[bitcoin-dev] One testnet to rule them all

[bitcoin-dev] One testnet to rule them all

[vjudeu]

In testnet3, anyone can become a miner, it is possible to even mine a block on some CPU, because the difficulty can drop to one. In signet, we create some challenge, for example 1-of-2 multisig, that can restrict who can mine, so that chain can be "unreliably reliable". Then, my question is: why signets are introducing new coins out of thin air, instead of forming two-way peg-in between testnet3 and signet?

The lack of coins is not a bug, it is a feature. We have more halvings in testnet3 than in mainnet or signets, but it can be good, we can use this to see, what can happen with a chain after many halvings. Also, in testnet3 there is no need to have any coins if we are mining. Miners can create, move and destroy zero satoshis. They can also extend the precision of the coins, so a single coin in testnet3 can be represented as a thousand of coins in some signet sidechain.

Recently, there are some discussions regarding sidechains. Before they will become a real thing, running on mainnet, they should be tested. Nowadays, a popular way of testing new features is creating a new signet with new rules. But the question still remains: why we need new coins, created out of thin air? And even when some signet wants to do that, then why it is not pegged into testnet3? Then it would have as much chainwork protection as testnet3!

It seems that testnet3 is good enough to represent the main chain during sidechain testing. It is permissionless and open, anyone can start mining sidechain blocks, anyone with a CPU can be lucky and find a block with the minimal difficulty. Also, because of blockstorms and regular chain reorgs, some extreme scenarios, like stealing all coins from some sidechain, can be tested in a public way, because that "unfriendly and unstable" environment can be used to test stronger attacks than in a typical chain.

Putting that proposal into practice can be simple and require just creating one Taproot address per signet in testnet3. Then, it is possible to create one testnet transaction (every three months) that would move coins to and from testnet3, so the same coins could travel between many signets. New signets can be pegged in with 1:1 ratio, existing signets can be transformed into signet sidechains (the signet miners rule that chains, so they can enforce any transition rules they need).

Re: [bitcoin-dev] One testnet to rule them all

[Jeremy Rubin]

[-- Attachment #1: Type: text/plain, Size: 3167 bytes --]

There's no point to pegging coins that are worthless into a system of also
worthless coins, unless you want to test the mechanism of testing pegging.

As is, it's hard enough to get people set up on a signet, if they have to
run two nodes and then scramble to find testnet coins and then peg them
were just raising the barriers to entry for starting to use a signet for
testing.


If anything I think we should permanently shutter testnet now that signet
is available.

On Sat, Mar 5, 2022, 3:53 PM vjudeu via bitcoin-dev <
bitcoin-dev@lists.linuxfoundation.org> wrote:

> In testnet3, anyone can become a miner, it is possible to even mine a
> block on some CPU, because the difficulty can drop to one. In signet, we
> create some challenge, for example 1-of-2 multisig, that can restrict who
> can mine, so that chain can be "unreliably reliable". Then, my question is:
> why signets are introducing new coins out of thin air, instead of forming
> two-way peg-in between testnet3 and signet?
>
> The lack of coins is not a bug, it is a feature. We have more halvings in
> testnet3 than in mainnet or signets, but it can be good, we can use this to
> see, what can happen with a chain after many halvings. Also, in testnet3
> there is no need to have any coins if we are mining. Miners can create,
> move and destroy zero satoshis. They can also extend the precision of the
> coins, so a single coin in testnet3 can be represented as a thousand of
> coins in some signet sidechain.
>
> Recently, there are some discussions regarding sidechains. Before they
> will become a real thing, running on mainnet, they should be tested.
> Nowadays, a popular way of testing new features is creating a new signet
> with new rules. But the question still remains: why we need new coins,
> created out of thin air? And even when some signet wants to do that, then
> why it is not pegged into testnet3? Then it would have as much chainwork
> protection as testnet3!
>
> It seems that testnet3 is good enough to represent the main chain during
> sidechain testing. It is permissionless and open, anyone can start mining
> sidechain blocks, anyone with a CPU can be lucky and find a block with the
> minimal difficulty. Also, because of blockstorms and regular chain reorgs,
> some extreme scenarios, like stealing all coins from some sidechain, can be
> tested in a public way, because that "unfriendly and unstable" environment
> can be used to test stronger attacks than in a typical chain.
>
> Putting that proposal into practice can be simple and require just
> creating one Taproot address per signet in testnet3. Then, it is possible
> to create one testnet transaction (every three months) that would move
> coins to and from testnet3, so the same coins could travel between many
> signets. New signets can be pegged in with 1:1 ratio, existing signets can
> be transformed into signet sidechains (the signet miners rule that chains,
> so they can enforce any transition rules they need).
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>

[-- Attachment #2: Type: text/html, Size: 3786 bytes --]

Re: [bitcoin-dev] One testnet to rule them all

[vjudeu]

> There's no point to pegging coins that are worthless into a system of also worthless coins, unless you want to test the mechanism of testing pegging.

But testing pegging is what is needed if we ever want to introduce sidechains. On the other hand, even if we don't want sidechains, then the question still remains: why we need more than 21 million coins for testing, if we don't need more than 21 million coins for real transactions?

> If anything I think we should permanently shutter testnet now that signet is available.

Then, in that case, the "mainchain" can be our official signet and other signets can be pegged into that. Also, testnet3 is permissionless, so how signet can replace that? Because if you want to test mining and you cannot mine any blocks in signet, then it is another problem.

On 2022-03-05 17:19:40 user Jeremy Rubin <jeremy.l.rubin@gmail.com> wrote:
There's no point to pegging coins that are worthless into a system of also worthless coins, unless you want to test the mechanism of testing pegging.


As is, it's hard enough to get people set up on a signet, if they have to run two nodes and then scramble to find testnet coins and then peg them were just raising the barriers to entry for starting to use a signet for testing.




If anything I think we should permanently shutter testnet now that signet is available.


On Sat, Mar 5, 2022, 3:53 PM vjudeu via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
In testnet3, anyone can become a miner, it is possible to even mine a block on some CPU, because the difficulty can drop to one. In signet, we create some challenge, for example 1-of-2 multisig, that can restrict who can mine, so that chain can be "unreliably reliable". Then, my question is: why signets are introducing new coins out of thin air, instead of forming two-way peg-in between testnet3 and signet?

The lack of coins is not a bug, it is a feature. We have more halvings in testnet3 than in mainnet or signets, but it can be good, we can use this to see, what can happen with a chain after many halvings. Also, in testnet3 there is no need to have any coins if we are mining. Miners can create, move and destroy zero satoshis. They can also extend the precision of the coins, so a single coin in testnet3 can be represented as a thousand of coins in some signet sidechain.

Recently, there are some discussions regarding sidechains. Before they will become a real thing, running on mainnet, they should be tested. Nowadays, a popular way of testing new features is creating a new signet with new rules. But the question still remains: why we need new coins, created out of thin air? And even when some signet wants to do that, then why it is not pegged into testnet3? Then it would have as much chainwork protection as testnet3!

It seems that testnet3 is good enough to represent the main chain during sidechain testing. It is permissionless and open, anyone can start mining sidechain blocks, anyone with a CPU can be lucky and find a block with the minimal difficulty. Also, because of blockstorms and regular chain reorgs, some extreme scenarios, like stealing all coins from some sidechain, can be tested in a public way, because that "unfriendly and unstable" environment can be used to test stronger attacks than in a typical chain.

Putting that proposal into practice can be simple and require just creating one Taproot address per signet in testnet3. Then, it is possible to create one testnet transaction (every three months) that would move coins to and from testnet3, so the same coins could travel between many signets. New signets can be pegged in with 1:1 ratio, existing signets can be transformed into signet sidechains (the signet miners rule that chains, so they can enforce any transition rules they need).
_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Re: [bitcoin-dev] One testnet to rule them all

[Jeremy Rubin]

[-- Attachment #1: Type: text/plain, Size: 4371 bytes --]

Signet degrades to a testnet if you make your key OP_TRUE.


It's not about needing 21M coins it's about easily getting access to said
coins for testing, where it's kinda tricky to get testnet coins.

On Sat, Mar 5, 2022, 6:17 PM <vjudeu@gazeta.pl> wrote:

> > There's no point to pegging coins that are worthless into a system of
> also worthless coins, unless you want to test the mechanism of testing
> pegging.
>
> But testing pegging is what is needed if we ever want to introduce
> sidechains. On the other hand, even if we don't want sidechains, then the
> question still remains: why we need more than 21 million coins for testing,
> if we don't need more than 21 million coins for real transactions?
>
> > If anything I think we should permanently shutter testnet now that
> signet is available.
>
> Then, in that case, the "mainchain" can be our official signet and other
> signets can be pegged into that. Also, testnet3 is permissionless, so how
> signet can replace that? Because if you want to test mining and you cannot
> mine any blocks in signet, then it is another problem.
>
> On 2022-03-05 17:19:40 user Jeremy Rubin <jeremy.l.rubin@gmail.com> wrote:
> There's no point to pegging coins that are worthless into a system of also
> worthless coins, unless you want to test the mechanism of testing pegging.
>
>
> As is, it's hard enough to get people set up on a signet, if they have to
> run two nodes and then scramble to find testnet coins and then peg them
> were just raising the barriers to entry for starting to use a signet for
> testing.
>
>
>
>
> If anything I think we should permanently shutter testnet now that signet
> is available.
>
>
> On Sat, Mar 5, 2022, 3:53 PM vjudeu via bitcoin-dev <
> bitcoin-dev@lists.linuxfoundation.org> wrote:
> In testnet3, anyone can become a miner, it is possible to even mine a
> block on some CPU, because the difficulty can drop to one. In signet, we
> create some challenge, for example 1-of-2 multisig, that can restrict who
> can mine, so that chain can be "unreliably reliable". Then, my question is:
> why signets are introducing new coins out of thin air, instead of forming
> two-way peg-in between testnet3 and signet?
>
> The lack of coins is not a bug, it is a feature. We have more halvings in
> testnet3 than in mainnet or signets, but it can be good, we can use this to
> see, what can happen with a chain after many halvings. Also, in testnet3
> there is no need to have any coins if we are mining. Miners can create,
> move and destroy zero satoshis. They can also extend the precision of the
> coins, so a single coin in testnet3 can be represented as a thousand of
> coins in some signet sidechain.
>
> Recently, there are some discussions regarding sidechains. Before they
> will become a real thing, running on mainnet, they should be tested.
> Nowadays, a popular way of testing new features is creating a new signet
> with new rules. But the question still remains: why we need new coins,
> created out of thin air? And even when some signet wants to do that, then
> why it is not pegged into testnet3? Then it would have as much chainwork
> protection as testnet3!
>
> It seems that testnet3 is good enough to represent the main chain during
> sidechain testing. It is permissionless and open, anyone can start mining
> sidechain blocks, anyone with a CPU can be lucky and find a block with the
> minimal difficulty. Also, because of blockstorms and regular chain reorgs,
> some extreme scenarios, like stealing all coins from some sidechain, can be
> tested in a public way, because that "unfriendly and unstable" environment
> can be used to test stronger attacks than in a typical chain.
>
> Putting that proposal into practice can be simple and require just
> creating one Taproot address per signet in testnet3. Then, it is possible
> to create one testnet transaction (every three months) that would move
> coins to and from testnet3, so the same coins could travel between many
> signets. New signets can be pegged in with 1:1 ratio, existing signets can
> be transformed into signet sidechains (the signet miners rule that chains,
> so they can enforce any transition rules they need).
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>
>

[-- Attachment #2: Type: text/html, Size: 5200 bytes --]