Applications of modern, computer (software) controlled measuring systems are various, from automation of testing and calibration laboratories to legal metrology measuring instruments and systems. Requirements to be met by such instruments or systems are usually defined by national or international legislation or international technical, safety, or laboratory competence standards. Validation of fulfilment of these requirements is essential for the conformity assessment of these instruments and systems. The complexity of validation and resources needed in the validation process depend on the construction of the object of validation and may be very extensive, especially in the case of computer based measuring systems. Suggested approach for modular validation of complex measuring systems may reduce the validation efforts, as well as costs and time consumption.

The paper is on the usefulness of global optimization for developing variational algorithms of spectrophotometer calibration. A new nonlinear method for reconstruction of absorption spectrum, on the basis of spectrophotometric data, is proposed and used for demonstrating that the quality of reconstruction depends on the ability of the calibration procedure to reach a global minimum during optimization of the parameters of the operator of reconstruction. The results obtained using calibration algorithms based on local and global optimization are compared in terms of measurement uncertainty and artefact content.

This paper considers the determination of measurement results and associated uncertainties when prior physical knowledge of the quantities concerned is available. The scientific concepts that provide a basis for determining realistic solutions to such problems are discussed, and implementations of these concepts are considered. To illustrate the concepts, an example concerning the determination of an analyte concentration in chemical metrology is used. This will be the basis for a discussion concerning the various approaches available to deal with such constraints. Results stating a coverage interval containing infeasible values (values the quantity cannot physically take) should be avoided, and this fact will assist in the comparison of the relative merits of each method.

The weighted mean has been used to estimate the common mean of several populations with unknown and different variances. However, the traditional estimator of the variance of the weighted mean estimator underestimates the variance. Two new variance estimators are proposed with smaller biases and correspondingly formed intervals that have much better coverage probabilities for the mean. Results are extended to the general case with both Type A and Type B uncertainty components being presented.

The experimental design, active statistical tool generally used for the optimization of processes, can also be considered for the evaluation of sample(s) homogeneity. This tool may be applied to Metrology for the analysis of large amount of repeated measurements permitting the "mining" of the results and include this "time-dependent sources of variability" information at the uncertainty calculation.