Start work on SL 2.0

I've already submitted it as a suggestion in jira. Please vote or explain why you are against it:

https://jira.secondlife.com/browse/MISC-255


LL should consider pushing certain features into SL 2.0 and rather than attempting to maintain complete backwards compatibility with SL 1.x, they should design from scratch with an eye to eventually making tools that allow existing content (prims, scripts, etc) to be converted to run properly in 2.0 sims rather than try to turn SL 1.0 into 2.0 directly. If a landowner converts to 2.0 he/she gets a discount on rent from LL. LL could even bring back the "first land" thing for a while for people willing to adopt 2.0.

AVatars that enter 2.0 land do so knowing that some/all current stuff won't work right, but see below for a possible workaround.

The rationale for doing this is simple: it is generally easier to design certain bugs out of existence than it is to "fix" a bug that is based on a fundamental design limitation. It is certainly easier to design from scratch for new features than to add them to a system that was never meant to use them.


An agent that enters 2.0 could have TWO inventories: one for 1.x and one for 2.0. ENtering/leaving 2.0 would activate/deactivate enable/disable the appropriate inventories, attachments, clothing, avatars, animations, etc.

This may sound cumbersome, but compared to the hoops that LL will need to go through to make SL 1.x work with a new physics system, and implement other changes of equal magnitude, this proposal is almost certainly an easier solution than implementing a new physics system that is backwards compatible with existing content and if they are designing from scratch, they can design all sorts of neat kool stuff that just wouldn't be practical to implement in the current SL, no matter how long they took or how many warm bodies they threw at the task.

People should vote/comment on jira concerning this (LL wouldn't even consider this proposal unless there is strong grassroots support from the existing community, I am certain), and this thread could be used to discuss what you would want to see in SL 2.0, and suggestions for how to do it.

https://jira.secondlife.com/browse/MISC-255



Cheers.

The most important new feature that almost requires (maybe it DOES require) a SL 2.0 strategy is a new physics engine. It would be nice to see better physics, AND a real link-hierarchy for prims. A special form of the hierarchy might be implemented for skeleton animation which would allow specific prims to be referred to as joints, and allow avatar-style (or even script-controlled) animations to be applied to a linked set of prims where multiple prims made up the space between joints. Special commands could be created to make joint-control and animation faster than regular scripting.

In other words, you could have Joint_A=>Bone_1=>Bone_2=>Bone_3=>Joint_B etc

and have other stuff linked to Bone_1 in its own sub-tree of prims. The animation/script would control the joints and control the other prims indirectly. If LL wanted to get really fancy, they could allow sub-animations of sub-skeletons as well. That flex-prim growing out of your pet robot's back is really a tentacle... is that a GUN???

Battle-oriented sims could get very interesting. With distributed processing, and/or client-side processing of animation-control/AI, things would get VERY interesting in a battle...





.

This depends a lot on how they've been coding the system so far, decent Object Oriented design allows a system to have parts changed easily. A physics engine, if kept modular enough, shouldn't be that hard to rip out and replace with a newer one so long as the function calls required haven't changed. Presumably Havok X has the same features as Havok 1, so it shouldn't be too hard to just replace them with the new version, hopefully getting improved performance as a result. Work can then begin on looking at any extra features that are offered that can be integrated.

If the system is poorly designed and not modular at all then yes, redesigning it could help considerably, as a good design from the start can make a piece of software limitless.

However, I'm not sure that an SL 2.0 necessarily needs to break existing content, it's perfectly possible to completely redesign the code of a system and still maintain the same functionality. While the end result may not be especially impressive, the result is a system can have additions made on a whim and broken parts easily isolated and fixed.

There is enough whining around already that I try to ignore. Too many features.. Not enough features... Avs too old... Avs too young, sheesh.
Now this!

Why don't you start you're own version. Instead of telling LL to do it now for you. It will when it's time.

Nothing personal, just get on with it...

I suppose when the coming competition gets here, LL will wish they had looked at the option for rewriting the base code and upgrading their physics engine. They apparently are really stuggling with scaling and I believe they have some real basic design problems probably dating back to the original code. A major update such as SL 2.x with lessons learned from 1.x avoided would help LL greatly in any future expansion or growth.

But, who knows..........maybe one Wednesday we will get Second Life 2.0

Of course struggling with scaling is their own fault so I don't have any sympathy. They knew it would happen when they opened registration.

Actually, according to the Lindens, the reason why they haven't replaced Havok 1.x is BECAUSE Havok 2+ changed so much that it will require a c omplete rewrite of all the code to handle it. The reason why they took features, like link hierarchies out of the game, according to the Lindens, is because SL 1.x and Havok 1.x weren't compatible in that area, so while you would THINK that what you say is true, it just isn't. What they have explicitly said is that they will rewrite things before they implement new physics.

But their usual way of doing things: making sure everything is backwards compatible, will obviously be MUCH more difficult than what I propose.

Also, while it is possible that the current system has an OOP design and implementation, there's a principle of "code twice and throw the first away" that applies to Smalltalk (pure OOP) as much as assembly language programming: you learn so much the first time around that you can make it so much better the second time around that it is worth the extra time to code it a second time without all the design-errors from the first.

ANd of course, this principle will continue to hold. Eventually, even SL 2.0 will be "outgrown," but hey can use what they learned from the transition from 1.x to 2.0 to make an easier time of it in the transition from 2.x to 3.0 when that becomes necessary in another few years after 2.0 is released.

Wait... Are they making a 2.0? I've never read anywhere that said they were making a new SL???

I don't think they have said they are developing a SL 2.0. But logic would surely dictate that such move is coming. That's assuming, of course, that LL wants SL to succeed.

What they appear to be trying to do is give SL 1.0 a complete makeover and morph the entire game into 2.0. That is just the wrong way to go, IMHO.

Yeah, but I think a big reason for that is how much people have invested in SL 1.0. Exactly how feasible would it be to make a completely revamped Second Life 2.0, and find a new audience? A lot would go to the new one, but what about store owners, land owners, people invested in Second Life? They'd have to start over with the new client, or would LL transfer all of what we are into the new one? Would it even be compatible at all? It probably wouldn't, and would frustrate a lot of people, which is why I think they are trying to morph SL itself.

Personally, I will be VERY surprised if they can morph 1.x in a stable way, but even if they can, it will be faster to design clean and add morphing/importing technology/tools later, rather than to include them from the start.

Quite possibly the hardest part of morphing will be to keep things compatible with all the patches and Band-Aids and pasted-on new features. The longer they take before deciding to throw up their hands and go with a clean design from the start, the worse off they (and us) will be.

It will be easier to design a new system with ease-of-porting in mind then to morph the old system into the new. At least, that is what every other computer software company has found out over the years.

Yeah, and meantime we all suffer for it. EQ is still running, despite EQ2 being in existence. A lot of older games remain even when newer versions are made. In a way it'd be almost nice if they'd just make SL stable, make a new version, and run both.

There's a nice (and somewhat lengthy) article about this sort of thing (rewriting from scratch).

http://www.joelonsoftware.com/articles/fog0000000069.html

While LL *could* spend lots of developer resources on rewriting the client and possibly backend architecture, what happens while they're doing that? Do they let the 1.xx client sit and rot while people get very frustrated with bugs for a few months, maybe almost a year?

I don't know how much developer force LL has, but if they had enough to work on 1.xx and 2.0 in parallel, they probably would've done it already. Instead, they're doing incremental updates. While things aren't getting fixed as quickly as some are desiring, it IS slowly getting better. Whether it'll take too long to happen before everyone leaves is another matter.

How is it slowly getting better? Visually? Minimally. Stability-wise SL is NOT getting better, it's getting worse. I understand they are working on it, and I give them a lot of props for taking on such a big and relatively new project as this, but it is NOT getting better. I could go through the same old list that many, many other people have gone through, inventory loss, presense issues, teleporting issues, the list goes on of things that are consistently not fixed, in-fact getting worse time and time again. A year ago before flexi and light prims, SL was more stable than ever. Now? I don't think so.

Don't get me wrong either. I do alright in-world. I get heavy lag in heavily populated areas, but most of the time I only lag in sims that are laggy themselves. I can live with the lag too, personally. It's frustrating, but it's not the end of the world. But compared to how it was when I made my first account? Things are definitely a lot worse today than they were then.

Admittedly, I don't use SL that much.

Regardless of whether it's getting better, the point about rewriting everything still stands. I'm guessing it would take much time and effort to just drop everything that they've done so far and redo the whole system. While it would satisfy everyone greatly if they did so, how long would they have to wait for it?

Actually, it takes LESS time to redesign from scratch then to patch the existing stuff, at least when you're talking about something as major as rewriting the entire system to allow for a new physics system.

Better to start from scratch and figure out how to painlessly and automatically transform everyone when they enter a new 2.0 sim, then to try to morph the 1.0 system into 2.0. That can be done on a sim-by-sim basis, while the way they are currently talking about doing requires that everything work perfectly from the start, or no-one can play at all.

Hmm I think it's improved vastly from November-Decmeber when it collapsed with only 12,000 online and search was disabled, teleports were 50/50 at best. Sim crossings usually crashed you etc.

Trouble with running 2 grids in this case compared to other worlds is the Land rush it would create to buy in the new grid and money lost to owners in the old grid, basic people would just switch to whichever runs ok. You would be paying the ridiculous tier still on land worth nothing. I think perhaps transfering existing land to the new grid might be fairer way to start.

Imagine that Island brought for $1500 USD now could be worth $100 within days of such an announcement. You could come back from holidays and find the island you rented on and all your stuff there gone.

That would be an important consideration also. Perhaps the highest priority should be given to converting existing sims to 2.0, but I still suspect it will be easier to convert the 1.x land and its content directly into 2.0 than to attempt the conversion incrementally within each sim. perhaps a compatibility layer that presents the 1.x functions could be implemented, and maybe that is what they are trying to do now, but the description of how they plan on implementing mono certainly sounds bug-prone, and I think that physics would be an even bigger headache if they try to do it the way they say they will do mono (parallel implementation running in the same sim at the same time).

Don't forget to vote!

https://jira.secondlife.com/browse/MISC-255

[bump ]

There's a companion thread discussing a "wish-list" for Second Life 2.0, found here:

/13/50/188821/1.html

[bump ]

They never paid attention to bug reports from beta grid before, I'm not bothering with JIRA. Trust is earned.

Eh. There are ways of bringing public pressure on a company if they don't pay attention to their customers and you can document it, which you can, now that jira is the official way of reporting bugs and suggestions.

Which they've still been shown to ignore. The public pressure is no different from before. The bugs are no different than before. JIRA is not a system I trust.

Jira only became viable last week, IIRC. Give it a month or 3 THEN call it a bust.