The Future of Pocket

I just chanced upon a little conversation in the comments of @pocket-pi's latest Pocket statistics post (If you want to support Pocket, you should check it out and upvote!).

@cryptick left this comment:

saying "Could we send [Pocket] as a type of second curation reward?"

This led me to ponder if it would be appropriate to update the POCKET protocol to do more things. I'm writing this post to talk about what kinds up updates would be ok, as well as a few things that I think we need to be certain to avoid.

What is the Pocket Protocol?

First, a tiny bit of background. The Pocket protocol is a document, a set of rules, that determine what kinds of Steem operations have what kind of effect on the Pocket system. When you submit a pocketsend command in a comment, it's the Pocket protocol that decides whether your command is valid or not. The bots you see confirming Pocket transactions (mostly @pocket-pi these days) are running Python code that I wrote (you can run it too!) that checks each Steem operation to see whether it matches the Pocket protocol and is thus a valid Pocket operation. If you want the high-level picture, read my meta-subchain post.

When I first wrote the Pocket protocol, I imagined that it would never change. I'd been involved in cryptocurrency projects where the people in charge were overly-flippant about their authority to change the rules that governed the systems, and there were many times that they decided to change things that alienated a lot of people. So when I originally wrote the protocol, my thought was that it would never change. Maybe people would come along and decide they could do better and copy it and do something else with it, but POCKET would always be POCKET.

What would be an appropriate upgrade for Pocket?

So, let's just suppose for an instant that we wanted to make some change to the Pocket protocol. What kind of change would be acceptable?

For me to support a modification of the Pocket protocol, I would have to be fairly convinced that (sorry, gotta go economist-speak for a bit) the change would Pareto-improve Pocket. This means it wouldn't make anybody worse off, and some people would for sure be better off.

An Example

Here is an example of a very simple Pareto-improving modification: give people the ability to transfer some of their Pocket into a "savings" account that is controlled by their Steem account's active key. One way to implement this would be to introduce a command that is executed by transferring SBD/STEEM to yourself, with a memo that says pocketvault:1000 -- this would mean "transfer 1000 POCKET to my personal savings account." Then if you wanted to withdraw your Pocket tokens from your savings account, you'd send a transaction to yourself with a memo that says pocketunvault:1000. Or something like that.

• Who would benefit? Anybody who doesn't want their posting key controlling their POCKET tokens.
• Would anybody be worse off? Only thieves would be worse off. Anybody who doesn't care about posting vs active authorities would just keep on using Pocket in the same way as always.

I have a vague feeling that something like @cryptick's idea about using Pocket as a curation reward could possibly be Pareto-improving.

What would be a bad upgrade for Pocket?

Here, we'd have to come up with a decent way of describing what it means for a Pocket user to be worse off. I think it's fairly reasonable to say that "subjectively" worse-off counts here, so in my mind this means for example that we should not do anything which increases the total token supply. Why? Because in the world of cryptocurrencies, people tend to believe inflation is evil (never mind the fact that almost all cryptocurrencies are inflationary). Since many of the holders of Pocket tokens are those people, I don't want to invite their ire by adding some kind of inflationary model to what was a token system with a static supply.

Another example of a bad upgrade might be a change to the fee for submitting a pocketsend. If the fee were increased, that would make people worse off who are sending Pocket tokens around, even though it could indirectly make everybody better off by increasing the incentive for confirmation bots to operate.

Is there a gray area?

Are there changes that are beneficial that might make people subjectively worse off but we can't agree on whether it's true? My approach in this case would just be to leave it well enough alone. Pocket is an experiment, let's let it run its course!

For more information

• Trade POCKET on Bitshares
• Pocket Announcement Post
• Pocket Protocol Specification
• @biophil's Pocket Confirmer Github
• My meta-subchain post
• Dan's original post on subchains

And as always, if you need a Bitshares account so you can trade POCKET IOU tokens issued by @virtualgrowth, you can sign up for Bitshares using this link.