3gl rerunnable batch, monte carlo entrypoints
is_vc?(account)
take https://twitter.com/mlevchin/lists/memberships as sample
strat sketch (inwards)
go into every list, extract accounts
if account is a (known with weight) vc: recurse
...and accumulate weights
result = new set {}
lists_processed = new set {}
function find_vcs(listname, initial_weight)
lists = twitter.com/[account]/list/memberships
foreach list in lists
break if list already processed
accounts = twitter.com/{list accounts}
foreach account in accounts
if account ~ vc
yield find_vcs (account, weight(account))
else
yield account, weight
strat coprocessing
parallel spider lists; yield process(list)
temporality or by assigned region or semaphores? (when mapping out: actively, overlap-detection: passive, sideline companion: queue deduplicator)
strat weights
the process is ~ endlessly looping
when finding a list -> weight+ ,
strat graph-processing - vizual dump
== now: reverse process
strat outwards
spider for depth n, cutoff, cluster: density.
collect, merge w/ density threshold
walk connectors.
