|
I know this was kicked around but,
....I tested this issue a few years ago on a smaller machine with a small user base and I could not determine a measurable difference in performance with a subfile of 2500 records....
This I can believe, because when talking about subfiles, I believe the real I/O processes are in the operations that execute on the subfile control record, not the subfile. Meaning the Reads and the Readc to the subfile record are all done in memory. My preference has evolved into a preference for reserving a hidden field in the subfile to contain original values, and sometimes the entire original record, because it makes it easier to (1) check to see if the record is changed without worrying about SFLNXTCHG, (2) no database read is necessary to refresh the subfile (I use F5 for this), (3) it's easier to make a change like adding F19 and F20 key presses (or other) to go "left" and "right", even if it's necessary to implement using two or three different subfile formats, (4) easier to add in other values, (5) easier to link this subfile record (such as an invoice header) with records to pull up from another file (such as invoice detail lines), and so on. Disclaimer: I still mix and match. :-) And a technique for subfiles I picked up on from (Kevin Vandever?) involves SFLPAG=SFLSIZ and using data queues, which may even cost time for possible extra disk I/O, but as a user I like it better. Efficiency versus effectiveness: Accuracy and precision within user-acceptable parameters are number one of course, which relates to "effectiveness". But efficiency should always be a consideration, and like Joe Pluta said, it's always a number one concern for a company that sells software. Staring at a screen is no fun even for us programmers! This is *especially* true for interactive applications that involve a customer waiting at a counter or on the phone. And if it's an application served over the Web, you have to add all that time that you can't do anything about. And even internally, sometimes the company LAN slows down. And if you don't have to worry about any of that, it's still a very good habit to have. When it becomes a habit, you hardly notice the extra effort. Very rarely do we get requisition specs that make response time a priority, but in part that's because users don't realize how slow we can make their computers work! :-) They notice how a program is faster if you make a good change, but more often how slow it is! Just my 2c too... --Alan
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.