# CIP21 - Dispensers

**URL:** <https://forums.counterparty.io/t/cip21-dispensers/5488>\
**Category:** Counterparty\
**Created:** [May 23, 2019, 12:07am UTC](https://forums.counterparty.io/t/cip21-dispensers/5488 "2019-05-23T00:07:13Z")\
**Posts on this page:** 1\
**Showing post:** 12

<div class="post-metadata">

**Author:** ![loon3](https://yyz2.discourse-cdn.com/flex004/user_avatar/forums.counterparty.io/loon3/32/1018_2.png) [@loon3](https://forums.counterparty.io/u/loon3)\
**Post date:** [January 23, 2023, 7:10pm UTC](https://forums.counterparty.io/t/cip21-dispensers/5488/12 "2023-01-23T19:10:57Z")

</div>

> [@jdogresorg](#):
>
> I support this idea, but I propose a tweak to this, which is that we only delay closing the dispenser if there is a pending dispense. This would solve the problem of dispenser operators closing dispensers with a pending dispense, but also allow dispenser operators to close their dispenser immediately if there is no pending txs (BTC incoming) to the dispenser address. Thoughts?

The problem here is you can’t do this at a protocol level because the mempool is not the same across nodes. So there’s really no way for counterparty-lib to detect pending dispenses. You could probably be closer to 5 block delay rather than 10 and still get the benefit of preventing scam dispensers. That should be a short enough amount of time that people aren’t inconvenienced too much.

---

_[View the full topic](https://forums.counterparty.io/t/cip21-dispensers/5488)._