About exception of redundancy in the data

It is saved as a reasonable demand of unitary data input in a DB for the decision of different problems{tasks} and protection against occurrence of contradictions (infringement of logic integrity) at actualization of the stored{kept} data. However, in conditions of global information space and componental designing the context of these requirements should be reconsidered. Undoubtedly, in operational DB it is rational to plan "islands" normalized and, in classical sense, bezyzbytochnykh clusters of attitudes{relations} or objects. These "islands" more often also will be for a long time known subject DB.

At the same time, association of historical given Storehouses, a DB of GIS-SYSTEMS, archives of text documents, the streams of the information received on the Information data highway, etc. in the general{common} production of designing by a corporate DB results in refusal from universal and vseobjazatel`nogo a principle of exception of redundancy: designing of a corporate DB at a level of the logic circuit and at a conceptual level does not lean{base} as on global criterion on the requirement and procedures of exception of redundancy in the data.

The new situation, including essential and beforehand independent stream of the information from the Information data highway in a corporate DB will demand development or increase in opportunities of " procedures of an identification " copies of information structures, i.e. definition of that these copies describe the same subject of the real world.

Problem of preservation of problems

Character of discipline of the designing, stipulated by the cascade circuit, methods of structural designing both hierarchical approaches and structures, pushes now designers to fix rigidly enough certain models About. The case-technology of designing of a DB so that to exclude preservation of existing problems of the enterprises in rigid, "cel`nointegrirovannykh" structures of a DB should be changed. For this purpose change not only technologies, but also tools of designing can be demanded.

Prospective approaches:

Opportunity to fix the description of attributes, suhhnostej, communications{connections}, functions, etc. with any degree of incompleteness, opportunities to make the description at a level of not detailed, in detail connected sets of information structures (" clusters suhhnostej ");

Designing or reconstruction of models of components IS and DB, their integration in the general{common} conceptual space;

Designing of the ordered sequence of statuses of a corporate DB as sequences of sets of the maintained subject DB including: nasledovannye the DB structurally predetermined by a DB of "purchased" functional components, designed specially for the given enterprise of a DB, and the DB of two last categories gradually replace nasledovannye and, then and in parallel, replace each other in development IS;

Openness of a repository of case-system, dictionary SUBD and systems 4gl, allowing to build on metaobjects and mechanisms required tezaurusnymi and deep semantic attitudes{relations} between elements, and also to make a bilaterial exchange of a metainformation with other systems 4gl and case, to connect models of different components in one to use and preservation of all necessary semantic attitudes{relations}.

Componental openness and semantic interoperability

The componental approach in development IS demands componental designing of a DB. Replacement with some functional components IS on similar, but designed by other developer, will demand structural replacement of some part of a corporate DB. Such replacement should be supported as constant process of a redesign of a DB. At replacement components of a DB interfaces with her of available applications and their users should receive precisely the same in the semantic attitude{relation} the information, as earlier.

Real componental designing of a DB can be based on formation and use of the general{common} for kompleksiruemykh a component of conceptual model and maintenance of conformity between models a component of a DB (and the applications connected to them) and the general{common} conceptual model. In a general view of the requirement to formalizmam such models were described earlier (see [zinder90]). Program realizations of the full formal systems usually based on the objective approach which can come nearer to the tool decision of this problem{task} (see for example, [Kalinichenko93]) recently develop.