<ji********@com pumarc.com> wrote in message
news:11******** **************@ f14g2000cwb.goo glegroups.com.. .
So is the thinking that the client would be bending things by asking
that this situation be handled.
James A. Fortune
In some cases, absolutely. For example, I have actually had persons before
wonder how come their attempts to enter data crashed when they tried to
enter the state in the Zip code. They enter some names as "O'Brien" and
others as "OBrien" and still others as "O Brien"--only to then wonder why
not all of those are sorted together as if they're the same. They wonder how
come a query they later add themselves for showing "O'Brien" returns
inaccurate numbers because of the other variations entered.
Duh! You have to be consistent, enter them all the same way.
Then of course you have drop-down lists for states commonly, because some
will enter "NC" some "N.C." some "nc" and on and on. You have to be
consistent, else search results and the like will be off. You wouldn't
expect me to design the query to say "NC" or "nc" or "n.c." or "N C" or "Nc"
and on & on, would you? I would surely hope not.
Simply enough--with data, you have to have consistency with such things. It
would be ridiculous to ask a designer to allow for these variations. What I
will do instead--and others apparently do likewise--is design the database
to FORCE the user to enter all such entries identically. Not sure how one
would do this with the O'Brien issue--unless they made it unable to accept
the ' character or spaces of any kind--but with states, again, commonly a
drop-down list of states would be attached to the "state" field, with the
"limit to list" attribute set to "yes."
To me, this is such a situation. If you have some who would call themselves
"Smith Barney," others who would call themselves "Smith-Barney" others who
would have some other variation--you end up complicating things. If you
search for those whose true last name is "Barney," for instance, you may
pull up "Smith-Barney" in there as well undesirably--if you are doing one of
those "Like "*" And [name] and "*" type of searches.
Simply enough, it adds complexity where it isn't needed. I think that is
why, for instance, I hear that when my brother in law who I shall call "John
George Smith"--we likes to go by George rather than John--he has to enter
forms "John G Smith," because they ask for his middle initial ONLY, they do
not care what his middle name is, even though that's what he goes by. To
MAKE SURE he gets found in important searches, they make him along with
everyone else enter his first name in full, his middle initial ONLY, then
his last name--period. No exceptions, whatsoever. I would guess if his last
name was "Jones-Barney" they would have him fill out the form "JonesBarne y"
and that's how it would get entered.
I like such designs, personally. And I figure--I don't know, but if people
like "George" can be understanding of such realities with government forms &
the databases that hold the data, why should we do any differently? The
conformity is a GOOD thing--increases the probability of the data being
obtained properly and quickly, and helps prevents "cross contamination" of
data with similar names. In other words--data integrity. Isn't that the
point? If overly-sensitive types are wanting to compromise this to save
their sensibilities, I say they need a reality check.
In closing--I typically post here with extremely practical questions, not
opiniated ones. I typically ask questions like "what syntax is appropriate
for creating a joint query in SQL." I will make sure to stick to such
questions in the future for the most part and hopefully not waste anyone's
time here; I realize Microsoft MVPs and the like are hanging out here to
help & don't want to get sucked into long-winded opinion discussions. I am
asking as a designer, not trying to be opiniated politically so much as that
the objective here is the protection of data integrity, not sucking up to
people's whimsical fascination with titles & so forth.
LRH
