> I don't see any benefit to
changing that. It is better that coins are burned.

I think this is our fundamental disagreement. People will burn coins to encode data, why allow this when there's a better alternative?

> You *always* need a key, to redeem inputs... regardless of values.

Correct, but with BIP70 that key is in the user's wallet and you can construct transactions on another machine (thus not needing a key during construction). Right now there's no way to do the transaction construction on another machine with zero value OP_RETURNs.

On Mon, Jan 25, 2016 at 7:04 PM, Luke Dashjr <luke@dashjr.org> wrote:
On Tuesday, January 26, 2016 3:01:13 AM Toby Padilla wrote:
> > As I explained, none of those reasons apply to PaymentRequests.
>
> As they exist today PaymentRequests allow for essentially the same types of
> transactions as non-PaymentRequest based transactions with the limitation
> that OP_RETURN values must be greater. In that sense they're basically a
> pre-OP_RETURN environment. OP_RETURN serves a purpose and it can't be used
> with PaymentRequest transactions.

OP_RETURN can be used, but you need to burn coins. I don't see any benefit to
changing that. It is better that coins are burned.

> > I have no idea what you are trying to say here.
>
> I think if you think through how you would create an OP_RETURN transaction
> today without this BIP you'll see you need a key at some point if you want
> a zero value.

You *always* need a key, to redeem inputs... regardless of values.

Luke