A decade ago I tried to initiate such a project --to wit:
http://igkt.net/sm/index.php?topic=1006.msg6785#msg6785
My first complaint is that the [u]ABoK Index gives page
vs. #'d-prg. citations/numbers. This is problematic in that
of course the latter are more specific AND there are
cases where the index’d item occurs more than once
on a given page (and w/o explicit reference to these
multiple places, one might be missed upon finding
another).
Also, it occurred to me that the Index could do more
than give mere location : it could also indicate what
sort of information --generally, at least-- was to be
found there (from a mere mention of some knot in
discussion of other things, to some fuller information
about the knot).
(I vaguely recall musing about more than just two such
categories, and some graphical means to distinguishing
them --partly in order or presentation (e.g., list the
fullest treatments first, then others in some order),
but with some indicator of the nature of the groups
(for if >2 classes possible, even if one presumes that
the primary info class is always present, there’d need
to be indication of what the other citations were for,
where not all classes were present). )
Somewhat related: one of my favourites is a knot in ABOK which is structurally identical to the Munter Hitch, but can't be reasonably called such. Anyone know which one I'm referring to?Ashley's [i]"crossing knot"[/i] fits the bill, at least for one of the states of the meso barcaolo / Italian knot / Munter hitch / Halbmastwurfsicherung (German, meaning "what-a-mouth-full!") --i.e., not in the braking state but where the ends head in opposite directions.
(It’s a bit of a philosophical problem as to whether
one considers this structure as a knot.)