Re: Looking for how to model 3D objects in 2D relational databases

JOG wrote:
> Mikito Harakiri wrote:
> > JOG wrote:
> > > Each Time(year, month, date) may be plotted as a point in a 3
> > > dimensional space, certainly making the entity 3D.
> >
> > The "year, month, date" part is a calendar artifact. It is indeed
> > formally 3 dimensional. Let me ask a related question though. Is
> >
> > 1024
> >
> > a 4 dimensional object? Because we can represent it as
> >
> > number of thousands number of hundreds ...
> > ------------------- ------------------ ...
> > 1 0 ...
>
> No, I totally agree. The term dimension is so abstract in terms of
> information modelling I struggle to define exactly what it means. It's
> often used as a synonym for "property of " an entity, which are rarely
> wholly independent, and often confused with the term as used with the 3
> physical dimensions, which are of course orthogonal.

Some possible glossary entries for "dimension" & perhaps also "multidimensional" ?

from mathematics, a tuple with n elements is said to have a dimension of n.
for numerical data visualization, n dimensional tuples can be plotted in an n-dimensional graph. For example, points in space can be graphed using three dimensions, such as x, y, and z axes.

from business intelligence online analytical processing (OLAP), a portion of a candidate key for a "fact table" serves as a key to a "dimension table". Each such key part is a dimension. Where there are two or more dimensions, the data is "multidimensional".

from data visualization, if a logical view (not necessarily SQL VIEW) of data maps directly to a visualization of that data in a fully populated (possibly including "nulls") table of single-valued row-column cells, that view might be classified as a two-dimensional view of data. In cases where some cells would be empty of any values or nulls from the view or where some cells hold multiple values, the data is "multidimensional".

------end of first shot at such entries in case mAsterdam can incoporate them into the cdt glossary

Anyone can feel free to fix up the wording or add other variations to the mix. I think the original question was how to take data that is n-dimensional from the first (mathematical) definition and use data visualization (the last definition) to view that data, possibly in order to then turn such a view into a relational view of the data (not necessarily the way to go about it, but that might be what was intended?)
--dawn Received on Wed Oct 12 2005 - 02:08:43 CEST