David said:
If there are enough 255-character fields populated with data to
exceed the record limit, I'd call it a design error, even if,
technically, it's not a denormalized structure.
Two of the main tables I work with on a day to day basis have a lot of
fields. They are for tracking quotes and jobs. The job costing table
has about 109 essentially unrelated fields that are tracked for each
job. So far, I have been able to avoid using one-to-one relationships
with additional tables by carefully managing how many characters are
allowed into the text fields (about 45 fields). Right now the character
limits are only enforced at the form level (the only place users are
allowed to enter data), but eventually I will enforce the limits with
table constraints in addition to the form constraints just to be safe.
I didn't know originally that the main growth of the tables joined to
those tables would be through division of departments into more
departments, then into tracking the subdepartments under those
departments. Naturally both the quotes and jobs need to have the same
departments and subdepartments so that proper feedback concerning
profitability can be obtained for each subdepartment relative to the
original quote. Even time tickets are broken down by job and subdepartment.
Although seemingly innocuous at first, tying the profitability of
individual jobs to the people managing them and working on them has
completely revolutionized the criteria used for performance evaluations.
Also, the people who prepare new quotes can evaluate the performance
of similar quotes to see if their subdepartment estimates were accurate.
My point is that it is possible to have quite a few fields in a
normalized table and that the record level byte limit can still be
managed. I recognize that keeping the number of fields in a table to a
minimum is desirable for several reasons, including the record level
byte limit. Given the number of badly designed schemata out there I
think that the rule-of-thumb warnings about too many fields are
necessary, but they are not all encompassing.
Some of the reasons for the record level byte limit were discussed here:
http://groups.google.com/group/comp.databases.ms-access/browse_frm/thread/6566e7c66ff4f4c1
James A. Fortune
(e-mail address removed)
It's possible that the movie "Pearl Harbor," which came out in May of
2001, was partly to blame for 9/11. Besides showing examples of
kamikaze attacks, the movie ended with a patriotic statement about the
U.S. growing stronger and the enemy growing weaker. Naturally, that
statement, taken the opposite way, may have given encouragement to
someone watching the movie and thinking about carrying out such a plan.