op_cat may seem simple but will just lead to more complexity off chain. Ideally we just get covenants that actually express what we want to do offchain so we don’t need to jump through hoops or have complicated interactive protocols that require servers

If we can't get even string concatenation we won't get anything custom or specific. That being said my position on the best opcodes is evident here

GitHub

GitHub - Nuhvi/sake: Script Army Knife Emulator

Script Army Knife Emulator. Contribute to Nuhvi/sake development by creating an account on GitHub.

Yes anything you express through the script army knife (or MATT) is a fraud proof system, but I have concluded that ANY L2 solution is guaranteed to require that minimum livenees to challenge at least Data Availability attacks. The only solutions I am aware of for DA is to either challenge withdrawals and demand full state on Bitcoin which flat out doesn't scale, or the good old merge mining that so many hold their noses at even though it literally allows honest people to challenge DA in the same way they challenged private forks in Bitcoin. If I am wrong and you are aware of black magic that doesn't require people to be vigilant and challenge withdrawals within a time window or risk losing their money, let me know. As far as I am aware only L1 is passively secure and that comes at the expense of scalability because it demands that everyone downloads everything forever and never prune or compress anything or else risk a reorg. But I am excited to hear if I missed something

i was just saying cat by itself, we should just do GSR and fix script

Man you are a more optimistic person than I could dare to be :) Script Army Knife is way simpler than GSR and it does all I think we will need, and still I can't imagine it getting any support. If you can get GSR anytime in this decade I would definitely love that. But we won't... Hell the GSR people aren't even pretending they will have a finalized proposal anytime this decade!

GSR is the right way to do it engineering wise. I don’t know why people would be so hesitant for doing things the right way.

I am not hesitant. I only say OP_CAT because it is absurdly simple change to avoid arguments about complexity and security. I then say SAK because it does all the fraud proof things I need and it is actually finalised spec and code. GSR is just a vague promise at this point so I don't talk about it for that reason. I also think OP_CCV is a very good compact way to do what we would need to do in GSR with many opcodes. But whatever... Just give me something. Anything. Because until then we are stuck with BitVM.