[Bitcoin-ml] Can we decide on a roadmap now?

[Erik Beijnoff]

> On 03 Sep 2017, at 09:46, Nagai via bitcoin-ml <bitcoin-ml at lists.linuxfoundation.org> wrote:
> 
> Schnorr advantages:
> 
> Faster transaction validation
> Smaller transactions
> (^= smaller fees)
> CoinJoin support
> (^= more privacy)
> 

Although I am convinced that there are benefits to implementing Schnorr signatures, a change to the signing algorithm should not be taken lightly. Even if added as an optional signature type. Algorithms harden over time and proof of their usefulness and security is partly in their application. The default signature scheme of Bitcoin has been proving itself over time, both through usage in Bitcoin and before.

As a side note, the implementation of segwit is a perfect example of how these types of changes should *not* be done. In short, betting the house on the reliability of a specific scheme. In that specific case with obvious potential security holes.

If Schnorr signatures should be implemented in Bitcoin, either Legacy or Cash, it should probably be done over several years, gradually letting in the new type into the system, and on a very limited basis for the first year(s).

To get some of the benefits of Schnorr signatures (smaller transactions) I’d recommend looking into alternative schemes where the potential side effects are more easily evaluated. An example of this is, as you mentioned, FlexTrans by Tom Zander.