We present Mavigator, a prototype of a graphical user interface to databases. The system is dedicated to naive users (computer non-professionals) and allows them to retrieve information from any data source, including object-oriented and XML-oriented databases. The system extends its core functionalities by the Active Extensions (AE) module, which assumes a trade-off between simplicity of user retrieval interfaces and complexity of output formatting functions. In AE the latter are to be done by a programmer using a fully-fledged programming language (currently C#). Thus the retrieved data can be post-processed or presented in any conceivable visual form. Another novel feature of the Mavigator is the Virtual Schemas module, which allows customization of a database schema, in particular, changing some names, adding new associations or hiding some classes.

The paper discusses a special phase, called static analysis, which we use for static query optimization. Static analysis, as a compile-time simulation of run-time query evaluation, enables us to gather all the information that we need to optimize a query. We present in detail all the major data structures used by static analysis and discuss how they emulate their run-time equivalents. The paper includes also the algorithm of static analysis in pseudocode.

An overlay cluster for a C program is a collection of C functions that are exclusively called from a given collection of functions. An algorithm (implemented on IBM XT/AT, MS DOS compatibles) for making overlay clusters is presented. It much simplifies design of overlays for large C programs.

The paper presents a simple algorithm for correcting lexical errors in data. The method employs the concept of a finite automaton. It effectively works in the experimental DBMS LINDA.

Semantics determines the meaning of language constructs; hence it says much more than syntax does about implementing the language. The main purpose of this paper is a formal presentation of the meaning of basic language constructs employed in many database languages (sublanguages). Therefore, stylized query languages SSL (Sample Selection Language) and J (Joins) are introduced, wherein most of the typical entries present in other query languages are collected. The semantics of SSL and J are defined by means of the denotational method and explained informally. In SSL and J, four types of expressions are introduced: a selector (denotes a set of addresses), a term (denotes a set of values), a formula (denotes a truth value), and a join (denotes a set of n-tuples of addresses or values). In many cases alternative semantics are given and discussed. In order to obtain more general properties of the proposed languages, a new database access model is introduced, intended to be a tool for the description of the logical access paths to data. In particular, the access paths of the network and relational models can be described. SSL and J expressions may be addressed to both data structures. In the case of the relational model, expressions of J are similar to SQL or QUEL statements. Thus J may be considered a generalization of relational query languages for the network model. Finally, a programming language, based on SSL and J, is outlined, and the issues of SSL and J implementation are considered.