Function to check quorum for arbitrary number of participants

[CICS-Oleg]

Suppose we have a set of stakeholders with corresponding stakes in %.

(= (spk_holder h1) 2)

(= (spk_holder h2) 4)

(= (spk_holder h3) 5)

(= (spk_holder h4) 2)

(= (spk_holder h5) 0.5)

To start а discussion it has to be a minimum quorum thresh based on the aggregate stake size of the participants. Arbitrary number of participants can come to a meeting. What is the best way/representation to make a function to determine if a quorum is present?

[Necr0x0Der]

Well, in a slightly awkward way, it can be done with something like this:

(= (spk_holder h1) 2)
(= (spk_holder h2) 4)
(= (spk_holder h3) 5)
(= (spk_holder h4) 2)
(= (spk_holder h5) 0.5)

(= (stake-size $participants)
   (let $lst (collapse (spk_holder (superpose $participants)))
       (foldl-atom $lst 0 $x $y (+ $x $y)))
)

(= (enough-stake-size? $participants $threshold)
   (>= (stake-size $participants) $threshold)
)

!(enough-stake-size? (h1 h3 h5) 7)
!(enough-stake-size? (h1 h3 h5) 8)

[Necr0x0Der]

What is slightly inconvenient here is that we need to use superpose and collapse. It seems like a common pattern that we want to fold over non-deterministic results. Having functions like foldl-nd or even car-nd might be nice, but the latter may need reimplementing non-deterministic results as streams, which was already discussed, but may require more considerations.

Another minor thing is that it would be more traditional to have lambdas, so instead of passing separate arguments $x, $y, (+ $x $y) we could write (\ ($x $y) (+ $x $y)) and, hence, just +. I'm not sure to what extent it is easily doable now.

So, I'm not proposing these changes specifically for the present issue, but just want to mention these points, @vsbogd , for a later discussion

[vsbogd]

I believe one can use map-atom instead of (collapse (spk_holder (superpose $participants)), but collapse/superpose looks even more compact.

Yeah, it is interesting point about non-deterministic fold/car. Let's discuss it further later.

[Adam-Vandervorst]

I think it's completely reasonable to add summation and other MPI reduction strategies to spaces/nondeterministic results.
I.e. (summation (spk_holder (superpose $participants))).