User blog comment:DrCeasium/Hyperfactorial array notation: Analysis part 2/@comment-7484840-20130530171941

Okay, I think I've made a slightly better version of the w/ operator (no magic ...'s, but the same strength). It's not brilliant, but it's certainly better (italic things may or may not exist):

@[@1[1]w(k)/[q@]@2]@ = @[@1[1(k)1(k)1(k)...(k)1(k)2]w(k)/[1@]@2]@, where there are q 1's, and @1 and 2 are just continuing w/ chains, and the arrays in @1 (if there are any).

@[@1[@]w(k)/[q@]@2]@. The [@] will be evaluated (doesn't really seem the right word since they don't actually get a value. Any suggestions?) as per normal, but when the brackets are nested, instead of just using a slightly altered version of [@], use [@]w(k)/[q@]@2 (with the slightly altered version of [@]).

@[@]w(k)/1@ = @[@]@ (remove trailing 1's from the chain).

If no (k) is specified for w(k)/, it defaults to 0 (leading to ,s being used).

Extra information can be specified about the w(k)/, for example, a w(0)/ could be specified to produce the string of 1,'s on the second row, or in the 3rd realm of the 5th flune.