Towards interoperability of heterogeneous distributed software repositories.
Item
-
Title
-
Towards interoperability of heterogeneous distributed software repositories.
-
Identifier
-
AAI9530918
-
identifier
-
9530918
-
Creator
-
Schreiber, Mindy Rosman.
-
Contributor
-
Adviser: Keith Harrow
-
Date
-
1995
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Computer Science
-
Abstract
-
This thesis deals with the problem of sharing software among heterogeneous distributed software repositories. Several issues must be resolved in order to successfully share software among many repositories. This thesis addresses two of these issues: (a) transferring information among heterogeneous distributed repositories; (b) establishing a classification scheme for the purpose of storing and retrieving the software or collections of software in packages.;During the course of this research, it was determined that the use of a data independent data structure, such as MARC records, would allow the efficient transfer of corresponding information among several repositories. Tools were created to extract package information from repositories to be placed into MARC records, and also to manipulate and extract information from the MARC records, thereby completing the transfer of this information to other repositories.;Then a taxonomy for a dynamic classification scheme, capable of growing and evolving to satisfy the changing needs of the software community, was developed. The hierarchical data structure designed for this purpose is physically a network; however, it can be viewed logically as a forest of trees. Two objectives are accomplished with the use of such a data structure. First, there is no redundancy of nodes in the network; in addition, because the network can be traversed like a series of trees, there is no infinite cycling at any point during traversal.;Two methods of generating a network which can be traversed like a forest of trees are discussed in detail. The algorithms and data structures provided to describe the taxonomy contain the first method of generating this network. The second method, which is very flexible, is discussed in a section of its own. Performance issues regarding the new taxonomy are also discussed.;Further research directions regarding the automatic generation of USMARC records and the pros and cons of distributed vs. centralized software union cataloging are discussed.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
