Talk:Second Life 2.0

Avatar is bigger than a normal human / SL 1.0 "meter" does not equal a real world meter

The standard avatar height does not match the size of average humans, in the measurement units used by the virtual world. The default avatar height is about 25% larger than a normal human. This may seems a small problem, but it affects the design of everything else in the virtual space.

Because all in-world objects have been built slightly larger than normal to compensate for the oversize avatar, this means the in-world measurement system cannot be used to accurately duplicate real-life objects; the resulting object is too small compared to the typical avatar. A house built using actual real-world measurements would be slightly too small compared to the avatar.

Due to this minor early design flaw of the avatar, the in-world definition of "meters" is meaningless. It would be better to stop calling them meters and use another SL-specific term (such as the "munge"), so that people don't confuse it with correct real-world measurements.

This isn't a bug or a design flaw. It's a social problem. People choose to have larger then average avatars which led to the average SL avatar being larger. Buildings were consequently made larger to compensate for peoples tastes in larger then average avatars. This isn't a technical problem. -- Strife Onizuka 12:26, 5 April 2008 (PDT)

Why would the people who started building in SL have chosen a "larger than normal" avatar? There is no reason for it. Space and distance means almost nothing in a virtual environment where nothing else exists as an initial reference point.

Assume you start in a sim with nothing but flat ground and an avatar. Set the avatar to the default male or female shape and begin building. If you follow the rules of normal meters, centimeters, etc, you soon discover that the avatar much bigger than the objects for some reason. It may not occur to people to shrink the avatar to fit, especially since it is the default size and shape as devised by Linden Lab. So instead they resize the objects to fit the avatar.

Regardless, this is a legacy problem that should never have been allowed to get this far since the measurement system is mostly useless for anything requiring real-world to virtual-world design precision.

Scalar Tardis 15:29, 5 April 2008 (PDT)

A few different responses I had to this.

• I don't dispute that it is annoying but it isn't a technical problem.
• The only "solution" to this is a procrustean solution. When it comes to procrustean solutions, they are only popular with those people who are holding the pruning shears.
• Why should a virtual world, where anything is possible, be bound by the limitations of the real world? When it comes to architecture, the design requirements are very different from those of the real world. The buildings can ignore gravity, security concerns & weather; not to mention there needs to be adequate space for the 3rd person camera to move around. You don't need bathrooms, closets, kitchens or even walls. In a world where the requirements and limitations of biology are nonexistent, it goes without saying that content that depends on those limitations will look out of place; they don't translate well.
• Depending upon the region of the world average human height can differer quiet dramatically. People will always use the avatar as a ruler but because differing preconceived notions about average height they will generate content to differing scales. There will always be content of differing scales.
• If you are to pick a new SL average/default size, how do we go about choose it? Regional average human height can vary as much as 15% of the total height. Considering such a large variance in just the averages it seem impossible to choose a single set that fits all the preconceived ideas about average height.  Human_height
  • Changing the default won't solve the problem, just shift it.
• The psychological aspect of this is something that could be very interesting to study.

-- Strife Onizuka 18:17, 5 April 2008 (PDT)

RL measurement terminology should be useful

Indeed a virtual space does not have any requirement to mirror the real world.

However, it was LL's idea to include the concept of meters into this virtual environment and that should therefore mean something in terms of measurement since a meter is so precisely defined in the SI measurement system.

The word should have some sort of meaningful relationship to avatar size or it should not be used in SL at all. This is why I say it should stop being called the meter because it isn't a meter but some randomly defined measure based on the average inworld avatar height and as such it doesn't correspond to anything else, sort of like like hogsheads and rods and pinches.

Scalar Tardis 00:24, 6 April 2008 (PDT)

How would you go about enforcing avatars and content of reasonable scales? If a group of avatars got together and formed a community they would have their own average height and subsequent content would be tailored for their avatars. Would you go in and tell them they were too tall or too short? That they were polluting the content pool? -- Strife Onizuka 18:51, 5 April 2008 (PDT)

The problem is NOT varying height. The problem is that the average height avatar is far taller than what could ever be considered average' in the real world and thus the world must be built to conform to the average height of these giants.

I am doing research on the beginnings of SL and this problem goes way back to the beginning, to the primitar: http://secondlife.wikia.com/wiki/Avatar_History

The primitar is said to have been accidently made a bit taller than the average human, which is why placing the sliders in the middle makes the avatar a bit taller than expected

Scalar Tardis 00:24, 6 April 2008 (PDT)

A few more thoughts:

• You have hit the nail on the head (or maybe your head on the door jam?). Have you noticed in SL how tall doorways are? It's because they have been built to fit giants. If you are trying to sell a house (in SL) if your clients don't fit through the doorways they won't buy it, they complain instead. In SL there is no way for avatars to duck or crawl though tiny doorways like people can in RL, (like when you go an visit a historic building built during the time when people were shorter). We see a similar trend in RL but with regard to the width of doorways, buildings are being designed to fit wheelchairs. I remember reading an article about how many funeral homes have had to buy wider crematoriums because the old equipment didn't fit the wider clients.
• I have edited that Wikia article, the context of your quote and the quote itself is critically flawed. The quoted text is hearsay ("[it] is said") and no source is provided for who is saying it; the sentence after the quoted text elevates that hearsay to the level of fact ("This fact may..."). Considering these flaws the quote cannot lend weight to your argument but I'm sure you know this and you cited it demonstrate that there is a line of inquiry that could result in information that supports your position.
• If we are considering changing the range of avatar sizes then we should take into consideration that there are people who want avatars that are outside the current range of avatar sizes. For example not only are there people who want Tiny avatars but those who want Micro avatars and at the other end of the spectrum there are people who want Macro avatars. Thus far they have had to squeak by with the clever application of attachments, animation overriders and invisiprims, which all give sub-par results.
• This is almost exactly what is happening to tailors in China and Indonesia, they export clothing for fat Americans. The clothing they are being asked to produce is outside the realm of their accepted range of normal human sizes. This is no different then builders in SL being asked to create things to fit larger avatars.

I had more to say by my browser crashed and eat my edit. ~_~ -- Strife Onizuka 03:54, 6 April 2008 (PDT)

Visual object size doesn't match physical size

Physical and visual sizes don't match because of the problems inherit with detecting collisions of small objects moving at high speeds. The math looks something like: Probability = (A.Size + B.Size + buffer_size) * FPS / |A.Velocity - B.Velocity|. If you halve the buffer_size, then to maintain the same Probability you must either half the max velocity, double the min size, or double the FPS. If you remove buffer_size altogether the changes required are much more drastic. To put things simply, you cannot simulate physics in real time perfectly, you have to cut corners. This is SVC-1139. -- Strife Onizuka 12:26, 5 April 2008 (PDT)

A better way to implement this limit is to state that any primitives smaller than a certain minimum will be treated as phantom and cannot be made physical. If the primitive is desired to be treated as if physical then it must be linked to a 100% transparent object larger than the minimum requirements, which will act as a placeholder for the primitives below the limit.

This would allow accurate visual/physical matching while allowing ultra detailed objects down to 0.0005 m with no physics engine impact.

Scalar Tardis 15:04, 7 April 2008 (PDT)

It's a nice idea but there is a flaw, the way I've described the math ignores the fact that we are dealing with something that is happening in three dimensions. The collision may not be happening straight on or along an axis. Example, take a box and rotate it 45 degrees on Z, we have an edge that is facing us, and then we take our gun and shoot just a little inside of the left or right outer edges, the probability of the bullet hitting the prim is low because there isn't much thickness the closer we aim to the inside of the edge. You need the buffer_size to make these edge conditions probable. Requiring a small prim to have a larger prim so that it can be physical would increase the sim cost for dealing with bullets and shell casings (rapid fire weapons historically have been able to cause sim lag, this would make the situation worse). I'm inclined to believe that the various solutions to the improbably collision detection problem have been weighed thoroughly by LL and they choose this course for the reason that it gave the best results. That said, I'll ask Andrew Linden about it tomorrow. -- Strife Onizuka 22:48, 7 April 2008 (PDT)

I asked at Andrew's office hour about it and I've got things all wrong. The reason for having the "collision tolerance" is to avoid penetration at almost all costs. -- Strife Onizuka 11:34, 8 April 2008 (PDT)

A better name for this page?

This article may now cause confusion with a very different real SL Viewer 2.0 on the way. I was going to move it but I am coming up with a blank on a better title. Any ideas? --Cerise Sorbet 18:04, 4 January 2010 (UTC)