Re: Diference between 4GL and 4GL-RDS

In article <260idlINNqrq@emory.mathcs.emory.edu> alan@po.den.mmc.com (Alan Popiel) writes: > Database access is normally the bottleneck in an application. >In one test that we did, the RDS version of the application consistently >ran faster than the compiled-to-native-code version. (My theory, for >what it's worth: The RDS runner DBMS accesses could be more optimized RDS optmizes SQL access in two ways: 1] Whenever it sees one of the following statements, it prepares a statement cursor and re-executes it each time it is encountered again: insert, update, delete, alter, create procedure, create table, create view (and set debug file.) 2] Additionaly, RDS stores all 4GL program variables in the same machine independent format as is used to communicate with the back-end. In compiled 4GL, when accessing host variables, data conversions are performed to and from the SQL communication format to the local C format used to represent esql/c vars. RDS manages to save data conversions with each host variable access at the cost of increasing variable access for non-SQL manipulations. >because the runner would filter out abnormal conditions. The compiled >code had to include more error checking--in loops, etc.--because directly >compiled stuff could contain errors that the RDS runner would never >commit.) RDS does just as much error checking (if not more...) Hope that helps, Girard Chandler (informix 4gl development)