>Adding a blank record and letting the users fill in the fields,
>or, Adding a blank record and letting the users fill in the fields, and making the users SAVE/ABORT.
Robert,
The answer to this question depend on a lot of things. There is no one answer that holds for all circumstances. One of the issues is the ovrall design of the user interface, do all of the interactions with the user require that the user confrim their choices, or are the choices automatically confirmed unless the user says different?
In addition to the user interface issue there is the issue of how you are handling your data, are you using buffering, are you using scatter and gather without buffering, are you using updatable views for data entry? All of these questions will influence how to handle the addition of a new record.
It is true that INSERT INTO is faster on a network than APPEND BLANK and REPLACE, but in this instance you are dealing with data entry and the usr will take much longer to enter their data than the difference between these two approaches of adding record to a table. Also the user will NEVER notice the difference between append blank and insert into, unlkess it is being done in a loop for a large number of iterations.
So, in asnwer to your question, both of them are better and both of them are worse, depending on the circumstances.