Here economy of means would lead me to prefer an SRST to Jon
Perryman's traditional approach, which searches for a delimiter or
delimiters using a TRT[E].

Just where the break point is depends upon how many delimiters are
being searched for concurrently.  When, as here, it is a single
character (a blank), I would use the SRST.  When it is many, a single
TRT[E] and its 256-byte table are often preferable to a sequence of
SRSTs even when that sequence occupies less space than the TRT[E]
table.

This analysis is of course appropriate only for a string of known
length, either because it is always B bytes in length or because it is
prefixed by a byte or halfword current-length count K.  For C or
C-like strings of "conceptually unlimited" length, I would always use
a TRTE, since it permits a stop on a nul delimiter, x'00', to be
specified.

John Gilmore, Ashland, MA 01721 - USA

Reply via email to