Saturday, May 3, 2008

Pissing in the Well

Most people who talk about free market economics could not define a free market if required to, and don't want one. What they want is the right to piss in the well, and charge people for not pissing in the well. This is against the interests of a free market, where people are supposed to get money for making things better, not for not making things worse.

Consider a sim. A sim is a common pool of resources, and everyone shares them. LL's model of renting sim space does not charge for many kinds of shared resources, and this has created the incentive to find ways of pissing in the well of those shared resources. Camping and ad farms are two examples. What LL has admitted is that there are scarce resources that they are not charging for, like the load that avatars create on clients. Imposters go some of the way to reducing this load, but not very much, which is why they now have a tool to "rate" avatars. I think eventually this will be formalized, and avatars above a certain weight will be restricted from sims.

Ad cutting is one of the most prominent examples of this, with many people foolishly thinking that ad cutting represents a free market activity. In fact it is the reverse, ad cutters make their money not by selling advertising, but by griefing others into buying out parcels. The value of a 16M ad parcel is... determined by how much work others have done. This is a perverse incentive, that is, it gives an incentive not to improve land, but to grief people trying to build or improve land. The ad cutter charges to stop pissing in the well.

LL is starting to realize that while ad farmers pay tier, and LL likes tier to be paid, they are also generating enormous negative publicity, and angering people. Up until now LL has seen ad cutters as allies. Ad cutters grief people, and LL sells sims to people to avoid being griefed. However, markets are about information, and the information that LL supports and protects land griefing gets out, creating a market for competing platforms. These platforms, like LiteSim, OpenGrid, Twinity and so on are going to put downward pressure on LL's product and prices.

I've had it explained to me by business people that the computer industry deals with falling hardware prices by having "price points" and continually improving what is available at a price point. Ad cutters, by degrading LL's mainland product, then, are going to have to be dealt with to allow LL to maintain it's price point.

At the meeting today, we created several working groups to try and put forward solutions to the ad cutting problem. Most rely on persuading the monarchy that runs second life to act in the general interest of people who own land. It will be an uphill battle, even many people whose wells are being pissed in, want the right to piss in other people's wells in return. They just want a special exemption for their particular well. Then they complain when people use private property powers on them. This hypocrisy is deeply entrenched, and will take a great deal of work to over come.

Here are the working groups formed:

Zoning Working Group Lillie Yifu
Linden Land Klaatu Congrejo
Splitting Rem Nightfire

Now zoning is a property right, it is the right to control how people use shared resources for which everyone pays. Most people don't understand democracy or capitalism, and instead want feudalism. Feudalism is, however, labor intensive. It chews up a great deal of labor, and even successful land earls are under pressure in Second Life. If Anshe Chung Studios, which pays about 2/3 of what the rest of us pay in island fees, has a massive network effect, and body of content to leverage, is having problems keeping "Angels," then that is a clear sign of what should have been obvious: labor is less efficient than capital.

It is true that there will always be people who want to piss in the well, just for the lulz. It is true that there will always be people trying to piss in the well for money, note I don't say profit, because it isn't profit. However, there will also always be people who sneeze in your face, but that doesn't mean that it has to be accepted.

One of Second Life's big problems is the lack of democracy, because democracy forces people to make coalitions, help others with their problems, and get problems solved in return. Monarchy, everything is about pleading with the monarch.

Let me outline the proposal I have for zoning in Second Life.

It works on a sim by sim basis. In each sim every land owner votes for the minimum parcel that can be created or sold. To move up or down requires unanimity. So that means that the default would be whatever the first land owner sets it at, and to move that number up or down, all land owners would have to vote that way. A landowner cannot vote larger than their smallest parcel. Since voting is not tallied, it does not matter how many land owners there are.

Let me take an example, the first sim buyer wins the auction, and sets the default at 512M. that means that no one can cut a parcel smaller that 512M. He sells out, there are a whole new set of owners. One decides to start ad cutting, but all he can do is vote for the size to go down to 16M. If no one else does, 512 is still the minimum size. However, if one land owner wants to push it up to 1024, the same thing applies, only if everyone votes upwards will the size got to 1024.

This will make it so that people who want to ad cut, or just want a single 16M in a sim to rez a bot, can buy in sims whose size is 16M. People who are buying land will be able to see what the minumum size in a sim is, and decide if they want to buy accordingly. There will be no one size fits all. LL will be able to see if sizes bring people to bid higher, because they will auction land abandoned in sims.

What this will not do is force out land owners who have small parcels, they can set the size at whatever number they like, and other people cannot force them to change it. However, it does mean that once a sim is zoned, the owners are not going to wake up and find ad farms in their sim, because new owners can't force the number down. With the auction system, LL will not hold land long that is abandoned, and there will be the possibility of improving a sim, and therefore selling for more.

No more land griefing once a sim has established a common zone. Even more so, one can find out what the zones around a mainland sim are, and therefore core areas will go for more than side ones.

LL gets out of this improved mainland product, and the ability of people to bid higher on auctions. By giving land owners the right to protect common resources, rather than having them open to what is called the tragedy of the commons, where common things are over used, each one has an interest in setting the number high enough to protect their interests, and low enough to have parcels sell. People who improve a sim, rather than degrade it, get rewarded. New sims will have a mixture of starting sizes, and people can migrate to new areas that are more to their liking. Once a sim is cleared, it can stay cleared if the owners want it that way.

This would not be expensive to code, it is a few throttle checks in the server side on, a few dialog boxes, and a simple list to keep track of. It would also allow the market, rather than fiat, to set what size parcels are going to be, and allow people whose business is buying sims at auction and cutting them to provide value to their target buyers.

3 comments:

  1. The weighting tool is entirerly useless for calculating the actual lag effect of an Avatar on a sim server. At the moment the main load component on the server is scripts, particularly those with high resource hits such as sensors and prim movement. In fact it would be quite easy to create a unassuming piece of jewelry with rotating prims and multiple sensors in each prim that would only kick up the load tool reading by a few points but would lag the sim to hell and back. This situation will change with mono as scripts run a lot faster, but the principal remains.
    I have also heard techie Lindons say directly that campers do not cause major load on sims. I am uncertain as to the precise mechanism of the avatar load, but if you think about it it could easily be very minimal. The avatar stays in one place running a simple animation script, therefor all the sim needs to do is get an initial shape, content and animation of the avatar and send these to the client with sync info, thereafter keeping track of the avatar for changes.

    ReplyDelete
  2. With mono most of the script problem will go away. Now it will proliferate again,as people load back up on scripts, but the current level of script lag will drop precipitously.

    Campers cause much more lag than this, because they have to have a script that pays them, and it tends to pay them in small amounts frequently. Bots cause much less lag that campers because they tend to be stripped of almost all content, are not running scripts to keep themselves logged in, and don't have to be paid. Which is why bots are replacing campers for generating traffic.

    There are people who run hundreds of lightweight bots on a machine, and rent these out for less than camping rates.

    The weighting tool right now is highly imprecise and imperfect, but the Linden Lab people have a big incentive to create, and refine such a tool, since their future is in selling to companies that want a chance to advertise to people in Second Life. From the perspective of the company, and individual's avatar is a load on their presentation and their build, and they want to pack more ciphers into their build and presentation.

    So this is less about camping than attendance at events. A company doesn't want one hundred or even 300 people at an event, they want it to be much more scalable. One big problem with scalability is the attendees themselves.

    ReplyDelete
  3. Lillie, a script which pays a camper Lx every few minutes is pretty trivial in terms of resources. The biggies are things like sensors, which need to track objects all the time. But you are correct, mono should reduce server overheads for scripts considerably.

    I still think you have this backwards however. Client side (i.e the SL viewer program installed on your computer) rendering is not a common resource because it happens on the GPU of your laptop, and this is mainly what the rendering meter is measuring. I can easily negate much of the effect of this by upgrading my computer, in particular the graphics card. Restricting sim access based on the Rendering Cost meter is senseless because it's not measuring the correct criteria to make a decision over any resource the Lindons control - no matter how much you refine it.

    This is not to say that there would not be a place for a complementary meter that measured the server side avatar cost. It's not just corporates who want to scale, there's enough Fashion shows, Gigs and Clubs in the mainstream SL that could really benefit from some means of identifying avatars consuming unfair amounts of resources. You may consider you have a constitutional right to descend on a gig with 1000 prim jewelry whenever you want, but I'd prefer a door policy myself :-)

    Of course restricting entry by server-side load isn't the only way to go. One of the issues with SL atm is that all sims are effectively equal. If it were possible for a corporate running and event, or manager hosting a gig, to buy or rent a more powerful sim (and standard sims atm are amazingly low-powered machines) which could support all the fancy jewelry that anyone wants then this would be an interesting approach. The advent of IBMs 'private' extension to SL possibly point a way to this, as does discussion about interoperability of OpenSim grids and the SL grid.

    ReplyDelete