I'm not too happy with the current state of the
documentation? ;) of the series package.
The implementation has more features than CLtL2
For example, the following are undocumented:
scan-range specifies :start but example &
implementation are using :from
The often cited AI-memos 1082 and 1083
are poor quality scans (.tiff embedded in .ps) of old
Richard Waters. Precious, but possibly inaccurate, out
of date (1989).
They contain valuable information. For example, the
mentions why I couldn't define my own transducer as a
&key or &rest parameters. &optional is all that is
possible. Using a
macro is suggested instead. This failed attempt cost
me (and Raymond
Toy) time and e-mail trafic.
There are rumours about papers from 1991 (ACM) or
I volunteer to participate in updating an eletronic
R.toy wrote: "I should ask Dr. Waters if he still has
the original sources for the
documents." -- with what shall we start?