We present the results of an experience in design and evaluation of multimedia guides for art exhibitions based on Apple iPod-class devices, gesture based interaction and rich multimedia content. Design is based on decoupling between content, interface and interaction design, adaptation to different user skills, and reusability; the evaluation is based on logging users' actions, time spent in the visit and content access. We report about the lesson learned in building and testing on the field two different guides during a period of six months.

We discuss the design and implementation environment of a family of multimedia guides for art exhibitions on Apple iPod touch devices. The project aims at developing a unique framework adaptable to different content types and presentation styles, while retaining a common interface model with consistent gestures and a fast development process based on standard environments and toolkits. Evaluation is based on automatic collection of information about the user behavior by logging user actions.

Galileo, a programming language for database applications, is presented. Galileo is a strongly-typed, interactive programming language designed specifically to support semantic data model features (classification, aggregation, and specialization), as well as the abstraction mechanisms of modern programming languages (types, abstract types, and modularization). The main contributions of Galileo are (a) a flexible type system to model database structure and semantic integrity constraints; (b) the inclusion of type hierarchies to support the specialization abstraction mechanisms of semantic data models; (c) a modularization mechanism to structure data and operations into interrelated units (d) the integration of abstraction mechanisms into an expression-based language that allows interactive use of the database without resorting to a new stand-alone query language. Galileo will be used in the immediate future as a tool for database design and, in the long term, as a high-level interface for DBMSs.

Given r positive integers s1, s2, …, sr with an associated profit pi two problems are at the root of several interesting applications. The ‘M-partition problem’, that is determining all possible combinations of these numbers which sum to M, and the ‘Knapsack problem’, that is determining a combination of these numbers maximising the pi sum subject to the condition that the si sum is not greater than M. A solution is proposed for both the problems based on a tree search approach which leads to algorithms with simple structure, linear storage requirements and an average computation time lower than other proposed algorithms.

The rectangular cutting stock problem is to determine how to cut a specified number of each of certain types of rectangular pieces out of a given stock of rectangular sheets, the object being to minimise the waste. A special case of this general cutting problem encountered in many practical applications is one in which all cuts have to be accomplished from one edge of the rectangle to the opposite one, i.e. the cut has to be of a ‘guillotine’ type. For such applications the problem can be formulated as a mathematical optimisation program and the optimal solution can be found efficiently in terms of computer time only for problems of medium size. A heuristic method is described for an approximate solution of larger problems, which gives suboptimal results but presents an interesting tradeoff between computing effort and solution optimality.