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

I've already posted LSL suggestions in another forum some time ago, so I won't bother repeating them here. Hopefully the Linden's have already noted them. (Though I could probably find them again, if pressed.)

Since then, I've gotten a few more ideas:

1. Upload bump maps, too; this can go a long ways to enhancing texture realism
   
2. Ability to "hollow out" existing polygons; supported in every other 3D environment, why not here? Let windows be windows, and not crude texture transparency hacks.
   
3. ability to take remote snapshots; e.g., implementing offline security camera
   
4. some primitive image manipulation for touch-ups
   
5. polygon meshes
   
6. NURBS surfaces
   
7. user-only overlay display
   
8. abilityt to "blend" more than one texture on same polygon face
   
9. image procedurals to dynamically create texture patterns
   
10. "advanced" options for 3d-heads to tweak, well, everything; e.g., back-face culling, control over specular, diffusion, etc.
    

However, the biggest thing to me is an LSL re-vamp. It's a a fine ad hoc first hack at a scripting language, but it's truly horrific. It's time to have a mature scripting language. Either leverage off an existing language (e.g., python), or create a nice clean O-O language.

Wow, you really are a gamer.

You do realize that this would make SL unplayable for many. Not just off-putting. If SL was run by an in-game cartel, places like ours simply wouldn't be able to exist. Money flows up in those situations, and if a build wasn't a casino, high-income store or club, it wouldn't exist. You'd see very few noncommercial builds. Not to mention I'm almost positive that it'd become open season on some categories of players. Yes, it'd be a very different world. And for folks who can't get enough of the might-makes-right/those-who-have-the-gold-make-the-rules mentality, yes, it'd be heaven.

Personally, I also am bothered by your suggestion of "earning rights to rez prims". If that were the case, everything we do at Luskwood wouldn't be able to happen.

But then again, I'm sure I'd be in the minority.

I have a new idea for what I want in future versions of SL.


I would like a completely working VR interface so that I could be my AV in world.

But there'd have to be some sort of in game interface for changing my AV's appearance.

With the VR interface I want to be able to really touch objects, create, delete (my objects). Think of the 'hands on' experience that would be.





of course the lag would drive me nut!
Be like walking/flying in pea soup!

I hope if there is an "SL 2" that they honor the original lifer program.

You can hope, Angel... but I'd somehow doubt it. The likelihood is that they'd say that you had a lifetime account to Second Life '1'.

I could be being overly sceptical of course... I'm just going by what the attitude of most corporates would be in similar circumstances. Who knows? Linden Labs might just rise above the pack should the fantasy become reality.

Actually, I have asked Lindens before if the original Lifer program will be extended to future versions of SL, and I was told they will. I just posted my hope here because I want them to remember to honor us in the future.

In that case...

Cool! I'm happy to be wrong!

1) 2d functions - I sware, I have spent more time in our lovely 3d world trying to find good ways to implement 2d images, text, etc. We should have the ability to directly write and draw on a prime's face. Uploading a freaking hoard of textures to handle stupid things like text is ridiculous. Waste of time, waste of hard drive, waste of bandwidth.

2) A bit out there, but before SL came about, I was very much into text based MUDs, in particular ones that used the LPC language. There were many great concepts in LPC, and in general was very much superior to LSL. May be worth checking out....

3) I agree with the ability to gave physicless sims, perhaps not the whole world, but if particular sims were designated "non-physical" there would be higher timeslices to things I believe most people use more. Just basic positioning, and perhaps a simple AV->object collision detection.

4) To be continued when I have more time....

I love lots of things about SecondLife. But, what I'd honestly want in a virtual world is a hybred of SecondLife and There.com. Neither has all the features that a person desires. Yet, I love both worlds!!!

In SecondLife I love:

The building tools, the fact that we can make our avatar into almost anything that we can imagine, the animations rock, parties, instant submission process, the freedom of mature zoning, the infinite possibilities to create here & much much more.

The things that I think need improvement:

Privacy - I am thinking about how houses work in "There". No one can see inside your house, no one can 'hear' what you are saying inside, unless they are inside too. You can ban ppl from your house, etc. So, once you are inside - you have your own private chat. I'd love to see something like this added to SL.

Additionally, I'd like it if we could appear as invisible to the map. I'm not fond of the current tracking devices via the IW map.

Vehicles - I'd love to see improvements in vehicles. I notice that when we try to cross a sim at a higher speed - it crashes me. The vehicles don't have the speed or feel of "There" vehicles. I really miss that alot.

Loading - Not sure how to express this. For instance, if you try to fly a plane at a higher altitude, you crash into debries before you can see them.

Atmosphere - I'd love to see a better rendering from up high looking down on the world. Somewhat "There-like". From on high looking down in "There" you can clearly make out the islands, oceans, etc. I would love to see something close to that in SL too.

Games - I'm used to being able to place quest clues around the world, as an example. And, we can not do that in SL - as ppl own the lands. I am wondering if there could be "LindenLands" which would allow you to apply for priveledges to drop items for games. This would be in various spots throughout the world. Sort of like public parks or something. Just an idea...

------thanks for listening--------

I'd like to see:

1. more control of my avatar. I want to point at anything, instantly, without planning ahead, I want to look like I'm reaching over and picking something up. I want to hug someone, or shake their hands. If I could just grab my avatar's hands with the mouse and move them about it would be great.
   
2. X11 client (NetBSD)
   
3. inworld spell checker
   
4. external compiler - if only for syntax checking
   
5. regex, all the stuff that's trivial in perl
   
6. ability to rez a dummy avatar and control it with a script
   
7. better documentation (at least run the current docs through a spell checker!)
   
8. increased max prim size
   
9. prim negatives/subtraction
   

First some coments on what other had to say:

LSL should not be simplified in any way, if anything, more functionality should be added; I aggree to teh fact that doing something like programming would take *ghasp* effort. Reducing this comment to anything less is just dumb.

SL should NOT be governed by the users, i dont know if you have looked around lately but i personlay have verry little faith in the general publics ability to have good discretion above the desires in the favor of individuals.

There should not be diffrent levels of functionality for diffrent client accounts, had i not be able to build and script right off the bat there would have been a gynormous chance taht i would not have registered / stayed with SL.

Custom object modeled outside of the SL system based on a ploygon system would completely defeat the pourpose of using prims, the hole point of teh prim system is to take the load off the server bandwidth wise, custom objects would double the badwidth needed to transfer content to users. The size of information to trastfer loc rot and deformation properties is exponentionaly less then if tey were to use actual vertext based geometry.

Improvements for SL itself:

File sharing / transfering system where the bandwidth is completely independant of the SL system and SL would simply be the method of connecting to other people. choise of protocals, compression and what not can be fought and died for later ;D

Voice chat system. I know this would increace the requred bandwidth load on the servers but it would be a wonderful feature.

A in environment GUI / scripting language, where instead of cerating imags and prims for everything you could use a kind-of HTML-esque markup language and display it all on the surface of an object, things like images, fourms, buttons, text etc. And, if yoru feeling crazy make taht information / meadia interactable i.e. like the Doom3 interfaces where instead of opoing up a window to do everything (though taht wouldnt be bad for say filling out inworld fourms) you could click on items on the surface of an object, mabey even replicate web documents onto teh surface / face of an object.

llSteamSound(string source, float volume, int local/global)
Mand would i like the ability to on the fly generate sounds from a streaming source to share with others or just make it so if i want to stream audio just to my local client without the need of using winamp etc externaly. Would open up the possibilities for DJ setups, boomboxes, vehicle "systems", personal headphones etc. and if you had a fourm creation system and llList2Notecard one would be able to create playlists etc.

Obvously every one wants a higher prim per m^2 ratio but im just going to say it any way becuase I know your all thinking it ;D

A tree based linking system where you can define a hirechial strucutre to the set of objects, so you can have parent, sibling and child objects. Also the ability to offset the "center" of an object. Assuming there will be improvements to the phisics engine this will add the ability to do very interesting and productive things.

Now for an idea that would be more seprate from SL:

How about the ability to have a personal server, kind of a local world, one that others could connect to but would not be apart of the SL universe but could still be linked to from SL. It would essentialy be the same as how a Activeworlds server worked, where the server had to be ran some where and reported to the SL master system weather it was online or not. everything like textures etc would still have to be uploaded to SL and still have upload fees but the environment would be able to run independantly of the SL world.
If not a seprate system then atleaste allow for an offline development system that would be like a small personal SIM where you can work on what ever it may be but when you got to move something onto the SL system it would notify you of all the resourses requred to transfer the media to the SL system and the appropreate L$ requred.

Another idea would be to make a completley LAN based version of SL where you could setup a server and have clients connect to it and have the entire system run independantly of the SL system.


Theres probably more but its 3:35 AM and the brain is begining to slow so I'll have additional posts at a later date

I agree as a programmer, but I don't agree as Joe Resident of Second Life. The language can be quite the barrier to someone who doesn't understand it. The curly brackets, the paren's, not to mention the cryptic syntax of the properties of llCommands ... it all points to a steep learning curve.

So, perhaps the answer is to offer two programming language options. One for us more seasoned coders and one for the less seasoned, a la Micr0$oft's Active Server Pages allowing you to use either VBScript (BASIC-esque language) or J-Script (M$'s javascript-like language). It only needs a language declaration at the top and you can code how ever you see fit.

Ideally, an object-oriented and more GUI WYSIWYG-type of scripting system may work out better for newcomers to programming, where all options are laid out in a context-sensitive palette, but that's prolly more like SL3

My Vision of the inWorld:

"Hammerspace". Along the lines of Az's idea, we should be able to define an area within the global SL grid as a link to a hammerspace, or where cartoon characters get hammers and things from seemingly out of thin air. Imagine building a quaint little cottage and you step inside and it's an immaculate mansion, 5 stories high, fountains, pool etc. It would solve a lot of problems with land if we could create micro-metaverses within our metaverse.

True Interavatar Interactions. Like TSO (and some of the amazing scripted attachments i've seen lately, OMG the hug attachment is the coolest thing EVER) we should be able to interact with other av's fluidly. We should be able to impact them more than just bumping into them. A slap should get the other av's head to fly back, stumble and then rub their face where they got hit. Damage enabled sims should actually damage you where you're hit -- visibly and from a movement standpoint. If you get shot in the leg, you should fall and no longer be able to walk, or you should stumble). Maybe this is better left to the scripters tho.

Better Prim Geometry. Building, while pretty intuitive most of the time, can become quite the headache when you are attempting to achieve complex shapes. We should be able to define the axis we cut on, hollow on, dimple on and "top size" (how we make cones and trapezoids, I'm not sure of the actual term) on. We should be able to choose to have a prim stretch from the center (default) or any of the grab nodes (using a simple diagram of a cube, you click the node you want to anchor or stretch from and typing 10 stretches it to 10 from that point, not the center which is sooooo annoying when lining things up). The UV map on funky shapes needs to be updated. Ideally, we should be able to use booleans (drool) but that'll never happen .

New User Orientation. Prelude Island needs an upgrade, badly. New residents come into SL with the most basic of skills and need to ask people tons of questions which we've all answered 80,000 times before. I'm a Mentor and I don't mind doing it, but the basics are the basics and if Prelude is supposed to get them ready to embark on life in SL, it should do that. Teach them to take an object out of a box, please! No more newbie cabins on their heads! Teach them about land and some basics about building, or let them explore a more advanced Prelude after getting to the Graduation Temple where if they have the time they can hit on these skills. ALLOW THEM TO GO BACK IF THEY NEED TO. Make Prelude hide all avatars from other avatars, and you won't need to worry about newbie griefing. Objects you rez in Prelude should only be client side to accomplish this. Or, build Prelude into the Client so you can visit it anytime (for that matter, build an offline simulator into the client. Oooh, I should make that a point )

Offline Simulator. Build an offline simulator into the client where we can experiment with scripts, crashing our client till our hearts content with hinge physics. You should be able to build something offline, play around with it, design it to fit your needs perfectly and then import it into the World for maybe $L 50/100. All those workshops people get ejected from would be a thing of the past. From what I can gather the client should be able to do this easily, having a prefab fake sim within it (sorta like the Construct in the Matrix).

I'll keep thinking, but these are my must-haves. Great thread idea, I hope the Lindens are watching

> A in environment GUI / scripting language, where instead of cerating imags and prims for everything you could use a kind-of HTML-esque markup language and display it all on the surface of an object, things like images, fourms, buttons, text etc. And, if yoru feeling crazy make taht information / meadia interactable i.e. like the Doom3 interfaces where instead of opoing up a window to do everything (though taht wouldnt be bad for say filling out inworld fourms) you could click on items on the surface of an object, mabey even replicate web documents onto teh surface / face of an object.

I think this is an original idea? It sounds pretty interesting; I think it's technically doable without too much difficulty. Not trivial but not mindnumbingly difficult either. Text is displayed from bitmaps, and you know the coordinates of each letter; we already know the coordinates (3d and screen) of where you click, so it'd be fairly easy to calculate which word you'd clicked on. The hardest bit would be the markup language interpreter, which is not really hard, at least at a first level of approximation.

> LSL should not be simplified in any way, if anything, more functionality should be added; I aggree to teh fact that doing something like programming would take *ghasp* effort. Reducing this comment to anything less is just dumb.

Dont let ease of use fool you. Easy to use != not powerful. LSL needs to be both more powerful and easier to use

> Better Prim Geometry. Building, while pretty intuitive most of the time, can become quite the headache when you are attempting to achieve complex shapes. We should be able to define the axis we cut on, hollow on, dimple on and "top size"

I agree with this, but it's harder to do in practice than it sounds

> We should be able to choose to have a prim stretch from the center (default) or any of the grab nodes (using a simple diagram of a cube, you click the node you want to anchor or stretch from and typing 10 stretches it to 10 from that point, not the center which is sooooo annoying when lining things up).

This is fairly easy to implement, once you know it's important.

> Ideally, we should be able to use booleans (drool) but that'll never happen

Booleans is also something which I think would be really useful but is really hard to implement in practice. Booleans are not an integral part of OpenGL, so youd have to run collision detection between everything then work out the resulting planes and polygons of intersection... ugh Adding more advanced cutting and hollowing is probably tons easier.

Azelda

P.S. Stream of consciousness on booleans: just thought: you might be able to create anti-objects by alpha-texturing them and rendering them *before* the object you want to cut, so that the z-buffering depth-test prevents the cut-away bit of the object from being drawn? It's an implementation-dependent hack, but...?

Physics could be tricky. Off the top of my head, I can think of ways to "hack" the physics to work more or less, so might not be impossible to do some advanced hacknig to get it to work somewhat realistically.

Hehe Az, I almost put "antimatter prims" but stopped thinking either a: noone would know what I was talking about and b: it would be too hard for LL to implement.

OOH forgot one of my oh-so-important ideas that I had today ... METRIC/IMPERIAL CONVERSION OPTION!!!!!!!! It really can't be THAT hard to make happen ... it's just applying a formula to the numbers really. I was working on a design for an RL object in SL trying to get the size right and the calculations were a headache, since materials are generally sold in imperial/english measurements and SL uses metric. While metric is hella easier to use it would be nice to work in inches optionally. It should be a check box in the build UI that swaps the numbers from metric to imperial. This would also make a lot of real-world builders see the appeal of SL as a layout tool for RL projects.

-- edited to add my metric/imperial argument
-- growl ... edited to add fix my lack of a /i on the previous edit comment

In reference to the scripting thing, as a beginner a sort of 'scripting wizard' would be nice. IE I tell it I want to make something disappear, it gives me a basic function to do so. Then later one, when I'm more experienced and I know how the nuts and bolts, I can forgo the wizard (or use it to set up a basic template) and just script 'normally'.

If LL would allow us to write to notecards we could make our own script generator ... :-/ ... but someone could have it dump output to the chat console. You'd hafta clear the Object whispers: stuff from the output tho, but a wizard is within our (residents) power.

If I were designing SL 2.0, I would want the following :

I. Create a scalable architecture

As it is now, a sim comes to its knees as soon as there are 20 or 30 AV's gathered. We currently have a situation where one sim might be coming to a crawl, while most of the other sims are mostly idle becuase there's virtually nobody there.

Create an architecture where computing and networking resources can be directed to where the AV's are.

If the sims could scale to 100's, if not 1000's of AVs, I think you would see some pretty interesting "Events".

II. Improve the physics engine.

Havoc 2.0 sounds kewl. Havoc 3.0 sounds kewler. Integrate these with SL, and who knows what the mad genius creators of SL will come up with.

III. Add more primitive shapes (OR... support wire-mesh models)

There's only so much you can build, given the limited number of basic prims we are given. A wider selection of "lego blocks" would allow for more attractive builds using fewer prims.

IV. Publish the LSL byte-code format.

This way, those that want object oriented languages, or more "BASIC" languages could implement them.

- Ace

edited to correct typo

Mmm, I've wished for this as well. A good arcing spiral, for one. Vertex manipulation would be nicer. You can stretch. You can rotate. But you can't grab a corner and move it -anywhere-.


Also... please, please, please... update SL's player models to use Poser 4 models, not Poser 2... Free up the face and fingers for animation/manipulation. Tired of holding stuff in an open hand.


Next, allow ingame manual manipulation of the limbs. I'd like to be able to adjust pivot points, or make the leg stance wider than the appearance editor allows (for my new vulture AV for example, which right now has its feet stuck together thanks to its oversized toes).


Finally... allow setting of some default limbs invisible. To keep people from exploiting make them keep at least one visible, but... well, it would be nice to be able to ''remove your legs'' and stick a horse body under ya or something, without having to mess with making human legs that match the look. Ahhh, centaurs!

Although I understand why they don't allow arbitary meshes, I can't believe that allowing constrained vertex manipulation on even just a cube basis would cause a huge overhead. Surely that just sending 6 verts.

It's not six verts...it's three settings that determine it's shape(same with all prims,read back later into the depths of fourm for more on this subject) if you add the ability to manipulate a box...it no longer has those three settings it has 112(is)now...how much easier is it to send three lines of data...then it is 112? Yah there would lots and lots of lag not to mention odd UV distortion...in theory for all I know they may scrap the current prim system for something more streamlined which I could go for, but it won't happen with in the current engine I'm guessing(Prove me wrong Lindens please!)

I put a vote on those feature:

1. Vectory system used for hair, grass, fur. These system should define
- how many vectors for a m2,
- average lenght (and % random variation),
- Color (and % random variation),
- elasticity (and % random variation), this define how fast is the response to gravity ro wind or other forces
- texture to be used (where alpha define if there is more or less vectors)
Movement and position of vectors are calculated client side as they count as an object, not per vector.

2. Of course, bigger prims

2bis. Boolean operations between prims (add, subtract, multiply)

2tris. More prims type, latheing, extruding and extruding on a path operations.

3. interactive textures, like sort of webpage on a face of a prim (well explained by others b4). this should be wonderful for programming personal devices or vendors and use less prims

4. Better collision detection between avatars

5. sims with active weather, clouds and fog (not haze, fog , as already said b4

6 no lag. Means faster servers, more bandwidth, better optimized streaming protocols. While sl is already good, I think user experience should improve.

7. Reflective surfaces

8. better lighting system

9. Video conference in IM, at least between 1 to 1 IM

10. ability to use movies as textures

11. ability to apply textures to land too.

12. scripting animation (like make a finger point to an eye or a knee touch the ground)

--------

Some of things i said could already been talked about, some better explained, also maybe i wrote stupid things. Please talk about the feature not the poster.

Several of you are confusing the syntactic complexity of LSL with its semantic complexity. Syntactically, LSL is one of the simplest programming languages on the planet today. Old BASIC is slightly simpler (though not by much), but its slightly smaller syntax has the effect of making the meaning of BASIC programs pretty unintelligible, even to experts.

Syntax vs semantics is somewhat of a tradeoff in computing, although that isn't really the key issue since humans are good at learning new words. What makes programming hard isn't mainly the syntax of languages, but the semantics: what things do, how they combine, the dynamics of their interactions, and how one uses the library functions of a language to achieve a goal.

A non-programmer would benefit from a reduction in the syntax of LSL for about 5 minutes ... and then find him/herself at a total loss because any kind of flexible programming always has pretty complex semantics, regardless of language. The object/state/event environment that is controlled through LSL is very flexible and complex, and it would be of no interest to anyone if it were not because the world would be terribly boring. That implies a minimum degree of semantic complexity in the language used, and you can't go below that.

This isn't theory: I've programmed in something like 30-40 different languages over time (languages is one of my research areas), and I've taught software engineering, computer science and programming (which are 3 very different disciplines) as a university lecturer. And it's pretty clear to me that no matter how simple the language, a very substantial proportion of people have difficulty grasping the semantic concepts of programming. They are non-programmers by mental predisposition, pretty much like I am a non-artist by mental predisposition.

And that is why LSL programmming is for programmers, not for non-programmers. Non-programmers already have the ability to "program" their objects in a simplified way that doesn't require them to be LSL script writers: by attaching ready-made scripts to their objects, and configuring them to do what they want as per instructions.

Will I get smacked if I said I wanted a LISP-like extension language?

I'd also like to see some kind of bytecode definition, so that those of us who like our own outlandish script styles can create compilers to it on our own machines.

I'd allow MUCH more access to the properties of objects (not just name, shape, etc -- but also allow objects to persistently store variables). I'd make every capability on land accessible via a hierarchal database structure. I'd make EVERYTHING on primitives available via script, also as a function of a hierarchal property tree.

I'd modify the behavior of primitive groups to easily see a list of all items in the group in edit mode, and query their properties interactively (I want, for example, to be able to query for all sittargets defined on members of the group). I doubt I'd want to allow SQL (unless it were all done on the client side, populated with data sent by the server when it's put into edit mode) or any other kind of 'standard' query language, but this is a low-level issue.

Permissions WOULD CHANGE. No more UNIX-like "owner/group/everyone" permissions bits -- and for God's sake, if I was the object creator, and I change the ownership of the object to someone else, WHY shouldn't I be able to retroactively grant modify and copy permissions to the new owner?

A much larger, and discrete, grid. Say I want to do something like the beginning of Piers Anthony's "Apprentice Adept" series, where there's two frames that exist in the same space but different quantum states -- I want to be able to create the 'curtain' and have it be possible to script the jump from one frame to the other.

I can't do this in SL. When I can do it, it'll be a marvel of engineering by whomever makes it possible.

The ability to create and edit objects locally, and then upload them to the server. The ability to see how textures will work on all primitives before I upload them. The ability to have real-time player-voice... for those projects built within the framework that need them. (I realize this is a point of massive contention, and probably part of why There went downhill, but there are reasons why it might be appropriate in limited circumstances.)

The ability to customize the interface. The ability to script the interface itself, instead of just scripting the server.

Oh well... all dreams.