Browsing by Author "WANG, FJ"

Now showing 1 - 9 of 9

A WORST-CASE OF CIRCULARITY TEST ALGORITHMS FOR ATTRIBUTE GRAMMARS
(1995-03-01)
WU, PC
;
WANG, FJ
;
交大名義發表
;
資訊工程學系
;
National Chiao Tung University
Although the circularity test problem for attribute grammars (AGs) has been proven to be intrinsically exponential, to date, a worst case for the existing circularity test algorithms has yet to be presented. This note presents a worst-case AG in which the number of incomparable dependency graphs induced at the root is exponential. The worst case can help to clarify the complexity of the problem.
AN OBJECT-ORIENTED APPROACH TO CONSTRUCTING COMMUNICATION PROTOCOLS
(1995-05-01)
CHEN, JL
;
WANG, FJ
;
TING, YC
;
交大名義發表
;
資訊科學與工程研究所
The development of communication protocols for computer networks and distributed systems is an increasingly complex and cost-sensitive process. This paper presents an object-oriented concurrent (OOC) model for the development of communication protocols. This model consists of three kinds of entities: data entities, which represent the communicated data units, state entities, which describe their behavior, and connection, entities, which are responsible for communication services. The state transitions in a protocol are modeled by using state entities, and a communication service is performed by a group of connection and state entities. A C++ library based on our OOC model is described that contains three class hierarchies, for data, state; and connection entities, respectively In addition, an approach to constructing a protocol using this OOC model and the library is presented. An example in which a T.62 protocol is constructed shows that this approach provides a high level of modeling, concurrency, and reusability.
AN OBJECT-ORIENTED SPECIFICATION FOR COMPILER
(1992-01-01)
WU, PC
;
WANG, FJ
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
In this paper, we present an object-oriented approach to compiler specification. Our method treats grammar symbols as templates which instantiate nodes of parse tree for object-oriented semantic analysis. To have better classification and inheritance for semantic description, it uses restricted CFG to specify grammar rules. The semantic specification is done based on a class hierarchy generated from the restricted CFG. Besides, that a path expression is booted to describe the possible receivers of a message releases the restriction on the specification of message (value) propagations in attribute grammar methods.
DESIGNING A REUSABLE SYMBOL TABLE LIBRARY
(1994)
WU, PC
;
LIN, JH
;
WANG, FJ
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
IS CIRCULARITY PROBLEM FOR ATTRIBUTE GRAMMARS EXPONENTIAL-TIME COMPLETE
(1994-07-01)
WU, PC
;
WANG, FJ
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
OBJECT MOVEMENT IN DISTRIBUTED OBJECT-ORIENTED SYSTEMS
(1994-05-01)
LEE, YS
;
WANG, FJ
;
CHANG, CH
;
交大名義發表
;
資訊科學與工程研究所
In a distributed object-oriented system-a system containing more than one object-oriented system (OOS)-different OOSs may contain classes with different definitions, different definitions but the same name, or the same definition but different names. Moving an object between two OOSs includes locating a suitable class and creating an instance of the class with the value of the moved object in the destination system. This study discusses what is a suitable class for a moved object in the destination OOS and presents the algorithms for finding such a suitable class. It also discusses the activities of copying the value of the moved object into the instance of a located class in the destination site. A general object movement mechanism and its implementation for a domain specific model are then incorporated into a prototype of a distributed OOS, called DOOSPE.
PADDED STRING - TREATING STRING AS SEQUENCE OF MACHINE WORDS
(1994-09-01)
WU, PC
;
WANG, FJ
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
A string is a sequence of characters. The operations such as copy and comparison on strings are usually performed character by character. This note presents a data type called padded string, a string type with faster operations. A padded string is a sequence of machine words. For 32-bit machines, four characters can be operated in one machine instruction. Operations on padded strings can then run faster than traditional strings such as char* in C language. An experiment sorting an array of strings shows speedup 24% using padded string.
POOL - AN UNBOUNDED ARRAY
(1994-09-01)
WU, PC
;
WANG, FJ
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
Many collection objects, e.g., stack, can be implemented with tabular representation (an array). Such an implementation provides fast index and saves space on linked structures. However, its space is bounded at creation time. This note presents a hybrid data structure, called pool, which combines tabular and indexed data structures. A pool is a number of segments of contiguous memory space. It provides unbounded space. Its first segment is allocated initially. When more space is needed, a new segment is allocated in double size of previous segment. All segments allocated can be accessed via an index table. The size of index table is log n, where n is the number of elements. A pool can provide index operation in O(log n) and iterations almost as fast as tabular implementation.
SCANNING REGULAR LANGUAGES BY DUAL FINITE AUTOMATA
(1992-04-01)
WU, PC
;
WANG, FJ
;
YOUNG, KR
;
資訊科學與工程研究所
;
Institute of Computer Science and Engineering
A regular language is generally accepted by a single finite automaton. An approach of dual finite automata is presented here. An input string is scanned by two deterministic finite automata (DFA's): reading from the string's head and tail respectively. One of them accepts the regular language itself; the other accepts the language's reversal. Whether a string is accepted depends on the states of both automata, when their reading heads meet. Dual finite automata can be applied in compiler generation and parallel computing.