<< So, would anyone have advice regarding this, especially with regard to my basic design and/or your experiences having done something similar? >>
Some time ago I did a "proof of concept" of a fake R> prompt in which the user would type in an SQL statement in a memo field, the statement would be checked, executed, and stored, and the results displayed in a larger memo field below. It functioned very much like the regular R> prompt except that a statement that took some time to execute (like a long, slow SELECT) would not display ANY results until execution was finished (output was sent to a file, then the file was loaded into the memo field). The only way in which it differs from what you layed out is that I didn't include a separate keyword table, but I did provide for the username to be stored with the statement so people could have private SQL libraries. The purpose was three-fold: to stop users from executing certain statements (eg "vSQL LIKE 'DEL%'"), to log all SQL executed to find certain miscreants, and to begin to build a library of common SQL to help us determine what functions could most productively be automated. I never put it into production although I have a feeling it's going to be revived sometime soon. If it does get done, I can probably share the work. -- Larry
