Upgrading Second Life, Or Starting a Third?

This is pulled from a Linden Hotline thread I began, but just as I'd like to hear the Linden's views on this subject, I would also like to know what we, the players, feel. My questsions to YOU are at the end.


Second Life is growing. It has been since day one, and at an ever increasing pace. The more time that passes, the more difficult it will be to change aspects of the environment and the tools to modify it that so many have become familiar with or even dependent on. Existing content that represents a large investment on the part of the user base can break when significant changes are made to underlying systems.

I see this as Linden Lab's most significant challenge at this time. Not because the problem is bad now, but because the time to prepare for when it does become bad is here now... Or even has passed.

Today there exist problems, great and small, that have persisted for long periods of time. Some from the very beginnings of Second Life. The reason why the needed changes have not been made in many of these cases is that to do so would cause prior work to become invalid or broken. There is a view that all the effort that has been made so far in this virtual world is nothing compared to the weight of future efforts, perhaps spanning decades. But it is far easier to ignore future needs in favor of the present's comfort. This is something that needs to always be in the forefront of the developer's minds.

That said, actually designing a system that can accommodate change without breaking down or causing unrest is a tall order. One solution is to provide for concurrent versions of behavior and data, but this can increase data sizes in the short term and adds a level of complexity. In its favor, however, over the long term the benefit more than outweighs the cost, and if a pattern is developed and rigidly held to, the learning curve would not be hugely affected.

( As an example: Radical changes could be made to the relatively isolated environment of the Linden Scripting Language if the scripts could be identified with a particular version and the servers that run them could support every recent version concurrently. )

Making the shift in thought and design to permit seamless upgrades to what is a VERY dynamic environment can not be a simple transition. It would be a massive upset to the game, in all likelihood. But I ask you this:

Isn't it better sooner, with 20,000 potential headaches to come, than later when who knows how many will be affected?

Does Linden Lab consider Second Life to be a long term platform, or a shorter-lived experiment and proving ground? Do you feel it is worth the development time to build for tomorrow? Given that Second Life, as it stands now, has no visible path to take it significantly forward, are you ready to rest on your laurels and see just how long this limited format will last?

I don't knock the experiment, if that's what it is. I find it fascinating, and look forward to each new development that arises from SL's diverse player base. But part of me wants to see it grow, not due to novelty, which wears off all to soon, but due to growth of the environment, tools, and customization. And as things stand, I have doubts about that progressing much farther. But if there is to eventually be a "Third Life", one that is ready to extend and grow with the world that uses it... I'll be the first to jump ships.




And now my questions for you, my fellow residents...


Are updates to SL that require you to re-address or repair previously created content worth the fixes, improvements, and added features they bring? Or do you feel that previous creations are more important than things you, or others, may make in the future?

Do you think that adding complexity to the environment in order to simultaneously support multiple versions of behavior and appearance is acceptable if it means prior works are less likely to be affected by future change?

If, in the end, it is decided that Second Life has painted itself into a corner that no incremental upgrade can pull it out of... Would you be willing to start anew in a "Third Life"? What about a fourth? Is one total do-over acceptable if it means things learned can be correctly applied from scratch, or would the introduction of any new environment based on, but different from, Second Life be unacceptable to you?

When should we, players and Lindens alike, consider ourselves ready and knowledgeable enough to start from scratch again and do it right from the beginning? Do you think we can ever be sure we are ready, or do you think the leap has to be taken blind?

What can we do now, in this Second Life, to prepare for a long-term future evolution of virtual worlds?

Thank you for muddling through all this. I look forward to this discussion!

I feel the changes are, in most (if not all) cases are warranted. I feel that, if we as residents had the option to an Inventory-Wide "search and replace" for objects and code, this would make things more or less moot... but that would be difficult to accomplish.

On current methods, however, I will again say that it's worth it. Because the net benefit is typically larger than the net loss to us as scripters and content creators, I'm all for said changes.


Isn't this already supported? I know I typically receive scripts tagged "lsl2," for example.


I would support a version rollback in this instance, with 24-72 hours of notice. If it came to this in a more major sense, though, I would be supportive of building a platform above or parallel to Second Life to further production. This is thinking very-long-term here... ultimately, a "next big thing" in this vein is more or less assured.


That's a value judgment that I can't profess to make for everyone. I would say when things are so obsolete as to necessitate this, or if the leap to a new technology was important. And I would not start from scratch - I would write an importer/exporter from Second Life and store all the information on user content for later use. I've been tempted to write a resident version of that, by the way.


Besides pray? This pretty much applies to SL content creators instead of pure scripters, because "conversions" for scripters are far easier. I'm already preparing for the evolution, by learning the skills that I deem will be viable in that leap - 3D mesh-based modelling, free-form scripting logic, mesh-based animation, texturing, sound production... I would call myself a Renaissance Man if I wasn't so jaded.

I think there are several "sure" things to be on the lookout for as to the "Next Big Thing:"

1) Mesh-based and NURBs Production - this is pretty much assured, it's just difficult to regulate.
2) Surface Texturing Changes - emissive, bump, trans, normal, specular, reflective, multi-tex... added texture layering is ensured.
3) Larger integration with Reality - ranging from rapid prototyping to 3D positional systems.
4) Software API Packages - nothing to add here.
5) Consumer-Level Productions and Storage - maybe not in Second Life, but eventually.
6) Improved Streaming Audio and Voice Support - again, technology limit here in the short term.
7) More "Intuitive" Logic - as these things become more familiar and more enmeshed in reality.

I will heartily endorse starting over if the changes made are significant and worthwhile. I would love to start with an empty lot and a bare inventory if it means vehicles work, sim boundaries are not an issue and movement made more realistic. On the other hand, if it means losing accounts, wealth and accumulated land then, no, I'm not in favor of it or I would have to think long and hard about it.

Devistating short term change can sometimes have long range benifits so I am not opposed to them on principal. If it makes for a significant improvement in the future, I am not going to worry too much over rebuilding my house.

Yes to all.

If the change would bring a great deal of improvement I would be willing to deal with anything it may break and have me to remake it. However I would A) Like advance notice of such changes to prepare and B ) know of the advantages such changes will bring.

Also, I would switch to a "Third Life" but only if the same development and creative team that are behind Second Life would be behind "Third Life" as well.

Scripts made in alpha still work today. Was this lsl1?
Somebody ask bp (he was the one who showed me stuff made in alpha) or BBC...

Nice post.

I would rather eat the broccoli now and enjoy the cheesecake later.

great post tiger. yes, make the changes in lsl, dump the telehubs in the trash, explode the grid and cancel all land ownership if that's what it takes to put the software development on the right track to being useful and productive in the real world.

you'll only hear kudos from me.

i'm convinced that the primary reason ll don't pursue necessarily destructive enhancements is because they fear the lawsuits people who've invested thousands of us dollars might file for loss of income. i used to believe it was because they were testbedding second life and would produce other software later. they've well beaten into my head that i'm wrong to think that.so they have, apparently, "painted themselves into a corner."

now we'll just have to see if they're able to fight their way out or if someone new will have to come along and be the bad guy.

to me this reads as, "should ll take what they've learned from sl and build something better and tighter from the ground up?" (and even if that's not the point... that's the question i'll answer )

i think that would be a great idea. and i think it would take 1-2 years to do right.

given that i'm sure ll learned quite a lot from secondlife, a fresh new code base would surely be tighter and more efficient. it would allow ll to fix deep architectural issues that probably make certain fixes difficult address in an active code base.

Good thread.

Our desire is to balance changes with backwards compatibilty. We take content as built today very seriously, since there are so many great works in-world already. Given that, we think we've architected fairly well for allowing incremental changes. To answer Tiger's original question - we are very much behind SL expanding and retaining content - we don't see the existing version as an experiment that can be wiped out. As discussed, we think versioning and adding functions can allow expanding LSL while still retaining all previous functionality. Admittedly, this is often a challenge. We've already been able to do 'rolling' deployments, where we upgrade some sims but not others (we are doing this now with updates to the OS version under the sims, for example).

Longer term, we are trying to architect in such a way that upgrades can be made progressively to some sims and not others - this should allow a balance between preserving older content and allowing more radical or non-backwards-compatible changes. One has to imagine that with thousands of sims this will become the norm. A challenge to this sort of model is to normalize the sim-sim messaging and interactions to a standard protocol that will support such changes.

Linden Lab never guaranteed that it was forever. We've been very lucky so far in our upgrades; some scripted objects broke and panties became a little skimpier than they should have been. We managed to recover. With each major innovation, we came out better off than we had been before. I think that personally, I would be delighted to start off fresh - totally clean slate - so long as we didn't really lose out by it, and there were real improvements in the performance and general appearance of the medium. I would be really upset if I lost my two years worth of textures (most of which I never use) but I guess I would get over it with enough new features to distract me. I think you are correct in that it would be much less disruptive to make a major break soon, rather than later.

Wow, the implications from that idea will take me a bit of time to digest. Some new features and changes only coming to new sims... Older sims with content that depends upon older methods being valued for those legacy elements, in some case, shunned in others for being backwards.

It would really mix up the land valuing system, far more than the single existing example of this, land modification limits. I think that decisions to potentially split feature changes like this should be heavily discussed with the player base. I'd hate to see an unnecessary split take place.

A sample might be: a graphical bug that's been around forever (topsize texture distortions) but might not be fixed yet because a few people have made custom textures to work around the bug.

Boggle boggle...


Lessee, I have 5000m2 in an old, legacy sim that I will have to give away so I can pay premium prices to move to a sim where vehicles work or Havoc2 physics are installed?

Wow, its one thing to make a clean slate for everyone, it is quite another to screw those who can't afford to move or who simply can't because of the complexity of their build. I can only see this working in any equatable way is to vary the tier rate so older sims are attractive due to lower cost. Even then it becomes a rich entitlement debacle. The old server performance controversy all over again except 10x worse.

Bah...

Kathmandu,

This suggestion is NOT MEANT for every feature or upgrade. Just for changes to existing systems that some people may have been relying on. Even then, it would only be used for medium-sized changes of that type. I don't think Linden Lab wants VAST differences between sims.

The different landscape modification ranges between the 20-30 or so original sims and the vast majority of sims added since is an example. It doesn't affect gameplay at all. The tradeoff is that in the new sims, you can't distort the land as much as you want, but you also won't be bothered be neighbors who DO. In the old sims you can, but have to put up with neighbors that HAVE.

Big things like Havok 2 have to be all or nothing.

Small things like a change to how a script function behaves would also have to be all or nothing, since scripts can move from sim to sim.

The possible options for things that might be different between sims is FEW. Pretty much limited to land-specific changes, as far as I can imagine. Like how Olive was used as a testbed for tying prim limits to parcel sizes.

Fantastic way to put it.

For me, I would rather eat the brussel sprouts now and enjoy the watermelon later.

Hey, do I hope I and at least some of us are still here five years from now in a thriving Second Life-of-the-future? Yupsureyabetcha!

I'm a person who believes intensely in solid bedrocks and fixing the foundation before building the rest of the house. That's a crude analogy, but can lead to some evocative mental images.

I like brussel sprouts.

The other side of this would be losing a part of history as it were. And an idea such as Philip gives with some sim's remaining un-touched, would preserve this in some way, such as Clementina.

I wasn't thinking in a maner as this, but it occured to me again when re-looking over this thread.

So then the problem then becomes, how does one keep stuff that may be of importance from the past, when the system is changed so drasticly? One way of going about it would be to have all previously made items, stay as they are, but any futurely made items would be using the new system. That could break the system in some way, or make it totaly pointless though.

I guess I was thinking a bit selfishly, in that, I wouldn't mind if any of *my* items had to be rebuilt, because I could do as such. But items I have from others, that may become broken, and the owner may not be around, or may not wish to rebuild the item, well then I would be in quite the pickle. Sure I would have images of such items, but those are not as good as the items themselves.

So I'm back to square one.

Also you bring up another thought Philip, lets say an "old sim" that uses a non-upgraded system, represents an object I'm wearing in one way, but a "new sim" represents the object I'm wearing in a different way, how would that carey over? Would the item be able to transition between the two seamlessly? Or could the constant going back and forth between its old look and new look somehow effect it negatively? Say, just as an example, I have an item that uses a prim type that SimOld does not have any data on, what would happen then?

Not saying you may know these things now, but its an interisting problem to think of when you have old systems next to new and transition between the two is so easy to happen.

Oz brings up interesting points.

In the future, would there be some sort of EMULATOR to run previous versions of SL within -- given the massive compuhorsepower of tomorrows -- or would we otherwise have some sort of "wrapper object" functionality to "see what things were like"? It is a fascinating thoughtline...

Hah. It is indeed. Why do i doubt this sort of crazy backwards system could compete with a smoothly running third life that is bound to exist in the future - whether the Linden's create it or not - built on technology that wasn't available to them when they began.

Maybe they can have an eventual 3-5% market share like Apple.

Is that what the smell in my sink was?


Related to the post, I agree that almost anything is acceptable as long as the world continues to evolve.

That is the the thing though, are small changes what he meant or large? Geeze, its like getting cryptic messages from the oracle...

The question, as I undrstood it, was if it took starting over with SL with a clean slate to fix some major, persistant problems, would the residents go along?

Philip answered something to the effect that LL is going to stucture it so problems get solved in new sims while maintaining older sims. That is what I got from it but hey, I ain't so smart.

It ends up being the same thing even for smaller changes, like "free trees" sims (if it were ever to be implinented) would raise the value of the sim and cause the 1 prim tree sims to have less value yet the tier remains the same. The people owning the new sim property get added value simply because they get a new sim. Its the same as the Luskwood crew pointed out, they get boned because they were first and have paid thousands of dollars over a long period of time yet they get stuck in a cruddy, obsolete server even though their tier payment is the same.

There again too, what if the sim owner or owners don't want to be part of the "Museum" of how things were? I'm sure Luskwood is going to be really happy to know they get to remain on a laggy, feature backward sim so we who live on fast, feature rich sims can relive the past...

I'd love to see at least 4 or 5 other companies attempt some new 3D virtual environments where you don't have to kill people, looking closely at SL and the other ones like There and older ones like Active Worlds and Cybertown, examining them all for good and bad points, applying differing viewpoints on what to make and how to do it.

I would be perfectly happy with a massive changes that break existing content based on a different scheme from the ground up that would make radical improvement but I would be fearful that many of the changes in the interface would be for the worse, changes that make things harder to to do, less intuitive, require more keystrokes and more mouse activity, reducing the ability to use the keyboard to accomplish tasks, covering the screen with more clutter.

sure if it has the stuff that made me join SL in the first place. but i don't really think it's possible to design something that will never hit a point where it can't grow anymore or get so complicated it breaks. i mean if it's possible why didn't they do it the first time so this post would never have to happen?

but yeah if someday it gets to the point where it's easier to start from scratch than keep fixing poor old SL i'd have no problem switching to TL and starting over.

going along with the trend......i'd rather eat the spinach now and enjoy the coffee ice cream later.

I hope they make the skin tats easier to put on without sacrificing detail and or having the skin creator have to modify it.

Yep, me too. I'll take the brusselsprouts now please, for my ice cream dessert