Relationships between elastic and plastic theory, design and analysis methods, and the upper and lower bound theorems are discussed. It is proposed that the use of modern analytical techniques by structural engineers in the assessment of existing structures may lead them to a false conclusion that the structure is inadequate. This is illustrated using a reinforced concrete rectangular slab simply supported on all sides as a hypothetical example. It is assumed that the structure was designed in 1970 by an engineer familiar with methods for the elastic analysis of slabs, and also with the Hillerborg strength method of design. The slab is assessed today by engineers using finite element analysis incorporating the Wood-Armer equations. It is pointed out that a failing structure must have a collapse mechanism, and that the collapse load may be computed by an upper bound analysis. Also, a computer analysis is only as accurate as the underlying assumptions. 8 refs.