Mitchell Marks and Kristian J. Hammond, A review of 
James F. Allen, Henry A. Kautz, Richard N. Pelavin, and Josh
D. Tenenberg, <I> Reasoning about plans</I>, Morgan Kaufmann, San Mateo,
CA, 1991; <I> ACM SIGART (Special Interest Group in Artificial
Intelligence) Bulletin</I>, 1993, Vol. 4, No. 2, pp. 8--11.  
<P>

The book that Marks and Hammond review describes the use of intervals
for planning. Marks and Hammond emphasize that the interval approach
to planning (originated from the works of Allen) is based on the
assumption that expert knowledge about time is <I> consistent</I>. If
expert specifications turn out to be inconsistent, then we must
inform the experts about this inconsistency and ask them to correct
their specifications. This tedious procedure is absolutely necessary
in critical situations, when every requirement is absolutely necessary
(<I>hard</I>). In real life, however, many requirements are <I> soft</I>
(recommendations rather than requirements), so inconsistency and the
resulting impossibility to satisfy all the requirements may be quite
admissible. 
<P>
Maria Beltran