Search This Blog

Yesteryear

Friday, December 20, 2002

December 20, 2002


           5:30 a.m., Miami. Again today, I was reading and researching database, so you can skip this page if you hate computers. Mind you, there is always a twist if you pay attention, always something to learn. This picture of Paris Hilton added in 2016, by which time she was another rich nobody.
           There is still the odd nagging around the office that my database is inaccurate. I finally talked to Keith Mitchell about this, that the database is very accurate and it is merely portraying the inaccurate information that is on file at the office. I've been tinkering with how the table links can be used to follow an employee around without the usual delays of the system now and being. After 10 hours of deep thought I've only got an inchoate theory.

           Here's the problem: each department is keeping its own records and not necessarily communicating changes. Plus where those departments are using a computer many records have only one date, allow deletions, and have no audit trail. This is an ill thought-out list of employee categories similar to a university student list. That is, a person coming on their first day and someone just finishing a Ph.D. are both called “students”. Our records have "employees". These type of lists don't allow for different levels of involvement.
           Each department can enter an employee record without checking to see if that employee is active and where she or he is supposed to be. My favorite is people who pull a paper out of the cabinet, white out to make a change, then stuff the paper back thinking, "There, I've done my job." Of course, such a system is far too cumbersome to track the little things that help keep the company human.

           [Author's note: for those who've never studied database, my descriptions of certain problems can be an eye opener to those who wrongly blame computers for errors. For example were as I define changes as any activity which causes any part of a file to be different most people in conclude I mean additions, deletions and alterations. But in fact, I mean even saving a file in a different format or location is a change that should be recorded.
           Example: I mentioned above the error of a record having only one date. This seems to be an error that 100% of people make before they study database. I won't tell you the solution, but every record needs at least two dates. Otherwise, somebody can enter a change today that alters a record made six months ago, rendering all subsequent records and decisions based on that original now null and void.
           In the end my database was never adopted. The reason for it was that unless I was around there was nobody to keep it current. There is also a factor that I suspect certain parties like the old system where it was too difficult to keep track of corruption. Like who got twice as many of the free football tickets as anyone else last month, or which employee got 12 company T-shirts from six different sites.]


++++++++++++++++++++++++++
Return Home
++++++++++++++++++++++++++