Secondlife declining player base

[Coffee Pancake]

19 hours ago, Ashlyn Voir said:

I’d say it’s because SL has a serious lag problem and not everyone can afford high-end PCs to play a chatroom type game. Not to mention, there’s hardly anything to do in the game itself, sadly.

This is correct, and has been a problem since day one.

Although I wouldn't put the blame entirely on the need for a high end PC. Lag and latency is a byproduct of how SL works and it does have a massive impact on what works and what doesn't.

19 hours ago, Rolig Loon said:

Even after all these years, I can hear my mother: "Nothing to do?  For heaven's sake, girl.  Use your imagination.  It's not the world's job to entertain  you." 

You have this field, why don't you make a theme park? 

Money, time, skill at construction, skill at social engineering (badly overlooked), uncertain return on investment (not just fiscal),  contacts, marketing, artistry and so on. The staggering number of fully built empty regions is a testament to hard working failure. "Build it and they will come" is not true.

2 hours ago, Love Zhaoying said:

If Experiences caught on, it could help gamify SL more, and people may get friends to check it out so they can see/experience it too.

Experiences alone wont solve this. Animesh is one more peice of the puzzle but again .. it's not the last missing piece. Go try the Horizons games (land mass in center of Horizons residential areas). There is an hours worth of play there, do the whole thing .. some if it sort of works, some if it will have you wanting to murder me, none of it will satisfy anyone wanting to come to SL and play a game. It's in the realm of a tech demo and really shows the tech is woefully inadequate, best experienced for yourself.

Oh .. and don't accept the cash prize at the end, it's 20L .. get a thingy, at least you can rez it in your home as a souvenir.

Edited August 2, 2018 by CoffeeDujour

[Phorumities]

19 hours ago, Ashlyn Voir said:

I’d say it’s because SL has a serious lag problem and not everyone can afford high-end PCs to play a chatroom type game. Not to mention, there’s hardly anything to do in the game itself, sadly.

 

 

And yet you still log in 

[Tari Landar]

2 hours ago, Love Zhaoying said:

If Experiences caught on, it could help gamify SL more, and people may get friends to check it out so they can see/experience it too.

First you'd have to get people that know how to optimize experiences, and everything surrounding them(ie, the sims/parcels they are on) too, though.

I went to a sim(collection of sims perhaps?, not sure, lol) not too long ago that had experiences, and while it was VERY well built from a scale perspective and somewhat a visual perspective-that which I could see anyway (honestly, it was, and I'm someone that pays far more attention to how things look than how they function when I visit sims, quite often, it makes me smile to see the things people can do in sl), I couldn't actually "experience" the experiences(why is typing that far more funny to me than it probably should be?, I think I need a nap) because it slowed my machine to a crawl. I rarely have issues like that, though I am well aware that MOST of these issues are entirely client side. I don't have top of the line anything, but I also don't even have middle of the ground anything, lol. I'm somewhere between the two. I know how to adjust things on my end so that I can experience the things I see around me, the things I do in sl, etc..in the best possible way. I know what to lower, what to raise, what to increase, what to decrease..ok those things mean the same..whatever...I have gotten used to it after playing sl for a few years on a VERY sub par machine with integrated graphics, long before I purchased the equipment I have now, which far surpasses anything I have ever used to access sl lol. 

But my machine, she was having NONE of it, lol. It didn't help that there are scripted agents-yes bots-and they are listed as such-using texture change scripts every so many seconds, that some of the textures used are "high res"(at least, it is my belief that the intent was for them to be what many consider "high res") and much larger than needed to be, and weren't being as reused as they probably should be given the environment, or that a LOT of stuff is crammed into small spaces-small being relative to one's draw distance, which I keep low when I notice things starting to take an odd turn-allowing my machine ample time to process everything. Those things, combined with my machine not being the absolute top of the line, made the whole area un-visitable for me...which was a shame, because I bet the parts I couldn't see were just beautiful-to someone whose machine could actually process them. I know others have experienced the same there, we were chatting about it while standing around since the few with me couldn't move either, lol. I also know some others who didn't...but, again, they had top of the line everything under the hood, and they did even chat about some minor issues they had in various places, though they were negligible to them.  I'm not sure it's realistic to expect everyone to have top of the line anything, it's about as realistic as catering to the lowest of the low..neither makes a lick of sense. I have attempted to go back on more than one occasion, respectively when very few to no others were there too, because it helps. But, it's still a no go...hmph. It sounds and seems so intriguing, and I don't want to knock the creators, in the least. I am absolutely certain they put a hell of a lot of time and work into it, and I respect that immensely, even if I don't think it's truly optimized for the average sl user. 

[Coffee Pancake]

Object density, or clutter, is easy in SL and we expect a contagious world. The SL render engine is terrible at not rendering stuff you can't see.

The solution is to make your location from a series of discrete self contained locations connected via experience teleport doorways, Don't decorate the inside of your home, instead make the entry point teleport the user to a second copy of the building physically separate from the outside location .. but this goes against the basic expectation of a contagious world.

People expect the ability to explore by flying their cam around.

I would argue that LL should remove this ability and then suffer the screams, SL will be a better and more active place for it. But it will be an especially bitter pill.

[ChinRey]

2 hours ago, Pamela Galli said:

Years ago they had a great web based version just for beginners, that had built in tutorials.

But the tutorials would have been the first thing to go. No, I'm not talking about making better ones, I'm talking about getting rid of them completely - make SL easy enough to use you don't need to be taught the basics. This is an absolute must and it would require a total revamp of the entire user interface. That's not going to go down well with the existing SL'ers who are used to the quirkiness of the current viewers.

Next, the content would have to be simplified considerably. There are places in SL where a high end gaming machine is struggling to keep 10 fps. We need at least 15 consistently on a tablet. So forget about gigabytes worth of raw texture data and megatris worth of mesh. 500 MB textures adn 200 ktris of geometry, that's it. Oh, and it has to be done with proper LoD. Old SL'ers may have learned to turn a blind eye to disintegrated models - newcomers sure haven't.

Even if the graphics processors can handle it, everything still has to be downloaded - through a cellphone connection, not even wifi. Wait for the scene to rez? Ok, Second Life, you got ten seconds. If you can't cut it, I'm out of here.

Then there is a question of style. Modern SL content tend to fall into two stylistic categories.

The most popular independent G&L merchants tend to go for a toned down color pallette and builds with more than a hint of nostalgia and - perhaps even more important - items that are intended to be studied and admired on their own, not as part of a context. This is probably perfect for the current SL population but probably not that interesting for the youngsters.

Linden Lab's own games and most of their recent builds, are very much based on 1990s computer games - Mario Bros meet Bilbo Baggins in Disneyland kind of stuff. Again, perfect for people close to retirement who want to relive their innocent childhood. Possibly also great for actual children (if they were allowed here) but for those inbetween - nope.

This is just the start of a long list of requriements for a virtual reality with a broader appeal. Getting all of it would be a Herculean task. But far more important, so much of it would alienate many established users and there is no guarantee it would work at all, so the risk would be very high.

Edited August 2, 2018 by ChinRey

[Penny Patton]

Regarding the OP's second question: No, but if only to make their platform more open and friendly to new users they shouldn't make it easy and they should take action against offenders where and when feasible.

Regarding the observation about SL's decline, that's actually a large part of this other active thread.

To summarize what I've posted there; LL has always had a problem with polishing the usability and presentation of Second Life. They've also neglected to develop SL's content creation tools to allow for a broader range of content and activities in Second Life. As a result they quickly lost the staggering amount of momentum they had on their side in 2007-2009.

• Second Life has a lot of problems, but all of them can be fixed giving time and effort on LL's part.
• Second Life can run so much better than the average SL user has ever experienced, LL simply needs to provide better tools and reasonable resource caps to curb the worst habits of content creators. LL is already attempting to at least partially address this with new Land Impact calculations, which could help, but they need to do so much more.
• The SL UI has so much room for improvement. Throw in some UI interactivity features and LL could even create proper tutorials covering a wide range of SL's features and popular activities, and provide content creators the tools to create their own such experiences. This would both make SL as a whole easier to get into, but also allow for much more engaging content.
• SL's social tools are woefully underdeveloped, making it harder than it needs to be for people to find content, events, activities and communities that would appeal to them, but it's well within LL's reach to address this.
• Likewise, SL's content creation tools are underdeveloped in ways that limit the interactivity and level of engagement content creators can provide. LL is already working towards improvements in this area with Animesh, EEP, and Gridwide Experiences.

 For the past 15 years most of SL's shortcomings have been left as they are. If LL can look closely and learn where those shortcomings are they can begin to take steps to make SL a way better experience and that would be the best way for them to bring more people into Second Life.

[Coffee Pancake]

2 minutes ago, ChinRey said:

But the tutorials would have been the first thing to go. No, I'm not talking about making better ones, I'm talking about getting rid of them completely - make SL easy enough to use you don't need to be taught the basics. This is an absolute must and it would require a total revamp of the entire user interface. That's not going to go down well with the existing SL'ers who are used to the quirkiness of the current viewers.

The light viewer UI was an experience that created more problems that it solved, even the most casual of users would invariably need to do something requiring they use the full client.

As for UI .. Please! Open up photoshop and go to town. The bulk of TPV work is building and hooking up custom UI and the design side is where we're lacking the most, programmer art and all that. (I really can't stress that if you fancy the idea of doing this, please do).

2 minutes ago, ChinRey said:

Then there is a question of style. Modern SL content tend to fall into two stylistic categories.

The most popular independent G&L merchants tend to go for a toned down color pallette and builds with more than a hint of nostalgia and - perhaps even more important - items that are intended to be studied and admired on their own, not as part of a context. This is probably perfect for the current SL population but probably not that interesting for the youngsters.

This is to counter the lack of scene lighting bakes. We can't just throw down stuff, place a few lights and bake the ambient in. So .. neutral off white it is, as that's the only way things from a different creators stand a chance of not clashing horribly when placed side by side.

If the goal is to make things that sell, this is your only real option.

I would LOVE to see sansar style parcel texture baking implemented the same way pathfinding was. Place stuff, place some lights, and hit bake .. the client spends 20 minutes crunching numbers generating a set of atlassed textures specific to your parcel's lighting & replacing those set by the creator. Add new things .. they wont look right till you rebake.

2 minutes ago, ChinRey said:

Linden Lab's own games and most of their recent builds, are very much based on 1990s computer games - Mario Bros meet Bilbo Baggins in Disneyland kind of stuff. Again, perfect for people close to retirement who want to relive their innocent childhood. Possibly also great for actual children (if they were allowed here) but for those inbetween - nope.

Iono .. retro games (stylistically and actual old games) are just as popular as they ever were, perhaps even more so.

[Penny Patton]

11 minutes ago, ChinRey said:

Next, the content would have to be simplified considerably.

I'm going to go ahead and say that ChinRey means "optimized" here. Optimizing content does not mean you have to sacrifice the quality of it, or take away features. For example, take any avatar you've ever seen, including the most intricate and detailed. I can pretty much guarantee that avatar's texture usage and polygon could be cut in half without you noticing any visual difference. And that's a conservative estimate.

 I know ChinRey knows this, I've just seen so many people look at terms like "simplified" and "optimized" and jump to the conclusion that it would mean an uglier Second Life and I just want to get ahead of that and point out that it does not.

16 minutes ago, ChinRey said:

This is just the start of a long lsit of requriements for a virtual reality with a broader appeal. Getting all of it would be a Herculean task. But far more important, so much of it would alienate many established users and there is no guarantee it would work at all, so the risk would be very high.

 This I disagree with and want to point out why, because I totally get where ChinRey is coming from.

 Getting all of the improvements she describes here all at once would certainly be a Herculean task, but LL doesn't have to do it all at once. In fact I'd argue that LL needs to take the long view when it comes to developing Second Life. LL can introduce these changes slowly, over time, and take care not to alienate their current users.

They can introduce the tools to help make better optimized content along with official guides people can easily point to showing why certain content creation habits are bad and how to avoid them without restricting creativity or sacrificing quality. Following this up with more reasonable resource caps after people are already making better content so the userbase isn't thrown into upheaval.

 They can introduce new features one at a time, try out new UI ideas with various groups (including new residents) to see how they fare before committing to them.

 And I'd also argue that adding in the features that would allow for more variety in content and activities would not mean any of SL's currently popular types of content and activities would go away.

 LL doesn't need to change SL away from what it is, they just need to polish it up and give it the features that would make it so much more on top of what it is now.

[Ashlyn Voir]

1 hour ago, Phorumities said:

And yet you still log in 

Yep. You have a problem with that?

[Penny Patton]

1 hour ago, Phorumities said:

And yet you still log in 

Like I said over in the other thread; You're allowed to love something and still be critical of it.

[Phorumities]

1 minute ago, Ashlyn Voir said:

Yep. You have a problem with that?

Not at all, second life is big enough for everyone.

[Chic Aeon]

6 hours ago, Pamela Galli said:

Years ago they had a great web based version just for beginners, that had built in tutorials. I thought it was just what was needed as an introduction to SL. It just disappeared. 

 

Did you try it? It was pretty darn awful. You couldn't do anything but look around (well a bit more but not much). I was living in Winterfell at the time not far from one of the drop off points. The new users seemed just as confused as before. 

Anytime someone wanted to know how they could do something, an oldtimer had to explain that they needed to log out and change the mode of the viewer so that they could do it. I think even changing clothes required that (not really sure on that but it was very limited). 

And according to some recent official talks and videos The Lab is indeed working on getting that concurrency growing again. One of the priorities in a long range plan and he said they are feeling very positive about the possibilities. If you look at that chart the ChinRey pasted in (a nice one) the last three years look pretty even -- certainly not the Sky Is Falling.

That being said, even if the chart was showing a steady increase, SOME people would simply say, "It's not growing FAST enough". 

I am pretty satisfied with things of late. These last few years have been very good for me. 

Just sayin'.

 

 

[ChinRey]

13 hours ago, CoffeeDujour said:

The light viewer UI was an experience that created more problems that it solved, even the most casual of users would invariably need to do something requiring they use the full client.

Nevertheless, that's what is needed. But the fact the the light viewer failed, doesn't mean it can't be done, only that Linden Lab got it wrong.

 

13 hours ago, CoffeeDujour said:

As for UI .. Please! Open up photoshop and go to town.

You could also mention Blender or Gimp of course.

But those are not consumer products.

The day a 75 year old grandmother who has never used her tablet for anything but to check her bank account and to admire pictures of her grandchildren can log on to Second Life for the first time, be able to maneuver around within five minutes, have her first nice experience within ten minutes and never ever have any significant lag issues, that is the day SL will have the potential to broaden its userbase. If you say that's impossible, well it's not going to happen then.

Edited August 3, 2018 by ChinRey
Typo

[Pamela Galli]

2 hours ago, Chic Aeon said:

Did you try it? It was pretty darn awful. You couldn't do anything but look around (well a bit more but not much).

Yes, I thought it was ideal as an introduction and teaching how to navigate, chat, etc. but perhaps most importantly, people could get a taste without downloading a viewer.

 

[Chic Aeon]

16 minutes ago, Pamela Galli said:

Yes, I thought it was ideal as an introduction and teaching how to navigate, chat, etc. but perhaps most importantly, people could get a taste without downloading a viewer.

 

OK. Good for you. Apparently you were not the norm as I talked to a huge contingent of newbies (in Winterfell) that were completely confused :D.  

[Pamela Galli]

2 hours ago, ChinRey said:

Nevertheless, that's what is needed. But the fact the the light viewer failed, doesn't mean ti can't be done, only that Linden Lab got it wrong.

 

You could also mention Blender or Gimp of course.

But those are not consumer products.

The day a 75 year old grandmother who has never used her tablet for anything but to check her bank account and to admire pictures of her grandchildren can log on to Second Life for the first time, be able to maneuver around within five minutes, have her first nice experience within ten minutes and never ever have any significant lag issues, that is the day SL will have the potential to broaden its userbase. If you say that's impossible, well it's not going to happen then.

It could happen — I see my five yr old friend running around in Minecraft  and  Roblox, building stuff, on an iPad. Maybe some kind of two tiered SL, with more limited tablet access and a PC option. 

[Pamela Galli]

2 minutes ago, Chic Aeon said:

OK. Good for you. Apparently you were not the norm as I talked to a huge contingent of newbies (in Winterfell) that were completely confused :D.  

I wasn’t new at the time, I just checked it out and thought I would have found it less confusing than my newbie experience had been. But then I find I quite often don’t experience things like most ppl do.

[Love Zhaoying]

13 hours ago, CoffeeDujour said:

Go try the Horizons games (land mass in center of Horizons residential areas). 

No. Lulz

[ChinRey]

18 hours ago, Penny Patton said:

I'm going to go ahead and say that ChinRey means "optimized" here. Optimizing content does not mean you have to sacrifice the quality of it, or take away features.

Yes and no. I don't mean reducing the visual quality, quite the contrary, but I'm talking about a paradigm shift in our approach to 3D modelling and scene design and even our concept of visual quality. Better optimization would certainly help but it wouldn't be nearly enough.

To keep it short, I'll focus on the 3D modelling for now and leave the questions about scene design and concept of visual quality for later...

Creating individual objects from polylists (what we call "mesh" in SL) is inherently a very inefficient way to do 3D modelling. It requires a very large dataset and it's rather inflexible. As Avi Bar-Zeev put it:

Quote

It’s like trying to express the function a*sin(b) as a long series of undulating (X, Y) sample points instead of, well, just “a*sin(b).” The sample points are invariably not the ones you really want to ideally reconstruct the curve. And god help you if you want to change the key parameters to alter the waveform on the fly. The sample points are missing the essential mathematical (trig, in this case) relationship.

Avi Bar-Zeev: Death to Poly

3D models with low streaming cost and flexible level of detail need to be procedural as much as possible to make them more computer friendly.

---

Unigine's Valley Benchmark is a very good example of what can be achieved. For those not familiar with it, it's a 4x4 km (the size of 256 sims) nature scene full of details so lovely it'll make a brave G&L builder cry. Yet for my computer at least it's considerably lighter than even a fairly simple SL scene and the streaming cost (if it had been streamed) would only have been a fraction of what it would have been in SL. I'm not sure exactly what methods they use (I think it's fractals for shape generation and perlin noise for object placement) but that's not important. What matters is that instead of feeding the computer a long list of detailed instructions, you tell it how you want the scene to look and let it decide the best way to achieve the result.

---

Second Life's prim system is another very good example of procedural object generation. It's a very simple principle: you take a line (called a profile), extrude it along a path to make it three dimensional and then add (or not) a few simple modifiers to change the shape.

The current prim system officially supports four profiles (square, triangle circle and semicircle) and two paths (line and circle) and those account for all the seven basic prim shapes plus one that isn't normally used. If I counted right, there are 15 modifiers to create all the many prim twists. That means that a prim shape - any prim shape - can be defined by less than 20 bytes of data. The same shape as mesh would take kilobytes.

The prim system is brilliantly expandable too. Bar-Zeev hints at some of the possibilties in How SL Primitives [Really] Work but he's barely touching the surface. There are literally endless possibilities and some of them would - in principle at least - be very easy to add. (One has  been added even. There are two rather weird looking alternative paths not officially supported but available for builders with most thrid party viewers and viewable with any viewer.)

---

Procedural object generation and object placement are key factors for creating a truly efficient 3D environment. PBR is another one. We do have a few very crude implementations of normal and specular maps now - three completely different ones for normal maps and two for specular maps. But the code that handles them is so inefficient that it often is better to emulate those effects with mesh geometry and baked textures and besides, most of them are horrendously incompatible with Windlight (that too would have to drastically change of course). They're not nearly enough anyway.

We need an effective way to handle surface maps and we need PBR. Even Unity's "PBR Light" is probably not enough, we want the real thing.

---

All this add up to a completely different way to do 3D modelling. It's not compatible with any of the current Second Life content (except possibly some prim builds), It's not compatible with the current SL economic model and I doubt it is even compatible with the current SL culture.

Edited August 3, 2018 by ChinRey
Correcting typos

[Phorumities]

14 hours ago, ChinRey said:

Nevertheless, that's what is needed. But the fact the the light viewer failed, doesn't mean it can't be done, only that Linden Lab got it wrong.

 

You could also mention Blender or Gimp of course.

But those are not consumer products.

The day a 75 year old grandmother who has never used her tablet for anything but to check her bank account and to admire pictures of her grandchildren can log on to Second Life for the first time, be able to maneuver around within five minutes, have her first nice experience within ten minutes and never ever have any significant lag issues, that is the day SL will have the potential to broaden its userbase. If you say that's impossible, well it's not going to happen then.

grans RL tab in her profile:

young at heart senior looking to meet similar for nice experiences

*insert nude pic of 75 year old gran here*

[animats]

 

3 hours ago, ChinRey said:

Unigine's Valley Benchmark is a very good example of what can be achieved. For those not familiar with it, it's a 4x4 km (the size of 256 sims) nature scene full of details so lovely it'll make a brave G&L builder cry. Yet for my computer at least it's considerably lighter than even a fairly simple SL scene and the streaming cost (if it had been streamed) would only have been a fraction of what it would have been in SL. I'm not sure exactly what methods they use (I think it's fractals for shape generation and perlin noise for object placement) but that's not important. What matters is that instead of feeding the computer a long list of detailed instructions, you tell it how you want the scene to look and let it decide the best way to achieve the result.

I downloaded and ran that a few weeks ago. Not only is it very good looking at close range, it really is as big as it looks. You can get to the snowy cliffs in the distance. Those are not flat backdrops.

There's work of comparable quality in SL.

Hayabusa Design.  Trees and shrubs for larger outdoor areas. These all move with the wind. Visit and check your frame rate. Not too bad.

Level of detail, though, is botched. As usual. This island has about the same density of trees all the way to the shore, but we're not seeing that here. This is on the highest graphics settings. Trees are the ideal case for billboard impostors, but SL doesn't have those, so the distant trees look sparse. You can't have a good forest in SL.

Lack of modern LOD tools is SL's biggest visual quality problem. Distant objects look bad, disappear, or overload the viewer.

Edited August 3, 2018 by animats
Cleanup

[ChinRey]

1 hour ago, animats said:

There's work of comparable quality in SL.

...

These all move with the wind.

No they don't. They rotate back and forth in a way that looks thoroughly unnatural.

As for comparable quality in general, sorry but no. It's in a completely different league. Here's the demo video:

 

(Edit: I just have to say it. Valley Benchark was released in 2013, the year before LL announced the development of Sansar and Unreal Engine 4 turned up and revolutionised 3D rendering quality. Ebbe promised us cutting edge graphics in Sansar and it isn't even close to what Unigine could deliver in 2013, let alone what UE 4 came up with in 2014. Oh and before anybody comes up with the "but SL is so old" argument, the Unigine engine was launched in 2005, Unreal Engine as early as 1998. What makes the difference, is that those two engines have been continuously and systematically updated.)

Edited August 3, 2018 by ChinRey

[Love Zhaoying]

3 minutes ago, ChinRey said:

No they don't. They rotate back and forth in a way that looks thoroughly unnatural.

As for comparable quality in general, sorry but no. It's in a completely different league. Here's the demo video:

 

This made me sneeze due to realism and allergies.

[Phorumities]

52 minutes ago, ChinRey said:

No they don't. They rotate back and forth in a way that looks thoroughly unnatural.

As for comparable quality in general, sorry but no. It's in a completely different league. Here's the demo video:

 

(Edit: I just have to say it. Valley Benchark was released in 2013, the year before LL announced the development of Sansar and UE 4 turned up and revolutionised 3D rendering quality. Ebbe promised us cutting edge graphics in Sansar and it isn't even close to what Unigine could deliver in 2013, let alone what UE 4 came up with in 2014.)

Very pretty, but so what? 

I can watch travel videos that look better than that

[animats]

On high settings, the "Valley" demo is much better looking than SL. On "Low", it's roughly comparable.

"Valley" demo on Low Quality. 20 FPS on a NVidia GT 640.  No shadows. Comparable to "High" in SL, but not "Ultra". If I run Valley in "High" (which I will post later; I've hit picture upload quota) I only get 13 FPS.

This demo assumes you have programmable shader hardware - OpenGL 3.2 or above. SL doesn't yet require that.

Chin is right that Unreal Engine can look much better than SL, but the hardware spec to get that is higher. Not that high; a $75 NVidia graphics board is enough. Valley is Unreal's benchmark for 2013. There's a new one every year; you can download the 2018 benchmark if you like. Download page. This is a good GPU test; if it won't run properly, you have hardware or software problems.

You don't need a $500+ board for Valley. But you're not going to get this look on midrange laptops with default graphics.