Todd Fleming

I like that idea

That, together with write conflict checking, could also be a solution to read-write sync calls.

Scratch read-write; I see it causing a different kind of confusion. It's too far removed from what people tend to think of with synchronous models; e.g. write X then read...

Hmmm. Contract A write X then read X through a read-only function would also violate typical synchronous models.

If we place that burden on them, then we could avoid changing database semantics during read-only calls.

I see an advantage to keeping an Alternatives section. It provides a handy lookup. Someone is using them, they hear we're deprecating them, so they look this up and (hopefully)...

This also isn't the proposal to remove. It's a deprecation notice.

> unless we produce a version of the code that cannot replay from genesis This is a possibility. It'd be nice to eventually yank that code out.

> Trying to micromanage pair-wise interactions where we have created a false dependency to disabling deferred transactions is still a lot of effort. A possible approach is for new features...

They do deserve separate proposals. That doesn't prevent the short descriptions from being in the alternatives section.