    • Like 13
    • Thanks 14

  2. On 8/4/2019 at 6:59 PM, Wulfie Reanimator said:

    Second Life has far too much technical debt to survive into the future. VR is not for Second Life and vice versa, on a very technical level.

    It would be easier to fix the technical debt problem than to get a critical mass of users on a new platform. See: Sansar, High Fidelity, SineSpace, Sominium Space, Worlds Adrift... None got to a profitably large user base.

    The technical debt doesn't get fixed because SL has a tiny dev team.


  3. Looks like a nice build. Looking forward to visiting.

    Roleplay sims often need a full cast to work. It's hard to maintain interest. Maybe have active hours and days, and encourage people to show up then.

    I like Cocoon, which is a very well built futuristic roleplay sim. There are roles, factions, character types, and an in-game currency. That sim needs a minimum population to work. Same for Sanctuary. Hoodlum. Not sure how Crack Den is getting along. All those places have complicated rules, some have damage meters, and there are ongoing story lines. If you visit any of them at a random time, they will be almost empty.

     


  4. 10 hours ago, Nika Talaj said:

    Old SLRR trains used llSetPos based movement, and some had many primmy cars, depending on who made them.  Newer technologies exist.  Low-prim vehicles help a lot; a train or trolley system may not need a long lineup of cars following the tracks. 

    I made a Key Framed Motion ferry system, it did not impact region performance in any noticeable way.  Not physical, zero collisions.

    Janet's Viking sim, Folkvang, has keyframed trains, carts, and animals moving all over the place. There are switches, signals, loading and unloading stations, all running automatically. It's all very smooth. Good example to study if you're thinking of implementing trains.

    With overloaded sims being the new normal in SL, keyframe animation seems to be the way to go. Scripts which are constantly making corrections to keep things on track break in overloaded sims. Overload messes up pathfinding and SLRR trains.

    • Thanks 1

  5. It's a general purpose library for managing push button touch events. By itself, it won't do anything. You need other scripts to make it work. Those other scripts receive link messages from this script and send link messages to it. They tell it which buttons do which function at startup. Then, this script sends a link message when you push the button.

    All those string handling functions are to pack up multiple parameters into a text string to be sent as a link message. I'd use the JSON functions for that, since they're built in, but this code is old enough to predate those.

    This is part of some set of scripts. You only have one piece here.

    This is complicated for what it does. I've written bike scripts, and button management does not require this kind of complexity.

     


  6. Need some shoe or boot textures for system avis. I need actual texture files, because this is for animesh. Classic old style. A basic sneaker or boot will do. Thanks.

    lowertexture.png.8097445ef5bbebc2d6c9309ec5991c30.png

    Classic "Lower" format. Just need some shoes; have other garments. Not heels; sneakers or similar. Thanks.


  7. 2 hours ago, Mollymews said:

    it always comes back to how many avatars can fit in a shard/region without killing the client viewers

    There are ways to scale that up. It works better if you control the environment. Suppose that, in a game, you wanted to show 50,000 people in a stadium. You could render each section of the stadium on a server, and feed the distant sections to each viewer as video. Only the nearest section or two would render locally. It's a form of impostor generation.

    Here's a tech demo of a 30,000 person crowd, from 11 years ago.

    This was real-time rendered on a modest PC by today's standards. Yes, they all have the same clothes. Today we have more GPU memory.

    Making this work in a more general environment like Second Life's is harder, but not fundamentally impossible.

    SL doesn't use impostors enough. I've written on this before. It's just way too much for SL's undersized dev team to tackle.

    • Like 2

  8. 3 hours ago, Vanity Fair said:

    More information on Tim Sweeney's vision of the metaverse is here (he gave a presentation at SIGGRAPH, the annual computer graphics conference in Los Angeles):

    https://ryanschultz.com/2019/07/30/is-the-metaverse-going-to-look-like-fortnite-kent-bye-reports-on-tim-sweeneys-siggraph-talk/

    Interesting. He has a SL/Metaverse type vision.

    "Need virtual worlds to scale beyond a 200 players on a shard. Need 1 shared world w EVERYONE. Needs a programming environment to scale to unlimited sized. Not single thread C++. Large-scale concurrency w safe transactions that are consistent, durable, isolated."

    "A viable Metaverse is going to need a successful economy so that creators can make a living, which is absolutely essential. We need a rich set of different economic models. The app store with microtransactions is merely one model. Ad models are dysfunctional."

    So he gets both the technology and the business model.

    He also has a personal net worth of $7.2 billion, heads the company that owns Fortnite and Unreal Engine, and got 10 million people online watching an event in Fortnite. So if he wants to do this, he has the tools.

    (Then again, Zuckerberg made Facebook come out with Facebook Spaces, which is generally agreed to be a dud.)


  9. 1 hour ago, Bloodsong Termagant said:

    i am trying to calculate the angle of a slope between two points.

    Unclear what that means. Two points do not define an angle. It takes three points.

    You want ground slope, right. OK.
     

    vector p1; // point on ground under avatar
    vector p2; // point on ground ahead of avatar
    ....
    vector movedir = llVecNorm(p2-p1);	// direction moving between points
    float zlen = movedir.z;			// height of triangle
    float xylen = llVecMag(<movedir.x, movedir.y,0>); // base of triangle 
    float slopeangledeg = RAD_TO_DEG*llATan2(zlen, xylen); // ground slope as an angle in degrees.
    //	Traditional slope (1 m rise in 10 m horizontal => 0.1 slope) if you want that	
    float slope = zlen / xylen;		// but will divide by 0 if vertical.

     


  10. 1 hour ago, RaeLeeH said:

    I'm just going to slip in here while the adults keep talking over my head to answer this original question.

    Just like the quote, I am at present trying to learn Blender 2.80 and it's freaking overwhelming,

    On that note, is there a reasonable guide to read if you know Blender 2.79 and have to convert to Blender 2.80? Not a video, something where you can look up stuff.


  11. It's more successful than High Fidelity, Sansar, and Sinespace put together. About 7,500 paying users. Enough to make it go, not enough to make it great.

    The technology isn't bad. It's using Unity 5, with physically based rendering running in multiple threads in the viewer. If we had that in SL, SL would look much better. They have some nice building and UI features from which SL could learn. Yes, it's a sex sim. Not a clueless sex sim, though. There's some good engineering.

    Being a vehicle builder, I look at that sit target system and think "How could that be extended to vehicles". Suppose you could add a "car" behavior outline to a static vehicle model, get the wheels to line up, get the seats to line up, and drive away. If you need a door, you'd drag a door behavior to a static door, align the hinge, and you'd have a door. They have a "snap to" magnet option in their World Editor, so you can align objects, a basic capability SL lacks.

    Simple in-world building plus external parametric building (resize stairs, get more steps, not bigger ones) could be a powerful and usable combination. SineSpace has some of that. Many users complain that building in SL is too hard. The usual answer is "suck it up and learn to use Blender". That doesn't really scale.

    Maybe LL should buy 3DXchat.

    • Like 3

  12. 22 minutes ago, KanryDrago said:

    I can tell you 3dx does not in anyway have all of that. Every 6th avatar looks identical because it follows the mmo format of chose from these 6 heads..... It has a choice of about 6 or 7 for everything

    They did check off the items that Penny Patton listed for a successful virtual world. In each area, they have the basics.

    They don't have a content marketplace and permission system. If you put something in world, anybody can copy it. So there's no creator business model.

    There's one other item that Philip Rosedale has mentioned but Penny did not - crowds. SL doesn't do crowds very well. That's a big problem. If you make a popular place in SL, the sim chokes. The London City people now show up at every Server User Group meeting. With one of the busiest sims in SL, they have a big interest in performance.

    • Like 1

  13. 1 hour ago, Penny Patton said:

    ... virtual world developers all seem to overlook the key elements that made SL as successful as it's managed to be.

    1. Robust avatar customization. (No, not just the user created content you can decorate your avatar with, but the appearance editor itself.)
    2. In-world content creation. (Not only did simple content creation tools help make it easy for people to jump in to content creation, but it also made creating content a social feature. Remember sandbox building contests? When Sandboxes were always filled with people meeting and learning from each other in-world?)
    3. Land ownership. (Avatar customization aside, this is one of the stronger ways to get people invested in your platform.)
    4. Making money. (See above.)
    5. Allowing adult content. (People like the sex.)

    Amusingly, there's a virtual world which has all that - 3DXChat. It started as just a sex sim. Then they added building. Then users started building and visiting each others places, instead of paying for sex like they were supposed to.

    vtzSyIh.png

    Building in 3DXchat is with prims, plus a library of prebuilt objects. It's more flexible than The Sims, but less flexible than SL. The pre-built objects may be parametric, like SineSpace, but I'm not sure.

    3dxchatpositionmarkers.thumb.png.77b2dac4210e965f5706c048155ce509.png

    A bit of the World Editor. This shows a neat solution to the sit target problem. The "person with yardstick" icon brings up the sittable patterns. Here, the builder just built a couch. It's not sittable yet. They have to select a suitable sit target pattern, drag it to the couch, and resize it to match. Then characters can use the couch. The available sit patterns are shown. SL could learn from this.

     

    • Like 4
    • Thanks 1

  14. It's expensive. $49.95 per month, or $419.40 per year. It's not too clear from the site, but it seems you could be in SL 24/7 for that. Bright Canopy charges $17 per month, but you only get 10 hours.

    It would be interesting to try SL from a streaming server located very close to, or in, the SL data center in Phoenix, AZ. That would reduce the ping time to a very low value.

    There's a school of thought that streaming gaming is the future. That's what Google's Stadia is all about. It takes more bandwidth to get SL's content to the viewer than a video stream requires.


  15. 7 hours ago, Qie Niangao said:

    Any newly acquired land (acquired by any means, even merely deeding or buying back group-owned land) would become subject to the "Bellisseria rules" if that land had any point of adjacency to Linden Protected Land. So all waterfront, road and rail frontage, etc., would become "explorer tolerant" the next time it changed ownership.

    Now that's a really good idea. That would start to fix the most annoying ban line situations. The waterways from hell, with invisible ban lines on open water. Ban lines next to roads, or in a few places, due to road edge problems, onto the road itself. Perhaps add that parcels adjacent to Linden Protected Land that do not already have a ban line cannot raise one in future.

    Users would probably map all ban lines adjacent to Linden Protected Land. Drivers of SL does that now. The rallies guided by their HUDs warn you of ban lines ahead on your left and right. Once we knew that new ban lines along roads would not appear, it would be worth filling out the obstacle map.

    (There are a few parcels in SL whose coastal ban lines block off water access to large areas. Take a boat from Bellesaria to Sansara and head either east or west as far as you can. You won't get very far. I'm plugging for a Bellesaria expansion to the northwest to clean up that mess.)

    • Like 1

  16. 14 hours ago, Kurshie Muromachi said:

    Well, Tim Sweeney of Epic Games has this vision of an "Open Metaverse" planned out. They want to use open protocols, formats and standards and not a closed system where companies have a lot of power over things

    "Open Metaverse" exists, but it's just a C# library for talking to SL and Open Simulator. Last update 2011. Code on Github. Or is Sweeney talking about something else?


  17. 2 hours ago, Penny Patton said:

    I knew Facebook was working on a virtual world, but I didn't know it had a name yet, let alone that it was something you could sign up for. Worlds Adrift I've never heard of.

    I'm interested in big-world systems, and I know about these because I seek them out. None of them got much traction.

    Mavericks Proving Ground just went bankrupt a few hours ago. That was another Spatial OS based big world, or at least a medium sized one. (Spatial OS is a back end for running big seamless world systems. Mostly for MMOs, but it potentially could support something like Second Life. The game industry is waiting to see if their approach really scales like they claim. Spatial OS is a big project - 341 people and $500 million. But nobody major is using it.)

×
×
  • Create New...