Thoughts on laziness (long)

[matthias@REDACTED]

Pierpaolo BERNARDI writes:

 > I too implemented ralists.  If I remember correctly, my implementation
 > was about 5/6 times slower than built-in lists in a
 > cons/head/tail-based benchmark.

Thanks for the code. Your implementation is cleaner and more complete
than mine.  Running it with the same test driver as I ran the other
code, I get no significant difference in performance:

  N       Sum   Update   Original ra-list
  ----------------------------------------
  100:    981      89    (962, 84)
  1000:  9731     125    (9903, 127)
  10000:   98k    183    (93k, 168)
  100k:  1501k    248    (1357k, 188)

Matthias