Generation of namecoins vs. demand for namecoins

[gst]

I think the current way how namecoins are generated is not really optimal, as the amount of generated coins does not directly depend on the demand, but only on the block number. So if the demand for coins (or domain names) is higher then what is generated by the network, domains will be too costly, hindering adoption of the .bit zone. On the other hand, if there are too many coins that are generated, domains will be very cheap, leading to a large amount of domains being registered by squatters.

It should be possible to solve this problem by making the amount of generated coins dependent on the demand. A good metric for the demand might be the total hashing power of the network.

For example:

• Charge a fixed cost to register a domain
• Make the amount of namecoins generated per block proportional to the block's difficulty (which is indirectly given by the total hashing power in the network)

So if the demand for namecoins is higher than what is currently generated, miners will be able to make quite some money by selling namecoins. Thus, new miners will be motivated to join the network, increasing the difficulty and thereby the amount of namecoins that is generated.

On the other hand, if the demand for namecoins is only low, miners won't be able to make much money by selling them. Thus miners will leave the network, until the hashing power corresponds to the demand.

In addition, instead of having a fixed cost per domain name, it might be a good idea to set the domain's cost based on it's length. So we would, e.g., have a standard cost of 50 coins for domains of, e.g., length 6 or above. A cost of 100 coins for domains of length 5, 200 coins for length 4, 400 coins for length 3, 800 coins for length 2 and 1600 coins for length 1.

[drknark]

For example:

Charge a fixed cost to register a domain
Make the amount of namecoins generated per block proportional to the block's difficulty (which is indirectly given by the total hashing power in the network)

Yes, the amount of generated coins is more or less static (halved every 200k blocks IIRC). But the cost of addresses is reduced every X blocks, where X << 200k, I don't recall what it is. Reducing costs with static generation or static costs with increased generation is the same thing, though your idea depends on difficulty in a way that the current system doesn't.

Regardless, I think that something like this would be hard to implement well, and it would definitely be a big step away from the bitcoin idea of limited supply. Because of this, I don't think we will see something like this implemented.

So if the demand for coins (or domain names) is higher then what is generated by the network, domains will be too costly, hindering adoption of the .bit zone. On the other hand, if there are too many coins that are generated, domains will be very cheap, leading to a large amount of domains being registered by squatters.

It is possible (I think) that domains might grow to be too expensive. But if domains are too expensive, demand should decrease thereby decreasing the price. This link might not work the way we want it to, domains might get too expensive for anyone but corporations to buy.

Being too cheap however should be adjusted by the free market, cheap domains -> more registrations. If squatting becomes a problem, the squatters will themselves increase the cost of domains by increasing demand.

So if the demand for namecoins is higher than what is currently generated, miners will be able to make quite some money by selling namecoins. Thus, new miners will be motivated to join the network, increasing the difficulty and thereby the amount of namecoins that is generated.

As for getting more miners, the miners will make the same amount of money whether they get many coins with a low price, or few coins with a high price.

Disclaimer: I am not an economist

[misterbigg]

I think the current way how namecoins are generated is not really optimal, as the amount of generated coins does not directly depend on the demand, but only on the block number.

I agree with you, and I noticed this myself.

So if the demand for coins (or domain names) is higher then what is generated by the network, domains will be too costly, hindering adoption of the .bit zone. On the other hand, if there are too many coins that are generated, domains will be very cheap, leading to a large amount of domains being registered by squatters.

Also agree.

It should be possible to solve this problem by making the amount of generated coins dependent on the demand.

Yes, and I proposed something like this here:

http://dot-bit.org/forum/viewtopic.php?f=2&t=204

* Charge a fixed cost to register a domain

Good idea, I suggested this.

* Make the amount of namecoins generated per block proportional to the block's difficulty

I proposed generating namecoins in proportion to the number of registered domains times some constant (lets say, 5 or 10).

it might be a good idea to set the domain's cost based on it's length.

WONDERFUL IDEA!!!

[misterbigg]

Yes, the amount of generated coins is more or less static (halved every 200k blocks IIRC). But the cost of addresses is reduced every X blocks, where X << 200k, I don't recall what it is. Reducing costs with static generation or static costs with increased generation is the same thing, though your idea depends on difficulty in a way that the current system doesn't.

I see this as a huge weakness in the Namecoin system. There is no way to develop a static system of price adjustments. It has to be market and demand based. Thats why I proposed something to address this, that involves expiring namecoins and namecoin auctions. This provides the necessary market feedback to get namecoins priced "right:"

http://dot-bit.org/forum/viewtopic.php?f=2&t=204

[phelix]

we were also talking about the issue of names being to cheap here: http://nf.bit/viewtopic.php?f=2&t=360
and I think the latest discussion about fees: http://dot-bit.org/forum/viewtopic.php?p=2823#p2823

something has to be done: name_update is totally free

As a quick start I would set a hard limit for the minimum tx fee to 0.1 or at least 0.01

edit:
what about a network enforced minimum tx fee of 0.01 and default setting in clients of 0.1?
The default setting gives pool operators the possibility to actually drop free tx. The complex fee calculation makes this too cumbersome. Maybe the network enforced fee is not even necessary and the default setting with a change in fee calculation (hard limit) is enough.

[phelix]

I see issues in your proposal:
* it is complex. The namecoin development team is... small. I really don't see something like that coming.
* absolute difficulty is not a good indicator for demand.
With asics coming things might change by a factor of 100.

What you could do is control namecoin difficulty to a percentage of bitcoin difficulty, say 80%. Besides also being very complex it is still not enough. In the future difficulty will be driven by tx/name_op fees and tx/name_op volume. If you enforce fees to control difficulty you don't know how it will influence volume.

[virtual_master]

Any change in the Namecoin generation process would create a distrust among the users. It would be like changing the total available Bitcoin amount.
What we could regulate with the same effect the registration/renewal price which is 1 USD cent/domain at the moment and it is to low.
If the registration/renewal fee would be dependent of the length this could act also as a regulator without changing every time the fee when the namecoin prices are changing.
When the namecoin price is very low like now everybody could afford to spend (or invest as locked coin) a very short domain which would cost multiple of the basic fee.
When namecoin prices go 20 times higher everybody could afford a long domain for private purpose. For organizations or companies it doesn't matter anyway.