Hi ZmnSCPxj,

> Just ask a bunch of fullnodes to add this 1Mb of extra ignored data in this tiny 1-input-1-output transaction so I pay only a small fee

I'm not suggesting that you wouldn't have to pay a fee for it. You'd pay a fee for it as normal, so there's no DOS vector. Doesn't adding extra witness data do what would be needed here? Eg simply adding extra data onto the witness script that will remain unconsumed after successful execution of the script?

> how do new jets get introduced?

In scenario A, new jets get introduced by being added to bitcoin software as basically relay rules.

> If a new jet requires coordinated deployment over the network, then you might as well just softfork and be done with it.

It would not need a coordinated deployment. However, the more nodes that supported that jet, the more efficient using it would be for the network.

> If a new jet can just be entered into some configuration file, how do you coordinate those between multiple users so that there *is* some benefit for relay?

When a new version of bitcoin comes out, people generally upgrade to it eventually. No coordination is needed. 100% of the network need not support a jet. Just some critical mass to get some benefit.

> Having a static lookup table is better since you can pattern-match on strings of specific, static length

Sorry, better than what exactly?

> How does the unupgraded-to-upgraded boundary work?

This is what I'm thinking. Imagine a simple script:

OP_DUP

OP_ADD

with witness

This would execute as 1+1 = 2 -> success. Let's say the script is jettified so we can instead write it as:

OP_JET

1b5f03cf # adler32 hash of the replaced script

with a witness:

OP_JET # Some number that represents OP_JET

1b5f03cf

A jet-aware node transmitting to another jet-aware node can transmit it as is (tho it would need to do a swap out to validate). For a jet-aware node to transmit this to a non-jet aware node, it would need to swap the OP_JET call with the script it represents. So the transaction sent to the non-jet aware node would have:

Script:

OP_DUP

OP_ADD

Witness:

OP_JET

1b5f03cf

And you can see that this would execute and succeed by adding 1+1 and ending up with the stack:

2

1b5f03cf

OP_JET

Which would succeed because of the non-zero top of stack.

When the non-jet aware node sends this to a jet-aware node, that node would see the extra items on the stack after script execution, and would interpret them as an OP_JET call specifying that OP_JET should replace the witness items starting at index 0 with `1b5f03cf OP_JET`. It does this and then sends that along to the next hop.

In order to support this without a soft fork, this extra otherwise unnecessary data would be needed, but for jets that represent long scripts, the extra witness data could be well worth it (for the network).

However, this extra data would be a disincentive to do transactions this way, even when its better for the network. So it might not be worth doing it this way without a soft fork. But with a soft fork to upgrade nodes to support an OP_JET opcode, the extra witness data can be removed (replaced with out-of-band script fragment transmission for nodes that don't support a particular jet).

One interesting additional thing that could be done with this mechanism is to add higher-order function ability to jets, which could allow nodes to add OP_FOLD or similar functions as a jet without requiring additional soft forks. Hypothetically, you could imagine a jet script that uses an OP_LOOP jet be written as follows:

5 # Loop 5 times

1 # Loop the next 1 operation

3c1g14ad

OP_JET

OP_ADD # The 1 operation to loop

The above would sum up 5 numbers from the stack. And while this summation jet can't be represented in bitcoin script on its own (since bitcoin script can't manipulate opcode calls), the jet *call* can still be represented as:

OP_ADD

which means all of the above replacement functionality would work just as well.

So my point here is that jets implemented in a way similar to this would give a much wider range of "code as compression" possibilities than implementing a single opcode like op_fold.

> To make jets more useful, we should redesign the language so that `OP_PUSH` is not in the opcode stream, but instead, we have a separate table of constants that is attached / concatenated to the actual SCRIPT.

This can already be done, right? You just have to redesign the script to consume and swap/rot around the data in the right way to separate them out from the main script body.

On Mon, Mar 7, 2022 at 5:35 PM ZmnSCPxj <ZmnSCPxj@protonmail.com> wrote:

Good morning Billy,

Changed subject since this is only tangentially related to `OP_FOLD`.

> Let me organize my thoughts on this a little more clearly. There's a couple possibilities I can think of for a jet-like system:
>
> A. We could implement jets now without a consensus change, and without requiring all nodes to upgrade to new relay rules. Probably. This would give upgraded nodes improved validation performance and many upgraded nodes relay savings (transmitting/receiving fewer bytes). Transactions would be weighted the same as without the use of jets tho.
> B. We could implement the above + lighter weighting by using a soft fork to put the jets in a part of the blockchain hidden from unupgraded nodes, as you mentioned.
> C. We could implement the above + the jet registration idea in a soft fork.
>
> For A:
>
> * Upgraded nodes query each connection for support of jets in general, and which specific jets they support.
> * For a connection to another upgraded node that supports the jet(s) that a transaction contains, the transaction is sent verbatim with the jet included in the script (eg as some fake opcode line like 23 OP_JET, indicating to insert standard jet 23 in its place). When validation happens, or when a miner includes it in a block, the jet opcode call is replaced with the script it represents so hashing happens in a way that is recognizable to unupgraded nodes.
> * For a connection to a non-upgraded node that doesn't support jets, or an upgraded node that doesn't support the particular jet included in the script, the jet opcode call is replaced as above before sending to that node. In addition, some data is added to the transaction that unupgraded nodes propagate along but otherwise ignore. Maybe this is extra witness data, maybe this is some kind of "annex", or something else. But that data would contain the original jet opcode (in this example "23 OP_JET") so that when that transaction data reaches an upgraded node that recognizes that jet again, it can swap that back in, in place of the script fragment it represents.
>
> I'm not 100% sure the required mechanism I mentioned of "extra ignored data" exists, and if it doesn't, then all nodes would at least need to be upgraded to support that before this mechanism could fully work.

I am not sure that can even be *made* to exist.
It seems to me a trivial way to launch a DDoS: Just ask a bunch of fullnodes to add this 1Mb of extra ignored data in this tiny 1-input-1-output transaction so I pay only a small fee if it confirms but the bandwidth of all fullnodes is wasted transmitting and then ignoring this block of data.

> But even if such a mechanism doesn't exist, a jet script could still be used, but it would be clobbered by the first nonupgraded node it is relayed to, and can't then be converted back (without using a potentially expensive lookup table as you mentioned).

Yes, and people still run Bitcoin Core 0.8.x.....

> > If the script does not weigh less if it uses a jet, then there is no incentive for end-users to use a jet
>
> That's a good point. However, I'd point out that nodes do lots of things that there's no individual incentive for, and this might be one where people either altruistically use jets to be lighter on the network, or use them in the hopes that the jet is accepted as a standard, reducing the cost of their scripts. But certainly a direct incentive to use them is better. Honest nodes can favor connecting to those that support jets.

Since you do not want a dynamic lookup table (because of the cost of lookup), how do new jets get introduced?
If a new jet requires coordinated deployment over the network, then you might as well just softfork and be done with it.
If a new jet can just be entered into some configuration file, how do you coordinate those between multiple users so that there *is* some benefit for relay?

> >if a jet would allow SCRIPT weights to decrease, upgraded nodes need to hide them from unupgraded nodes
> > we have to do that by telling unupgraded nodes "this script will always succeed and has weight 0"
>
> Right. It doesn't have to be weight zero, but that would work fine enough.
>
> > if everybody else has not upgraded, a user of a new jet has no security.
>
> For case A, no security is lost. For case B you're right. For case C, once nodes upgrade to the initial soft fork, new registered jets can take advantage of relay-cost weight savings (defined by the soft fork) without requiring any nodes to do any upgrading, and nodes could be further upgraded to optimize the validation of various of those registered jets, but those processing savings couldn't change the weighting of transactions without an additional soft fork.
>
> > Consider an attack where I feed you a SCRIPT that validates trivially but is filled with almost-but-not-quite-jettable code
>
> I agree a pattern-matching lookup table is probably not a great design. But a lookup table like that is not needed for the jet registration idea. After the necessary soft fork, there would be standard rules for which registered jets nodes are required to keep an index of, and so the lookup table would be a straightforward jet hash lookup rather than a pattern-matching lookup, which wouldn't have the same DOS problems. A node would simply find a jet opcode call like "ab38cd39e OP_JET" and just lookup ab38cd39e in its index.

How does the unupgraded-to-upgraded boundary work?
Having a static lookup table is better since you can pattern-match on strings of specific, static length, and we can take a page from `rsync` and use its "rolling checksum" idea which works with identifying strings of a certain specific length at arbitrary offsets.

Say you have jetted sequences where the original code is 42 bytes, and another jetted sequence where the original code is 54 bytes, you would keep a 42-byte rolling checksum and a separate 54-byte rolling checksum, and then when it matches, you check if the last 42 or 54 bytes matched the jetted sequences.

It does imply having a bunch of rolling checksums around, though.
Sigh.

---

To make jets more useful, we should redesign the language so that `OP_PUSH` is not in the opcode stream, but instead, we have a separate table of constants that is attached / concatenated to the actual SCRIPT.

So for example instead of an HTLC having embedded `OP_PUSH`es like this:

OP_IF
OP_HASH160 <hash> OP_EQUALVERIFY OP_DUP OP_HASH160 <acceptor pkh>
OP_ELSE
<timeout> OP_CHECKLOCKTIMEVERIFY OP_DROP OP_DUP OP_HASH160 <offerrer pkh>
OP_ENDIF
OP_EQUALVERIFY
OP_CHECKSIG

We would have:

constants:
h = <hash>
a = <acceptor pkh>
t = <timeout>
o = <offerer pkh>
script:
OP_IF
OP_HASH160 h OP_EQUALVERIFY OP_DUP OP_HASH160 a
OP_ELSE
t OP_CHECKLOCKTIMEVERIFY OP_DROP OP_DUP OP_HASH160 o
OP_ENDIF
OP_EQUALVERIFY
OP_CHECKSIG

The above allows for more compressibility, as the entire `script` portion can be recognized as a jet outright.
Move the incompressible hashes out of the main SCRIPT body.

We should note as well that this makes it *easier* to create recursive covenants (for good or ill) out of `OP_CAT` and whatever opcode you want that allows recursive covenants in combination with `OP_CAT`.
Generally, recursive covenants are *much* more interesting if they can change some variables at each iteration, and having a separate table-of-constants greatly facilitates that.

Indeed, the exercise of `OP_TLUV` in [drivechains-over-recursive-convenants][] puts the loop variables into the front of the SCRIPT to make it easier to work with the SCRIPT manipulation.

[drivechains-over-recursive-covenants]: https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2022-February/019976.html

---

Perhaps we can consider the general vs specific tension in information-theoretic terms.

A language which supports more computational power --- i.e. more general --- must, by necessity, have longer symbols, as a basic law of information theory.
After all, a general language can express more things.

However, we do recognize that certain sequences of things-to-say are much more likely than others.
That is, we expect that certain sequences "make sense" to do.
That is why "jets" are even proposed, they are shortcuts towards those.

Assuming a general language is already deployed for Bitcoin, then a new opcode is a jet as it simply makes the SCRIPT shorter.

Instead of starting with a verbose (by necessity) general language, we could instead start with a terse but restricted language, and slowly loosen up its restrictions by adding new capabilities in softforks.

Regards,
ZmnSCPxj