Secondlife declining player base

[Kia Kiyori]

8 minutes ago, ChinRey said:

But that doesn't answer one fundamental question: if people want fun, if they want to feed their imagination, if they want an outlet for their creativity, why on earth would they want to go to Second Life??? There are so many other places for all of that!

As I said 'fun' is relative. Those who prefer MMORPGS would do better with actual games like World of Warcraft or something similar.

For others, it's all about creating objects, buildings, clothing, landscaping, etc etc - and endeavoring to earn a rl income from them, which MMORPGS are incapable of offering.

Combine one's idea of 'fun' with earning an income (however great or small) is ultimately what SL was designed for.

[ChinRey]

4 minutes ago, AnyaJurelle said:

For others, it's all about creating objects, buildings...

It may come as a surprise to many but Blender works perfectly well as a standalone program too, not only as an SL plugin.

Seriously, I think that is a very important reason (although by no means the only one) why Second Life has declined. SL is no longer seen as a relevant tool for building, not even by SL builders. The creative work is done elsewhere and then you upload the result to SL ... or maybe you don't bother.

[Kia Kiyori]

1 minute ago, ChinRey said:

It may come as a surprise to many but Blender works perfectly well as a standalone program too, not only as an SL plugin.

Seriously, I think that is a very important reason (although by no means the only one) why Second Life has declined. SL is no longer seen as a relevant tool for building, not even by SL builders. The creative work is done elsewhere and then you upload the result to SL ... or maybe you don't bother.

I haven't the intellectual capacity (nor the patience) to learn Blender. Just saying the name gives me a migraine  But yes, I've heard many do create elsewhere and then upload. 

Still, I'm not ready to give up on SL. As long as there's beautiful sims people have gone to so much effort to create that can be photographed and land barons supplying land I can landscape, then that's really all I seek from SL. That's my relaxing and fun escape from the cares of rl.

[Penny Patton]

1 hour ago, ChinRey said:

We need to get that down if we want a virtual reality that is as complex as Second Life and still reasonably useable to somebody on a fairly modest connection.

I can't speak for the technical aspects you and Theresa are talking about, but I'm not sure this is quite as big a problem as you make it out to be, nor that it requires a complete change in how content is generated to make it work. In sims I've optimized content loads really quickly. Granted, I'm already using low-poly content with good LOD where possible but that's why I suggest the tools should push content creators more in that direction.

If this is what you've been suggesting then sure, I can see the problems in implementing this. However, I'm not convinced there's a need to get rid of mesh and old content entirely to get everyone onto the kind of procedurally generated content you're suggested. It's an interesting thought but probably left to a successor to SL rather than an update to the current grid.

1 hour ago, ChinRey said:

Not only do we need procedural objects, we also need procedural object placement and instancing.

This is a good suggestion, but you don't seem to be mentioning any downside to it.

For the rest of your suggestions, there's multiple ways to approach the problem. We don't need a super strict system for resource management. Curb the worst problems and SL runs fairly well. Maybe not as great a professionally optimized AAA game, but well enough that it would be fine for most users, even those spoiled by their experiences with professionally optimized 3D worlds.

 For example, the LOD slider issue you and Coffee bring up, yes, LL needs to deal with this. However, better balancing how LI is calculated with regard to LOD could also solve this problem. Coffee has brought up before that you want creators to be able to cut out low end LOD on objects designed to be only visible up close and I can understand that view, but I believe a good compromise can be reached. LI already drops when you make objects smaller, afterall. I'm not sure most content designed for interiors would be all that affected by closing that particular loophole.

As for the interface issue, I'd say LL should focus on making the best interface they can, regardless of what users are used to. They don't need to make this new interface the standard right away. We already have multiple viewers running side by side. LL can push the improved viewer on new users, and us old codgers can either adapt or stick with what we're comfortable with. In addition, I'd argue that UI changes are also likely to be introduced slowly, over time. It doesn't have to be that we all log in to a brand new interface one day.

1 hour ago, ChinRey said:

• Get used to a brand new unfamiliar interface.
• Rebuild everything (at least evrythingin public view) with brand new, more efficient content.
• Replace your mesh avatar with a new expensive one (no wait, forget expensive, it's all on the Lab so that at least is not a problem).
• Replace all those lovely items that suddenly don't look so lovely now that you can't crank up the LoD factor.
• Get rid of all those items that turned out to be far heavier than the old calculation system thought.
• If you're a content creator: learn brand new skills (actually, some may think that would be fun - I sure would)
• If you're a merchant: watch as all your merchandise becomes outdated overnight (if you're an old-timer merchant: watch how all your merchandise becomes outdated overnight again!)

in summary

• Not all that much of an obstacle as people have a selection of viewers to choose from.
• This doesn't have to happen at all, not the way you're suggestion. Rather, introduce better tools and standards and then let time take it's natural course.
• Personally, I think it was a mistake for LL to allow mesh avatars to become the go-to. They needed to update the system avatar when mesh was rolling onto the scene. They have announced they intend to do this in the future. We'll see how well they handle this.
• Again, no need to make old content obsolete overnight. Granted, no matter how LL tackles this there will be much wailing and gnashing of teeth, but it won't be a mass exodus from SL and it will pass.
• See my previous points on the phasing out of old content.
• Not actually a big problem. If you know how to make the content in the first place, then you have the skills you need to optimize that content. People just need to be pushed in the right direction.
• Again, and I really should not have to keep repeating this point, but nothing needs to be made outdated overnight. See my previous explanations on why. This is an imaginary problem.

 

 

[Phorumities]

5 hours ago, CoffeeDujour said:

LL have to embrace service changes that impact existing vintage content as as possibility. Should LSO script support be phased out. YES. Sculpts? YES. System avatar hair? YES. The list goes on...

 

Why sculpts? they have their place. Are they more a drain on system resources than badly made mesh?

[Phorumities]

1 hour ago, ChinRey said:

It may come as a surprise to many but Blender works perfectly well as a standalone program too, not only as an SL plugin.

Seriously, I think that is a very important reason (although by no means the only one) why Second Life has declined. SL is no longer seen as a relevant tool for building, not even by SL builders. The creative work is done elsewhere and then you upload the result to SL ... or maybe you don't bother.

My building is entirely made up of prims, and if i were to run short of prims, I'd convert some elements of the build to sculpts. Repetitive rows of columns are excellent candidates for turning into a sculpt. I find that very often mesh builds don't look "real", like you are standing in front of a picture of a building. They have a "flat" quality about them.

If i could build with mesh while in second life, i might try it, but making a mesh seems like a terrible lot of work compared to rezzing a prim, stretching it to fit, then adding a texture.

*no socialists were slandered in the making of this post*

[Phorumities]

3 hours ago, Theresa Tennyson said:

Why would you want to do this? Sculpts are very inefficient in the viewer because of how they're made but they're comparatively light on the servers and data transmission pipelines. Your plan will require a whole new system to make a lot of bad meshes that probably won't help the viewer that much and that won't be as light on the servers and data transmission pipelines.

It's not necessary to do something that will help, its just necessary to do something.

oooo sculpts are bad, but meshes are good, so lets convert all the sculpts to meshes even if it makes things worse.

And they say SL isn't like RL

*rolls eyes*

Edited August 5, 2018 by Phorumities
added a word

[Penny Patton]

12 minutes ago, Phorumities said:

Why sculpts? they have their place. Are they more a drain on system resources than badly made mesh?

Sculpts are pretty bad. They're a lot higher poly than they need to be, especially how most people use them. Are they worse than badly made mesh? Debatable, but decently made mesh beats sculpts hands down, so it would be a good thing to push people towards better optimized mesh content and let sculpts phase out of use.

13 minutes ago, Phorumities said:

If i could build with mesh while in second life, i might try it, but making a mesh seems like a terrible lot of work compared to rezzing a prim, stretching it to fit, then adding a texture.

I've always hated how LL abandoned the in-world tools. I love mesh, but, SL lost a lot of it's charm, and a lot of it's draw, by letting the prim based tools rot on the vine.

[Phorumities]

2 hours ago, ChinRey said:

I haven't seen anybody even trying to debunk any of my arguments but maybe I should explain them more in detail.

Based on the info at the SL wiki, with 500 kbit/s bandwidth assigned, it should take about a second to download 20,000 tris worth of mesh. For a million tris - which isn't unrealistic at all - that means almost a minute. That is just to transfer the mesh shapes from assets server to client. (The maximum amount of mesh - that is compressed mesh files ready to transfer - for a sim is 128 MB. That's more than half an hour's worth of 500 k/s transfer time but I really hope no sims actually are that overloaded.)

We need to get that down if we want a virtual reality that is as complex as Second Life and still reasonably useable to somebody on a fairly modest connection. Caching helps of course but only to a certain degree especially if people are moving around a bit - and we want them to move around a lot. Optimization, yes we need that, as much as we possibly can get. But I do actually kow a thing or two about mesh optimization and how far we can take it and I can assure you it's not enough - not nearly enough.

And that means procedural object generation. This has to be developed almost from scratch. The prim system may be a good starting point but as it is, it's way too simple for today's needs. We also need good tools for working with that new material (which as a great bonus should mean a renaissance for inworld building). But most of all, it has to be actually used. It has to become the main buiding material in Second Life. Polylist meshes should only be used the way they were originally intended, for the few exceptions where they are the only alternative. Do you think SL users will replace all their old meshes with brand new procedural content over night? And how will the big merchants react? Do you think they will take that change easily? Especially of course, those who don't make their own content. This new material will be unique to SL so you won't find it (legally or not) anywhere on the interweb.

---

Not only do we need procedural objects, we also need procedural object placement and instancing. We need them, among many other things, to fill up all the abandoned land and other stretches of barren land with temporary content. A handful of low lag plants with instances placed in a perlin noise based pattern would do the trick. Fast efficient and, since they are instanced and located in palces where there's nohing else going on anyway, they're not going to cause any load worth mentioning on any part of the systen.

---

We need an effective resource accounting system. It has to account for everything, including avatars. The balance between server, connection and client load has to be based on a fairly low spec computer on a wifi connection or less. And the numbers have to be as accurate as it is possible to make them, no compromises whatsoever.

That means, if an old build turn out to be too heavy, then it is too heavy. Period. Maybe it would be possible to "lock" overloaded sims, let them keep the existing content but not allow anything to be added until the load has been brought below the new limit. But that's all.

No compromises whatsoever also means the suggested band-aid solution to stop LoD butchery is dead. If you can save actual load with broken LoD models, well then you have saved load and that is reflected in the calculated weights. Linden Lab will have to find another way to get builders to make proper LoD models.

The RenderVolumeLODFactor will have to go. It seriously skews the load calculations since increased LoD factor increases the bandwidth and asset server load significantly. Of course, that should also be the solution to the previous problem. You can't tell people to crank up their LoD factor if that function doesn't even exist.

The fitmesh LoD bug has to be fixed one way or another. Either LoD models have to start working with fitted mesh the way they should or the weight calculation has to take into account that fitted mesh effectively is at max LoD at any distance. Both solutions would effectively break practically all fitted mesh, the first solution literally, the second because those items will turn up with so high land impacts nobody can afford to wear them. This is all Linden Lab's responsibility and the users and merchants have all rights to be seriously p'''ed if the Lab doesn't offer them full compensation for all the inconvenience and losses the change causes them.

---

Then it's the user interface. You log on to SL and your screen is filled up to capacity with a lovely scene - closeup, no birdseye view anymore. The only little disturbance as a row of a dozen or less buttons along one of the edges giving you access to the most essential functions. All the other bells and whistles may still be there but discreetly hidden away until called for. But SL'ers love bells and whistles, they can't enough of them! And they would having to learn using a brand new interface.

---

There is a long list of other improvements necessary to make SL attractive to new users too of course. I started writing one but gave up after two pages. Maybe I'll get back to it later.

But I think the ones I mentioned are the ones that would cause the most problems for existing users:

• Get used to a brand new unfamiliar interface.
• Rebuild everything (at least evrythingin public view) with brand new, more efficient content.
• Replace your mesh avatar with a new expensive one (no wait, forget expensive, it's all on the Lab so that at least is not a problem).
• Replace all those lovely items that suddenly don't look so lovely now that you can't crank up the LoD factor.
• Get rid of all those items that turned out to be far heavier than the old calculation system thought.
• If you're a content creator: learn brand new skills (actually, some may think that would be fun - I sure would)
• If you're a merchant: watch as all your merchandise becomes outdated overnight (if you're an old-timer merchant: watch how all your merchandise becomes outdated overnight again!)

I sum it's a lot to ask from the old-timers and I'm sure many of them will decide it's too much and leave.

So we are going to completely remake something that works perfectly fine (for me anyway since I'm not obsessed with running on ultra).

Basically, declare everything that currently exists obsolete?  Force people to redo EVERYTHING? Tear down EVERYTHING? And you don't think lots of people aren't gonna say eff this crap, i'm done?

Then why not just delete sl entirely and move forward with the brave new doa world of Sansar?

[Penny Patton]

You missed where ChinRey was saying doing these things would be too risky for SL and potentially push a lot of existing users away.

[Phorumities]

I admit, all this technical stuff is meaningless gibberish to me. All I know is I log in, and in under a minute tops the scene is loaded, I'm at my throttled frame rate of 25 fps, and i can move around and do what I want. I render jellies to see if they look better that way, I dance, I shop, I hang out with my friends, I just live my Second life.

SL might not be virtual reality quality, but the only virtual reality I want is the friendships and associations, not "oh my god I can't believe it's not real" scenery.

Bottom line, SL works just fine for me.

Edited August 5, 2018 by Phorumities
added a line

[Phorumities]

10 minutes ago, Penny Patton said:

You missed where ChinRey was saying doing these things would be too risky for SL and potentially push a lot of existing users away.

I didn't miss it, I was just confirming it.

[ChinRey]

13 minutes ago, Phorumities said:

So we are going to completely remake something that works perfectly fine (for me anyway since I'm not obsessed with running on ultra).

No. I have to repeat what my point all the way here has been: It is not about how to make SL run better for the current users.

That would be great too of course but this thread is about the decline in the user base and my replies have all been about how a virtual world for a large number of people who do not have effective access to Second Life as it is today, needs to be made.

My answer to that is that such a world would have to be so radically different from Second Life there's no point trying to transform SL into it. Don't worry abut the retention rate or the decline in the user base because there's nothing to be done about them and they're not going to become a serious problems for many years to come yet. Make SL as positive an experience as possible for the SL users. And if there is a need and a market for a virtual world for a broader audience, make it somewhere else.

[Phorumities]

2 minutes ago, ChinRey said:

No. I have to repeat what my point all the way here has been: It is not about how to make SL run better for the current users.

That would be great too of course but this thread is about the decline in the user base and my replies have all been about how a virtual world for a large number of people who do not have effective access to Second Life as it is today, needs to be made.

My answer to that is that such a world would have to be so radically different from Second Life there's no point trying to transform SL into it. Don't worry abut the retention rate or the decline in the user base because there's nothing to be done about them and they're not going to become a serious problems for many years to come yet. Make SL as positive an experience as possible for the SL users. And if there is a need and a market for a virtual world for a broader audience, make it somewhere else.

Yes, I doubt anything can stem the decline of the userbase. Every virtual world starts out fresh and new, it draws people to the shiny, but over time the luster wears off and people go looking for the next great new thing.

And let's face it, there's a lot of older people in second life. As real life takes its toll, so second life will lose its numbers.

Its because second life is so much more than any other platform, so much more than just a game,  that it has endured for so long, is still making a lot of money for its owners and will continue to do so for a few more years anyway.

 

[Penny Patton]

I'm pretty sure if LL fixed all of the problems mentioned in this thread, it would definitely result in a lot more interest in SL. How much? Hard to say. SL is pretty old, but it's also been upgraded quite a bit and with the aforementioned problems solved it could look and run even better. Toss in more features and it could potentially give the whole platform a second wind, finally delivering on the promise people saw it in back at its peak.

When people move on to the next shiny thing, that assumes there is a new shiny thing to take the place of the old. Right now there is no real alternative to SL, it's the only game in town.

[ChinRey]

1 hour ago, Penny Patton said:

I'm pretty sure if LL fixed all of the problems mentioned in this thread, it would definitely result in a lot more interest in SL.

Well, you can always hope.

Me, I'm signing off from this thread now. It's going nowhere fast.

[Coffee Pancake]

3 hours ago, Phorumities said:

Why sculpts? they have their place. Are they more a drain on system resources than badly made mesh?

Sculpts were a hack to get more organic shapes into SL. The idea being that you have sphere, and with a little pinching and pulling you have an apple. Sculpts in SL are hardwired to degrade into spheres.

The problem came when people started making complex geometry with them, making every possible vertex the format allowed (which wasn't many) count. Soon prim to sculpt converters appeared where you could slap down some prims, and then replace them all with a single sculpt (from apples, to .. a shipping pallet for example). This type of geometry didn't really work when the system tried to degrade them back to a spherical form, even if only by a little.

Sculpts allowed us to hack in complex geometry and keep the 1 Li rezzed cost, in effect we found a new unintended use for them, and as rezzed Li costs money, there was plenty of pressure to do this.

The problem .. they looked like a scrambled mess of triangles unless you were right up close to them.

The bone headed solution (given to customers after they complained about they new items being poop), change the viewers LOD debug setting!

In the end, as everyone was fiddling with a debug setting anyway, firestorm changed the slider in their viewer so it doubled the range of LOD values. No more users blindly changing debug settings and no more 1 prim furniture exploding into poly vomit. As a side effect, every single prim and avatar in SL didn't degrade as intended and SL got slower everywhere, most people didn't know that firestorms max detail was double LL's max detail. The end result was that sculpts looked ok in FS and junk in everything else, because if you don't know, you're not going to test your stuff.

Then we started create a sculpts that were shrunken down then scaled it up twice the size (and looked close enough, sculpts are not a precise format). But to the viewer this sculpt was twice as big due to the bounding box further forcing the viewer to keep it fully loaded, add in a smattering of creators telling people to mess with debug settings even further and basically switching off viewer attempts to manage LOD (people are still doing this out of habit).

We end up with tiny sculpts that look ok from half a region away and a viewer that's struggling to render everything at full resolution.

Fast forward and we have mesh, superior to sculpts in every way and really what we should have had in the first place. The problem .. viewer LOD settings are still jacked though the roof and pressure to make everything a cheap to rez is still there. So .. we complain about mesh objects degrading to balls of triangles, mesh creators are twigging that the same bounding box hackery that worked for sculpts also works for mesh, performance is at an all time low and somehow it's textures that are to blame.

Sculpts are a poor substitute for mesh and while they aren't bad in isolation when used as originally intended, the repercussions of them crippled SL performance for everyone, everywhere, and we're still living with the fallout today.

Bin the sculpts by rendering them as mesh, restore the viewers LOD scale in firestorm, cry about all the stuff we've bought that's objectively terrible but end up in a much better place.

 

Penny is right when she states we need limits, seriously, we will break everything just to save a prim or keep things rezzed at insane distances. Limits that don't hurt are simply not going to be effective at curtailing the runaway performance problems we have today. Object complexity is increasing exponentially along with tricks and hacks to get it rendered, we're knowingly bombing ourselves back to a text adventure. SL can bring a high end rig to it's knees and most users are running with integrated intel graphics and budget hardware.

This mess is absolutely hurting retention.

 

 

Edited August 5, 2018 by CoffeeDujour

[Tari Landar]

4 hours ago, ChinRey said:

No. I have to repeat what my point all the way here has been: It is not about how to make SL run better for the current users.

That would be great too of course but this thread is about the decline in the user base and my replies have all been about how a virtual world for a large number of people who do not have effective access to Second Life as it is today, needs to be made.

My answer to that is that such a world would have to be so radically different from Second Life there's no point trying to transform SL into it. Don't worry abut the retention rate or the decline in the user base because there's nothing to be done about them and they're not going to become a serious problems for many years to come yet. Make SL as positive an experience as possible for the SL users. And if there is a need and a market for a virtual world for a broader audience, make it somewhere else.

You said it better than my long winded self can, but, yes...this, exactly, which is what most of my posts were about, though obviously not very clear.

This is where what LL can do, gets garbled up with what people think LL should do. Hence why I kept saying that what they can do, is absolutely relevant to "what they should do" conversations, they go hand in hand, always have, always will. To do what they should do-that is, what a lot of people want them to do, or to have done for that matter-would ruin what sl is today, because it would be something completely different, and it would have a negative impact on current users/merchants/creators...even if some can't see that. If people are worried, concerned, whatever, about user growth and/or retention...they shouldn't be pressing so hard for these "should do" things, that would put both growth and retention in jeopardy. Stick with the things that improve current user experiences(and I mean the whole of sl experience, not, those relatively new "experiences" we have...why do they use that word anyway...but I digress) and make what we have now, work better using tools and techniques that aren't going to simply break things for a hell of a lot of people. Trying to make drastic changes, such as some have suggested, and you quite eloquently pointed the majority of them out, will without a doubt have a negative impact on the whole of sl, both from a growth and retention vantage point. 

[Phil Deakins]

9 hours ago, Penny Patton said:

Unless LL is allowing legacy prims [...] to continue to count as 1 LI per prim (which I don't believe they are)

Now that idea is very scary. It sounds like pushing people away from prims, and that would make SL a completely different thing to what it has always been. It's bad enough that you can't be decent creator of saleable goods these days without doing it externally, and spending a huge amount of time learning how to do it. I really don't want prims pushed out.

I know that's not what you meant, but the very idea is still scary.

Edited August 5, 2018 by Phil Deakins

[Phil Deakins]

5 hours ago, ChinRey said:

My answer to that is that such a world would have to be so radically different from Second Life there's no point trying to transform SL into it. Don't worry abut the retention rate or the decline in the user base because there's nothing to be done about them and they're not going to become a serious problems for many years to come yet. Make SL as positive an experience as possible for the SL users. And if there is a need and a market for a virtual world for a broader audience, make it somewhere else.

You mean........an SL2?

What you've described in this thread, including the videos, reminds me of what we (I, anyway) imagined when LL announced that they are starting work on "the next generation", saying that, "If we don't do it, somebody else will", and we all thought it was going to be a souped-up SL, so we refered to it as SL2. You've described a 'souped-up SL', and I think it's such a pity that LL meant something that has all the hallmarks of being a pipedream instead. Sansar.

Edited August 5, 2018 by Phil Deakins

[Phil Deakins]

I agree with those who suggest having limits. Until I read this thread, I didn't know that the sculpty system was pushed beyond what was intended, but I did know that it happened with mesh. Mesh avatars weren't expected or intended. I was aware that mesh is often made so that it has a significantly heavier impact than necessary. I didn't know that Firestorm changed a default setting to deal with bad design.

It all sounds like quite a mess, and I do think it's time to apply the brakes, in the form of the suggested limits.

Edited August 5, 2018 by Phil Deakins

[Coffee Pancake]

The trick is meaningful limits that we wont just hack our way around or don't burn the house down around us.

LL will never deploy a Li change that will result in auto return triggering when the region boots up after it's applied. There would be pitchforks.

 

[ChinRey]

8 hours ago, Phil Deakins said:

You've described a 'souped-up SL', and I think it's such a pity that LL meant something that has all the hallmarks of being a pipedream instead. Sansar.

Yes, I know I've signed off this thread but I have to correct a possible misunderstanding in one of my posts.

When I mentioned Sansar earlier, I didn't really want to criticize. Sansar started off with a lot of mutually exclusive goals. Some of them had to go and the mass appeal turned out to be one of those. I can't say whether that was a good or bad or even conscious decision and that's not relevant either. What matters is that it ended up being targetted towards an even narrower niche than SL.

[Coffee Pancake]

Even if you ignore all the marketing buzz about creating "experiences" and "VR" , just looking at what Sansar offers, I have no idea what or who it's for. 

Without some sort of communal contiguous lobby for people to start in  .. it's a collection of bubbles. Where is everyone / anyone, not here .. and that's all you know. How do you even tell people .. come look at this cool bubble I made / found / love.

It's like trying to convince someone to check out something awesome, and when they ask where it is .. you say .. oh, it's an obscure corner of the intertubes, you need this client, then search for this bubble, wait 20 minutes for the download .. and then hangout for HOURS for someone to show up, i'm sure they will, you know, eventually. Oh you need to make an avatar too .. 

Even the act of playing dress up is entirely isolated. What do you think of my outfit ? No?  Ok, I will be back in while after I tweak it .. how about now? No? 

[Nextio]

SL's biggest problem has always been it's game engine, because it was constructed by amateurs.  It was never built to have performance and optimization as it's top priority, unlike mmorpgs.  They would need to start all over with a new engine to handle new coding techniques and hardware technologies that deal with rendering, culling, draw distance, and so on.  People were hoping for SL 2.0 but we got Sansar instead.  Now the SL playerbase is stuck with this dying old game.