06:43 <CjS77> Sounds like great fodder for a blog series, andytoshi. Teaching about how Bitcoin really works by talking about all the things that you would have changed if you were doing a complete rewrite :)

06:48 <CjS77> Specifically, on the malleability of the input script witness data. I know segwit pins the txid, but there's still the other wtxid that's still technically malleable. In an ideal world, I presume you'd want to prevent that too, and segwit is a compromise that somewhat solves the problem while maintaining backward compatibility?

06:57 <CjS77> Though practically, would there be a simple way to sign / commit to the unlocking script and maintain the flexibility that bitcoin script offers?

16:26 <andytoshi> CjS77: well there's no way in principle to prevent wtxid malleability

16:26 <andytoshi> because transactions will always potentially have multiple witnesses

16:27 <andytoshi> but there is a lot of encoding stuff that could've been canonical or at least fixed-length. DER signatures are another thing i'd change that i forgot to mention above