IF doing multiple OP_RETURN, increase data carrier size

Consensus
#1

This is in reference to this topic: Multiple OP_RETURNS - This time for real!

If we start to allow use-cases for multiple OP_RETURNs per tx, in order to “merge” more than 1 protocol into 1 tx – it might be beneficial to also increase the data carrier size limit per tx from 223 to some integer multiple of 223, so that people can comfortably fit more than 1 full protocol message in a tx.

For example 446 - can fit 2 full protocol messages. Or 669, can fit 3, etc.

I think this may make the multiple OP_RETURN thing more useful.

2 Likes
#2

I agreed to multiple op_returns because there is no extra cost to the network while allowing more innovation.
Your suggestion still needs actual numbers, actual examples that run on testnet for some time and all that jazz. It has a real effect on non-coin data storage which is not something I want to just ignore either.

If you want to push this, do it the right way, please.

2 Likes
#3

Your suggestion still needs actual numbers, actual examples that run on testnet for some time and all that jazz.

True. Plus there are downsides (like just blockchain bloat).

If you want to push this, do it the right way, please.

Nah, it can wait. I think even just the low-impact multi-OP_RETURN thing you and @BigBlockIfTrue worked on is a huge improvement.

3 Likes
#4

I think it makes sense to reconsider these limits once we have actual use cases that are hitting up against these limits and see that the value proposition is clear. I personally think a larger size is perfectly fine but there’s a significant amount of push back against any byte added by the dev community in general. For our own protocol that runs alongside SLP, we can create 19 OP_RETURN outputs in a single transaction. That’s likely to be “good enough” for some time. I think the most likely source of pressure (if any) to make this larger will be the reality that protocols will have to start adding some identifying content in their OP_RETURNs since they can’t count on being the only ones any longer and will need to do some pre-processing to make sure which one is their’s.

Let’s get multiple OP_RETURN in asap and see how the experiments go!

2 Likes
#5

Yeah that makes sense. No disagreement here.

1 Like