Clojurians Log v2
clojure 2025-12-08
where do folks sit with the reducers library vs ham-fisted these days, especially on parallel reducers/fold/etc? ham-fisted well supported/liked? or lessons learned absorbed into reducers??
except for fold c.c.reducers are "old technology" and totally subsumed by transducers
that's what I was getting at ... where's "the best fold" at?
I don't have any experience with ham-fisted, but you may look at aphyr/tesser as well
Why are they totally subsumed by transducers?
Chronologically they predate transducers, and transducers are more general and decomplected
Reducers couple xform and source coll together
But is there an interface for running transducers in parallel?
Like I said, fold is the only thing that isn’t subsumed by transducers
Okay, but the whole point of reducers is to provide a fold/reduce interface. If transducers cannot do that, then surely reducers are not replaced by transducers.
I seem to recall a Zach Tellman conference talk about understanding the performance implications of seqs, but I can’t seem to find it now. Does anyone remember the name of the talk?
Thanks. This was good, but I have a vague recollection of a talk where Zach, I believe, walked through the class hierarchy of the collections (ISeq, IPersistentColl, APersistentVector, etc), to talk about cases where e.g. first is surprisingly much slower than (nth n 0)