If *you* had to design Secondlife 2, what would it be like?

I support everything Huns said just because I want to write LOL scripts.

I like this idea, but its not exactly polymorphism

What you describe is method overloading, declaring methods with the same names, but with different arguments.

Polymorphism is the uber cool typing system Java uses that allows devs to use an object as another type without thinking.

Take this example:

public class Animal extends LivingThing implements Feedable, Decoratable, Loveable {

    /** The method Feedable makes you implement: *.
    public void eat() {
        // Code to make animal eat.
    }

    /** The method Decoratable makes you implement: */
    public void decorate() {
        // Code that would make decorate() work.
    }

    /** The method Loveable makes you implement: */
    public void snuggle() {
        // Code that would make animal snuggle.
    }
}

public class Human extends LivingThing {
    public void love(Loveable thing) {
        // Code to make Human love thing. 
    }
    public void cleanHouse(Decoratable[] houseItems) {
        // Code to make Human be able to fix up her house..
    }

    public void cookFor(Feedable[] things) {
        // Code to make Human cook for many Feedable objects.
    }
}

Since Animal completely implements Feedable, Decoratable, and Loveable, it can be used in all three of Human's methods without any extra work. That's polymorphism (please correct me if Im wrong)

I think both polymorphism and method overloading will be useful for LSL. Espicially when llSay is concerned.



POINTERSSSSSS AHHHHH!!!!
Pointers really give me the creeps, and passing things like this really needs to not be implemented. Have *all* Objects be passed by refrence, like in Java, and all primatives (lists, strings, vectors, int) be passed by value.



Why not just keep it simple, just make people be able to define public functions. There's really no need for public variables, just use accessor methods, safer.


Also, there are some things Id like to see changed from your example.
Namely, where all the ll functions come from. (Changes highlighted in red)


Of course, all LOL objects would implicitly inherit from one central class (like in Java, its Object), if they dont explicitly extend from another class. This provides a totally cool set of benefets for everyone

In most of your other points, I agree with you. My main concern was with the C++ style Object Oriented stuff you were recommending. C++ is *not*, repeat *not* object oriented. Any language that allows for global *anything*, stuff that exists *everywhere*, stuff that doesnt have an originating class, is not truely Object Oriented.

==Chris

I loved the story Eggy =)

I wouldn't mind an alternate scripting language option, especially if it were done with ECMAScript.

Major changes I'd like include:

1. completely redesign ratings
   
2. completely redesign permissions
   
3. unlimited groups
   
4. hierarchial linking
   
5. hierarchial land ownership
   
6. add social networking (like orkut)
   
7. multiple attachments to the same point
   
8. more attachment points
   
9. way more info-gathering built-in functions
   
10. terrain texturing options
    
11. more-natural av animations (There has us heavily beat on this)
    
12. rework vehicles so they don't actually cross sim borders, but instead are handled by a more centralized spatial system
    
13. mute-by-source sounds
    
14. redesign the edit panel (more precision, more status options, etc)
    

It could be optional, as in perl. I'd definitely use it, since loosely typed variables incur extra overhead.

How do you know this?

It runs like a snail as it is! We don't need to slow it down!

Making it easy to understand and use for non-programmers? What for? Come on. That's like saying, hey, let's change the rules of the road and make everyone drive 10MPH so that people who got their learners' permits four hours ago won't feel so put-upon. Programming is not easy to learn. If someone finds it "too hard" to program in a language like LSL, maybe programming isn't their bag.

> As far as SL 2 goes, the biggest change I'd make is to remove the responsibility of social management from the Lindens' shoulders. Abuse reports would be gone, and Lindens would only respond to genuine software problems. While a poor business decision, it would be a far more interesting if we could find in-world solutions to our own dilemas.

Yes! This was actually how I got into MMOGs, because I'd been studying a little bit about politics, not a lot but more than nothing, and I was interested in the MMOGs as a way of experimenting and experiencing different political systems in a fairly safe environment, ie if it goes wrong you dont end up with tons of people with their heads guillotined.

I'd sortof forgotten about this, because it became apparent that in the anonymous environment that is MMOGs it would be a very big challenge to introduce player-run abuse-control. One challenge is that only a fraction of players are online at any one moment.

But it's definitely a big thing for me; it was the reason I always played on the PvP servers in Everquest. I was never particularly interested in killing people, but staring alive added a whole new social/political challenge.

One point: be careful what you wish for, because to have player-run abuse-control, the players must have the ability to "grief" other players: ie ban them, suspend them, or kill them. In other words, the world will probably end up being controlled by a powerful mafia. Not necessarily a bad thing at all, especially for people interested in the politics in the game, but potentially off-putting for many.

> I promise if Eggy's SL mythology turns into a religion/cult

I'll join

Azelda

You are correct, I was actually talking about method overloading.

Needs to *not* be implemented? Let me choose whether I want to pass something by reference or not. If you don't like it, don't put & before your variables. Damn near every language has support for explicitly passing by reference. The chief use of passing vars by reference is that you may want a function to work with more than one variable. Without pass-by-reference you have to use junky hacks, like passing by list or using globals when you wouldn't have normally.

The model I suggested was private non-encapsulated methods and variables, and public encapsulated methods and variables. This is consistent with C++, Java, PHP, and probably other OO languages. In any case, accessor and mutator methods are not "safer." They make sense sometimes, but they are not always necessary. You shouldn't say things like "there's really no need for public variables." It assumes too much.

This is gratuitous use of objects. One of the things I like about PHP that I don't like about Java is that it doesn't force every last aspect of the language to involve objects. We already have llGetPos(). We already have llGetFreeMemory(). There is no point in making someone type all that extra stuff. It would also make it lots more time-consuming to port existing LSL scripts over.

This is the sort of thing windbags like to argue about at coffee shops. It doesn't matter to me whether the program entry I specify is inside an object or not, because as soon as it gets compiled, the real program entry is definitely not going to be in an object. The program, once it is bytecode, has to be started up by (gasp!) free-standing subroutines.

The language either supports objects, or it does not. Anything beyond that is a matter of religion.

Eggy mentioned earlier that most people go through their lives on emotion rather than logic. He's right. I would never want to see anyone other than an LL employee holding any kind of enforcement power over me, with the exception of land access controls.

> Making it easy to understand and use for non-programmers? What for? Come on. That's like saying, hey, let's change the rules of the road and make everyone drive 10MPH so that people who got their learners' permits four hours ago won't feel so put-upon. Programming is not easy to learn. If someone finds it "too hard" to program in a language like LSL, maybe programming isn't their bag.

I dont agree, and I find this attitude a little snobby and elitist.

To use your analogy, simplifying scripting is like taking a Formula 1 car and putting in a clutch so anyone can drive. Replacing Brands Hatch by motorways that normal people can use to drive from A to B.

Azelda

They aren't? The Lindens bear a decided resemblance to the Greek Gods:
1) The Gods come down into our world, and do whatever they please, whenever they want to.
2) The whims of the Gods are frequently inscrutable.
3) The Gods do not always agree amongst themselves, and a canny mortal can sometimes play off these divisions.
4) People petition them with prayers, but only rarely do they acknowlege those prayers, and even less often do they act on them.


And on a parting note, "God created Man in his own image, and Man, being a gentleman, returned the complement"

Going back to basics. For me personally, what I'm looking for in a generic sense, like what I've been looking for since starting to play Everquest, is:
- a MMOG
- who's world I can modify, contribute to

Like, at a basic level, I had a picture for Everquest of being able to point at the ground, cast a MakeTree spell and Boom! a tree would appear.

My vision was that these would be spells taht would use mana, and larger creations would need more mana.

One of the bigger creations would be to create a new zone, and you'd actually need maybe a hundred people all meditating together in order to create enough mana to cast the CreateZone spell.

You'd be able to create a zone anywhere you like, and this would basically create a portal to the new zone.

This was my vision, and that's what I'm looking for.

To what extent is SL consistent with this vision?

- Well, we have the ability to create stuff. We can create trees, and we can build
- If we compare SL to my vision, in my vision we have the ability to create new zones, which would be empty 3d pockets of space where you can do what you like.
- We need a MMOG
- And we'd need building to be done by spells, based on mana

I think there's nothing here that SL is not capable of, and the amount of backend SL modifications to make this possible is very trivial in the grand scheme of things.

I see where you're coming from. You actually played EQ! Most people here havent, and the ones who did, would probably describe it as a "miserable treadmill".
The whole point of SL is to NOT be a "miserable treadmill" like most MMORPGs, TSO included.
The basic concept behind SL is freedom. This can never become a game, because a game has a strict set of rules, it forces predefined behaviors on you. In SL, the only rules are "Dont bother other people" and "Dont break the law". You can basically sum up all of the ToS in those two sentences.
The beauty of SL is, even not being a game itself, nothing prevents you from making games.
There's very little preventing you from implementing the annoying TSO needs, repetitive EQ monster-bashing, or even the most wonderful game you can imagine.
Just think, that its entirely possible to make an exact replica of Doom in here. It would take a lot of work, surely, but nothing compared to how much work it was to make the original one.
What you want is a GAME. SL isnt one. It's a world, a place, an environment, it's like the WWW but in 3D, where you can get a piece of a server (land) and put up your content for others to see.

Oh Im not expecting SL to become a game, but Id like it to become possible to make a game within SL.

It's funny that you are saying that it is possible to make Quake within SL. I seem to remember your saying:

"With regards to your first post, I acknowledge that I am not the best scripter in SL, but if you truly believe that any sort of interesting game can ever be made with the current feature set, then you're either easily amused or have absolutely no idea how much data needs to be stored and processed in even a simple DOS-era game from... 10 years ago."

( /120/d4/7901/1.html )

I liked Ama's reply:

"I believe at this time that you, Eggy, are entirely incorrect. I believe that SL has the potential to be a great development platform. LSL needs an overhaul (interface control, data structures, communication, perhaps an API for client side control of in world items/actions) but given the history of SL updates that is entirely not out of the question.

"Also I am interested in MMOGs, I did play EQ and I still play SL. Why? Because I am as interested or more in making games as playing them and SL does offer that. Just because you don't want MMORPG components in SL doesn't mean others don't or that they aren't possible or can't be popular.

"Just because you or others havn't been able to port a game in the past doesn't mean it won't be possible in the future. Or that other creative games, unique to SL, aren't also possible.

"If it is within Linden Lab's vision of SL then it is more than possible."

Azelda

I said Doom, not Quake.
Its pretty hard to make any sort of NPC in the game. That's why so many people want to be able to use avatars as prims.
I still stand behind my original assertion.
You can make games in SL, sure enough. You can make Doom. You dont see me playing Doom, do ya? You might see me playing the latest incarnation of Unreal, with far better graphics than anything you can build or script in SL.
Because anything that you can do in the limited environment of SL, you can do better with a real game engine without any constraint on how fast your code is going to work, how much memory it can take up, and above all, without the horrible lag inherent to your keypresses being streamed to the server.
A good game, for me, is one that takes up 5 CDs. Where are you going to fit the data for a real game, if the memory alloted to your scripts is less than what a computer had 15 years ago.
Even without these restrictions, you have to keep in mind that people here arent being paid. There arent any interesting games for linux because, even without the technical limitations, most people interested in making games for linux are teen-aged volunteers... who never finish what they start.
Show me something long, polished, good looking, immersive and thoroughly fleshed out like Baldur's Gate, Planescape Torment, or even Final Fantasy, and I will buy it off of you for L$150,000.

If it was my game i'd investigate ways to improve performance. Such as storing textures, and world data locally. Reducing band with consumption, and improving load time. I'd scale down down the gui windows to fit in 600x800 resolution, because many of the video card on the market perform almost 50% better at smaller resolutions. I'd add four point view for 3d modelers, and take a few other tricks right out of auto cad to make 3d alinement easer. Such as 6 click box definitions. I'd also look into the possibility of using another 3d engine, like q3a, because it typically take 3+ years to build an engine from scratch. IN allot of server based games such as ever quest, shadow bane suffer from horrible inefficient rendering systems.

You can have up to 1 gig of cache.
Me and a few others have whined about LL not letting it grow bigger, and at some point LL responded with something along the lines of "making it bigger would not result in any performance improvements, as it would take too long to search the cache"


I endorse this! I wouldnt mind running SL at 640x480 if it would give me better framerate.
Thing is, SL's performance is more limited by your CPU and network connection than your graphics card, and SL is designed to run well at only 10fps. Yeah, just ten. Amazing isnt it?

> You can have up to 1 gig of cache.
Me and a few others have whined about LL not letting it grow bigger, and at some point LL responded with something along the lines of "making it bigger would not result in any performance improvements, as it would take too long to search the cache"

This is maybe a key point. A b-tree index will give a search speed that runs something like logarithmically with search-space size. You can have arbitrarily large search spaces and a b-tree will be pretty damn fast.

On the downside, creating a b-tree index takes a certain amount of time, and you've got to rebalance it every so often and so on. In other words there's some sortof compromise between storage time and retrieval time.

I'd tentatively suggest that the indexing process used currently is optimized to use up to 1GB of cache, but that doesnt mean that a future version of SL or equivalent couldnt use a retuned cache index to use tons more space.

Given that many games *require* 2-4GB of diskspace just to install them, it could be useful to retune the indexing parameters (pagesize and so on I guess?) for up to 128GB of diskspace. (thinking Moore's Law here, 128GB drives will arrive pretty soon, might as well take some advance on them)

Azelda

Well I'm running a 120 gig drive myself and it was pretty cheap last christmas.
There's 300 gig drives out there already...

It isn't snobby, or elitist. Programming is a complex subject. It requires (gasp!) ACTUAL EFFORT to learn it properly. I don't want to be saddled with a crap n00b language. In fact, I suspect someone who learned programming on such a language would eventually tire of it, as most people who start on BASIC move on to bigger and better languages.

BTW... Formula 1 cars do have clutches, were you talking about automatic transmissions? F1 cars already have paddle shifters. The driver tells it to upshift or downshift. The transmission takes care of the dirty work from there on. I don't understand why an untrained driver would expect to compete in a proper Formula 1 race. You actually have to train to do this, just like you have to actually learn things to be a good programmer. I can't imagine how screwed up and dangerous a F1 race would be with untrained drivers. Sort of like programmers writing tons of spaghetti code. Wouldn't you agree?

So, I'll leave *that* end of SL2 to y'all....

Here's my suggestions for SL2, most of which have to do with increasing the aesthetic portion:

- increase max. prim size : 10x10x10 just isn't big enough. double that and we'd be (more) in business.

- ability to shift hollow on X/Y axis : why should the hollow cut be *only* in the center of the prim? i'm sure there's a good reason, i just don't like it.

- varied terrain textures : allow the property owners to set terrain textures and increase the terrain texture 'library'

- better trees. do something with them. i love trees.. loads of them, but we need more variety and some way to display them w/o bogging us down too much

- *all* linden created textures available to the public.. why shouldn't we have them?

- wider waterway preserves for boaters and a few Linden-owned runways for aviators

- switch the avatar model from 'male or female' to 'start with form male or female' : allow *all* options regardless of gender.

- add a slider for chest area to stretch clothing over breasts or have it cling to breasts... you get the idea

While most of you are concerned with the more specific features you might change or keep, I propose a simple change in philosophy while retaining most everything else about SL as it stands.
I think Linden really had something when they proposed a one-time ten dollar fee to just be an unlanded user indefinately. However, I think they should have taken it a step further. In my estimation it might be to their benefit to make a clear distinction between "developers" and "users". All the features that make SL so rich for entrepeneurs and developers make it clunky and daunting for people who just want to socialize and check out the creations.
If I could overhaul SL I would have ten dollar user accounts that offer NONE of the creation tools or the debugging options and make it as easy to control your avatar and use objects as it is in..say, Unreal Tournament. SIMPLE. Move around, press a single button to use the object you're facing, look around, and chat. That's it. Simplicity in control is the key to basic user accounts for those not interested in developement. For those who are interested in developing and making money you upgrade your account and basically get the interface everyone already uses now. That's the one major change I would make for an SL 2.0.
I think a lot more people would take to SL and stick around if they could very easily move about the world and try out the things other people create with less complication. Second Life has the potential to be as useful and ubiquitous as the web but, like the web, it should be simple for users, complex for developers . Just my thoughts on the subject.

Persig

Not sure what I'd call it

All the porn

More dance moves, easier stripping controls

I'd add the ability to charge money just for seeing my AV

Sex sells

Neon lights everywhere, giant billboards, nekked people everywhere

nope, herd everybody inside to the clubs and casinos


**just being sarcastic**

Personaly I only take other peoples scripts and change them how I please.

Perhaps if the Lindens gave us better scripting tutorials...

> You can have up to 1 gig of cache.
> Me and a few others have whined about LL not letting it grow bigger, and at some point LL responded with something along the lines of "making it bigger would not result in any performance improvements, as it would take too long to search the cache"

Just create a cache subdirectory per sim. Voila, scalable, indexed cache.

Azelda

The ability to animate linked prims individually creating realistic object animations.

Make a polygon limit and allow people to upload 3D shapes
from software like 3D Studio Max.

Allow linked objects to have a simplified, user defined physics shape to reduce strain on sims. Perhaps just a box you make over your object.

Allow scripted objects to return to inventory when they no longer detect your AV. Its too easy to loose expensive items.

Function similar to llScriptDanger that only retuns TRUE if the area = no script. I use llScriptDanger to prevent following objects from being stuck but they wont go into many places that allow scripts.


Improve sim transitions to allow better vehicle movement.

Functions to allow better AI object behavior. Something that will keep groups of animated creatures from banging into eachoter while moving in groups.

Function to let scripts detect collisions with AV attachments not just the AV. (swords). I want to create creatures you can battle with swords.


A way to sync character animations between AVs. Hugs, Kisses, handshakes.. regardless of size difference.

Maybe a set price on land in particular sims based on overall dwell. This would prevent land barrons from driving up prices and perhaps give residents motive to increase land value.

Allow me to delete more than one object in inventory at a time. Add an empty lost and found option.

A tool to create AV animations.


Really I can not count how many things I would not change.
I LOVE this game.

You rock Az!