Re: [RC] SADDLE "FIT" DATABASE strategy, for the record. - Robert Ferrand
Mike,
Your point is well taken, however, I would like to
point out your comment "the Garbage In - Garbage Out data model I routinely
have to fix for my clients". makes my point: that at the beginning of this
research we need to allow equestrians, "free-form text" to express their
understanding, and it can be fixed later. (which is your stock and trade
) If we lock them into a "matrix" which can be easily queried, we are
prejudicing the results from the beginning. If you have a better idea for a
design, I am all ears, however, please understand what I am trying to
do.
From my research so far, I see the
equestrian universe as a multitude of galaxies within which are a multitude
of solar systems, each revolving around a trainer, a breed, a philosophy, or
whatever, the challenge is: each of the solar systems really do not
communicate very well to each other. Case in point, Dressage riders have little
discussion with team penners. Some endurance riders may do hunt seat, but it is
not common. Therefore, in building this database, we are trying to see if there
is any commonality, in how different equestrian groups approach saddle fit and
whether they have any real understanding of the physiology and physics of saddle
fit. This is not an easy task with some many different views.
Now, that said: once we have a broader
understanding of what the equestrians universe is actually believing
about saddle fit, then we can focus on building a more easily queried database
to address specific issues. From the responses that we have already received,
I do not see any reason to restrict the data so early in the process.
Employing a "free form text" allows people to speak their mind, which yields
some interesting insight. As a database designer, you must appreciate that
once the data is acquired, it can be rearranged. First things first, let's
acquire the data, without prejudice.
You never know what develops, once you begin to
measure. What you may think is happening is not, but something else develops.
When I began measuring the interface pressures under saddles, ten years ago, I
had no idea that the vast majority of saddles "bridged". I would have thought
out of random selection a higher percentage would have fit. It did not make any
sense to me, at the time. Even when we developed the gauge to measure the
angles and arcs to describe the polyform shape of the horses back and
corresponding saddle, (so the data could be put into a database and queried)
I did not appreciate the profound effect that the rider's weight had on
the horse's back. Even as recently as two years ago, now that we have begun to
measure the shape of the horse's scapulae and build structures to accommodate
that shape, we are discovering the horse's moving differently. You can not
improve that which you do not measure.
So my point is, with due respect to all the
database managers out there, I do understand how to measure and at this point I
prefer to continue to allow equestrians to input their thoughts through
"free-form" text, even if I am building a database managers nightmare. Look
on the bright side, this could provide you a large "data model you routinely
have to fix", but who knows what will actually come out of this research. I
am already amazed as some of the input and it has not even been up a week. So I
will keep an open mind and will appreciate any suggestions you may have to
improve this data collection.
Subject: RE: SADDLE "FIT" DATABASE
strategy, for the record.
>> From: Robert Ferrand
[mailto:sadddle@xxxxxxxxxxxxx] Your point that the entry form is improperly designed to
provide statistically relevant data because all fields are "free-form entry"
is not correct. Currently, we are making a WIDE sweep of the ENTIRE EQUESTRIAN
COMMUNITY - WORLD WIDE to better understand what the equestrian community,
which comprises 80+ breeds of horses, and as many cross breeds, in a multitude
different disciplines, with an undefined number of different types of saddles,
so that we all can better understand what equestrians actually believe about
saddle fit. You are quite correct, this increases the complexity of querying
the database, (hence Oracle), however, I think this is a better approach AT
THE BEGINNING OF THE RESEARCH, to let the equestrian community define the
parameters, not the saint at the computer. To my knowledge, NO company,
institution, or individual has ever attempted an effort on this scale. As the
data comes in, we can easily alter the parameters accordingly. So bring them
on.<<
Unfortunately, your expertise and passion is in
the field of saddle fit measurement, not database design, and that was
the whole point of my response. You chose instead to mount a soapbox and
try to convince us that not only are you altruistic but that there is no
linkage between your "research" and your business, and that your research
design was correct.
I have a different
opinion. I can clearly see such a linkage, AND that isn't why I'm
responding now. I'm responding, once again, to your belief that your
research design is appropriate. With 15+ years as a database architect
working for the largest of companies, I can tell you flat out, without
hesitation it isn't...you will store data that CANNOT be queried for
VALID statistical measures. You have created the classic "Garbage In -
Garbage Out" data model that I routinely have had to fix for my
clients.
