Interest in the behaviour of carbonate soils has increased in recent years owing to the difficulties which have arisen with the foundations of offshore structures located on these sediments (McClelland 1988). Carbonate sediments are widely distributed on the continental shelves in temperate and tropical areas of the world and exist in many regions where there arc substantial petrochemical reserves (e.g., the Arabian Gulf and around the coasts of Australia and India).

Many partly saturated soils undergo a reduction in volume when their moisture content is increased. The phenomenon, which can occur without any change in applied total stress, was termed ‘collapse compression’ because it was considered to be associated with a collapse of the soil structure. It has also been termed ‘hydrocompression’, ‘hydrocompaction’, ‘hydroconsolidation’, and ‘saturation shrinkage’, but collapse compression remains the most widely adopted expression. Soils with a potential for collapse are referred to as ‘collapsible’ or ‘metastable’. The process causing the increase of moisture content has been described as inundation, permeation, saturation, sluicing, or wetting, depending on the circumstances. Collapse compression can have an adverse effect on the performance of earth structures and in some situations susceptibility to collapse compression may be the most significant geotechnical property of the fill.

Tailings is the name given to the waste output from the tail end of an industrial processing plant. About 5x10° tonnes/year of tailings are produced in the world. This amount greatly exceeds the amount of fill handled by the civil engineering profession in the construction of embankment dams, motorway embankments, and all other earthworks. The continuous disposal of tailings is essential for the continuing operation of the processing plant. In order to continue the profitable operation of the plant, the owners require methods of disposal that are as cheap as possible consistent with absolute safety. Any failure in the method of disposal could shut down the processing plant causing loss of profit plus payment to unemployed staff as well as cost of repair and any claims that could arise from the failure. The disposal of tailings can be regarded as a world problem because of their large volumes and the effect they can have on the environment.

Deeply weathered residual clay and swelling soft rock is widespread in Southern Africa (see Fig. 1 for the distribution) and, like heaving clay, can have a major effect on the performance of engineering structures in the area of intluence. It would appear that the distress in surface structures founded on deep soft rock, which swells, is similar to that experienced on transported clays, or even on shallow residual clays, being caused by differential heave of the underlying material. However, deep swelling rock may also influence other engineering structures, such as tunnels or mining works, by causing heaving in underground floor levels, deterioration (slaking) of roofs and sidewalls of excavations, and crushing of concrete linings.

The common approach to most geotechnical problems has been based on instituted rules that provide reasonably correct answers on the ground of a number of input data such as: in situ soil properties, anticipated loadings, geometrical constrains, and criteria that ensure reasonable margin of safety for the behaviour of the analysed structure.

At the 1985 Annual Convention of the American Society of Civil Engineers, Dr. Jorj O. Osterberg (1989) presented the 21st Terzaghi Lecture. His topic and the title of his paper was 'Necessary Redundancy in Geotechnical Engineering'. Dr. Osterberg quoted the American Heritage Dictionary in defining 'redundant': '. .. duplication or repetition of elements in electronic or mechanical equipment to provide alternative functional channels in case of failure'. He then presented his own definition: 'Designing, incorporation, and including physical and human processes into analysis, design, and construction in such a way that if one element, whether physical or human, fails to function or fails to function in the way intended, other elements take over in such a way that the structure will still function essentially as intended'. The lecture developed the theme of redundancy in the following phases of geotechnical engineering: Reconnaissance and preliminary exploration Soil borings Laboratory testing Analysis Design Construction

The operation of a machine results in the generation of unbalanced dynamic forces and moments which are transmitted to the foundation and the underlying soil. The foundation for the machine must therefore be designed to ensure stability under the combined effect of static and dynamic loads. The dynamic nature of the loads makes the problem of analysis and design of a machine foundation somewhat complex. Even though the magnitude of dynamic load is small, it is applied repetitively over long periods of time. The vibration response of a machine-foundation-soil system is defined by its natural frequency and the amplitude of vibration under normal operating conditions of the machine. These are the two most important parameters to be determined in designing the foundation for any machine.

Since ancient times natural geotextiles such as timber, bamboo, twigs, reeds. jute, hemp, and palm fibre have been used for the reinforcement of the earth together with the enrichment of it with additional bending and tensile strength in the construction of embankments, stabilisation of weak foundations, and flood control. In some cases these materials are employed to achieve structural effectiveness by means of laying them in the earth in one or multiple layers or by integrating the earth and gravels within mats, cells, and gabions (Yamanouchi 1992). These techniques were handed down from generation to generation and they are, at present, replaced by methods using steel and synthetic materials. However, in some countries the natural geotextiles are still used as effective reinforcing materials (Datye 1987). The use of soil walls for housing construction that gives the earth tensile strength by composing it with straw is known, at present, as the art of Texsol (Le flaive 1982): it integrates the earth by means of long filaments.

The first requirement in the control of uncertainty in geotechnical engineering is an appropriate attitude of mind which accepts that uncertainty will exist even in the most controlled circumstances. Those whose understanding of geotechnics centres around the text book, analysis, and controlled laboratory experiments expect a foreseeable pattern of behaviour, a set of defined relationships between cause and effect. But the engineering behaviour of soils and rocks is not of this type; on the inscrutability of nature has been superimposed the technical and litigious limitations of man as revealed by the evolving practices of engineering design and the often irrational complexities of contractual relationships. The objective of this account is to attempt to identify and classify the nature and causes of uncertainty, followed by advice on strategy to contain it, illustrated by a number of examples of how this may be achieved (including how not to do it). Many of the principles apply broadly to many aspects of engineering.