On Wed, Aug 23, 2006 at 03:19:22PM -0700, Larry Wall wrote:
: But I'd still probably use a hyper-fatarrow for this case rather than
: relying on interleaving.

Another reason for preferring hyper is that it makes promises about
parallelizability, whereas the zip/each solutions would tend to
assume the input streams must be processed in order, albeit lazily.
(Probably doesn't make much difference until someone actually attempts
to vectorize Perl though, and even when that happens, it's not clear
what the exact sequence of events would be if you feed a lazy pipe
to a hyper...)


Reply via email to