@techreport{R-89-27,
TITLE = {An Object-Oriented Database Approach to the Symbol Processing in an Incremental Compiler},
AUTHOR = {Peter Fritzson and Petr Kroha },
YEAR = {1989},
NUMBER = {R-89-27},
INSTITUTION = ida,
ADDRESS = idaaddr,
ABSTRACTURL = {/publications/cgi-bin/tr-fetch.pl?r-89-27+abstr},
ABSTRACT = {This report introduces an object-oriented database used as a symbol table for an incremental compiler. It is argued that a pure relational model is unsuitable for this task. The used model is conceptually simple, but powerful enough to model languages of the complexity of Ada. We describe the conceptual schema in the EER-model terminology, the external schema as a set of specific features of each programming language (visibility rules) and the object-oriented implementation.An important area of practical concern is addressed, which is the need for efficient implementation. The primary focus of efficiency is in storing and managing complex objects on disk. A simple but very efficient method for clustering objects on disk is described. A predictive method is used to guide the transfer of objects from disk to the heap. This clustering has a crucial influence on the efficiency. Finally it is shown how these clusters can be derived from the known (or supposed) structure of queries.},
IDANR = {LiTH-IDA-R-89-27}