CPS-0017? | Settlement Speed #922
Conversation
There was a problem hiding this comment.
@bwbush @abailly very good to see this, as a problem statement as well as another item for the growing Consensus category (please contribute about the related issue #898 when you can).
Though there are some TODO's I'm marking this for introduction at the next CIP meeting so the community can start following & participating: https://hackmd.io/@cip-editors/98
rphair
added
State: Triage
Co-authored-by: Robert Phair <[email protected]>
Co-authored-by: Robert Phair <[email protected]>
CPS-0STL/README.md
Outdated
| There are several definitions of "settlement" or "finality", and precision is important when discussing these. Two noteworthy scenarios can be defined precisely. | ||
|
|
||
| - *Ex ante* settlement probability: "What is the probability that a transaction that I just submitted will ever be rolled back?" | ||
| - *Post facto* settlement probability: "Given that I submitted my transaction $x$ seconds ago and it has not yet been rolled back, what is the probability that it will ever be rolled back?" |
There was a problem hiding this comment.
Nitpicking: wouldn't ex post be a better counterpart to ex ante?
There was a problem hiding this comment.
Are not the two scenarios just the same, where the first scenario is just the case x = 0?
CPS-0STL/README.md
Outdated
|
|
||
| Even better would be to have an on-chain indication that transactions up to a particular point have been settled with high probability. | ||
|
|
||
| If one is unwilling or unable to re-submit a rolled-back transaction, then the *ex ante* probability might be of most interest. This matches use cases where there is no opportunities for the parties involved in a transaction to resubmit it: for example, one party might have purchased physical goods and left the vendor's premises, leaving no chance to resubmit a rolled-back transaction. |
There was a problem hiding this comment.
suggestion: move use cases examples before the definition of ex ante/post facto to help the reader put the definitions in context
There was a problem hiding this comment.
This might make sense, but it violates the CPS template.
There was a problem hiding this comment.
@ghost @bwbush - moving the entire Use Cases section before the Problem section would definitely run against the template, but we do have precedent for moving some brief examples before a detailed specification (as in CIP-0013). The guideline I think authors & editors should use: Does it improve reader comprehension, while avoiding duplication of content or ideas?
rphair
added
State: Confirmed
There was a problem hiding this comment.
@bwbush the CIP meeting today was in consensus that this problem statement is valid & that this document will keep moving forward. Please update the CPS number and rename the containing directory accordingly (and also please update the document link in the first posting). 🚀
Co-authored-by: Robert Phair <[email protected]>
Co-authored-by: Robert Phair <[email protected]>
- Change "post facto" to "ex post facto" throughout. - Included suggested changes to text. - Added reference to Midnight and Bitcoin OS bridges. - Slightly strengthened discussion of use cases. - Removed outstanding "to do" notes.
|
@rphair, a4960f0 addresses all of the outstanding reviewer comments and strengthens the text in several places. There was a suggestion to move the use-case section upward, but I didn't do that because it would clash with the CPS template.
|
|
very good then @bwbush let's put this on Review for the next CIP meeting (https://hackmd.io/@cip-editors/101) and hopefully confirm this should be promoted to |
|
@Ryun1 @perturbing - did we have a reason not to promise this to @bwbush please let me emphasise, as I posted in #922 (comment), that moving up an illustrative example or two (to precede the full inventory of Use Cases) would not be a violation of the CPS template as supposed in #922 (comment): so you could follow that suggestion, or not, as you see fit. |
This CPS defined the settlement problem for Cardano and supports CIPs like Peras, Anti-Grinding, and changes to the active-slot coefficient.
(rendered latest document)