________________________________ From: Rick Fochtman <[email protected]> To: [email protected] Sent: Tue, June 15, 2010 3:04:06 PM Subject: Re: Developers' Use of Prod Data For Testing
----------------------------------<snip>------------------------------- One thing to remember is that production data may not always be the best test, as it may not contain all possible combinations of values in the various fields. Just because it is production does not mean it is comprehensive. ---------------------------------<unsnip>------------------------ Another drawback to using production data is this: all too often it's already gone through a pre-screening process where invalid data is filtered out. Such things as alphabetics in numeric fields, missing data, etc. Rick, I agree with you that is one point that I could never get across to our users. The only people that seemed to be into testing with data from outside learned early on editing was mandatory. For the most part our "foreign" people were on IBM MF's so the testing from them was pretty easy. Where it got to be fun was the PC (and other computer) manufacturers. I do vaguely remember some buzz going around about issues from these "Other" computer users as to how to handle "bad data". I *vaguely* remember that the handling decision went up the management ladder to the president and it came down (exactly how I do not remember as it was 15+ years ago). But it really did. Now we were (somewhat) unique as we were owned by the users, so the president was the person who dealt with them directly. I wish I could be more exact but it was not in my daily workings, so I only got tidbits most of the time from the management of the programming staff when we went out for drinks:) Ed ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
