yms@cis.ufl.edu (Yuh-Ming Shyy) (02/22/91)
Many works have been done in applying object-oriented database technology to software engineering. Currently I am also taking a similar approach to a DBMS-support software environment where we treat everything as objects. The problem is that if we treat type and method uniformly as first-class objects and store them in the database (as instances of class "TYPE" and "METHOD"), then the traditional programming style (using files) does not fit because information will be redundantly stored in both files and data- base, and each time the .h file is executed, how to handle those types and methods objects which have already been stored in the database from the previous execution? On the other hand, if we distribute these information as database objects without using files, then there is a problem if we want to execute the system using a persistent C++ (such as ODE and E). Any comment will be highly appreciated. Thanks. yms@reef.cis.ufl.edu
davidm@uunet.UU.NET (David S. Masterson) (02/23/91)
>>>>> On 21 Feb 91 21:32:13 GMT, yms@cis.ufl.edu (Yuh-Ming Shyy) said:
Yuh-Ming> Many works have been done in applying object-oriented database
Yuh-Ming> technology to software engineering. Currently I am also taking a
Yuh-Ming> similar approach to a DBMS-support software environment where we
Yuh-Ming> treat everything as objects. The problem is that if we treat type
Yuh-Ming> and method uniformly as first-class objects and store them in the
Yuh-Ming> database (as instances of class "TYPE" and "METHOD"), then the
Yuh-Ming> traditional programming style (using files) does not fit because
Yuh-Ming> information will be redundantly stored in both files and database,
Yuh-Ming> and each time the .h file is executed, how to handle those types and
Yuh-Ming> methods objects which have already been stored in the database from
Yuh-Ming> the previous execution? On the other hand, if we distribute these
Yuh-Ming> information as database objects without using files, then there is a
Yuh-Ming> problem if we want to execute the system using a persistent C++
Yuh-Ming> (such as ODE and E).
I didn't quite catch all that is behind this message, but I think there's
something here I may be interested in. Can you go into more detail?
1. Is the "DBMS-support software environment" an environment for supporting
the development of DBMS's (DB's?) or an environment for developing software
where the DBMS provides "persistent" information about the development?
2. How are you attempting to fit "traditional programming style (using files)"
into the "DBMS-support"? Why doesn't it fit?
3. What do you mean by ".h file is executed" and "previous execution"?
4. Maybe you can think of (4)?
--
====================================================================
David Masterson Consilium, Inc.
(415) 691-6311 640 Clyde Ct.
uunet!cimshop!davidm Mtn. View, CA 94043
====================================================================
"If someone thinks they know what I said, then I didn't say it!"