[QUANTITY] My Summary of the "arguments"/transform discussion/thread (Was: Re: [QUANTITY] Quantity "arguments")

[Brian Thomas]

On Friday 14 November 2003 01:11 pm, David Berry wrote:
> Pat,
>
> > While I'm sure all this is very interesting, it seems way off topic to
> > me. I am sure that's why only two people are left standing.
>
> I'm not sure why it is "way off topic". I thought "the topic" was what
> sort of Quantity we want and what we want to use it for, which is what Ed
> and I have been discussing. I think I may now be able to see what the
> basic difference is between out two models, so I at least have profited
> from the discussion, and I hope Ed and other have too.

	Well, I guess "Ed" means me too :) Here's what I have observed (summary)
	of the discussion of multi-dimensional quantities/transforms:

	Both models are not that different. In fact, it boils down to semantic and
	minor structural differences. The commonalities in approach (which look
	like requirements) are:

	a. Quantities may have dimensions. The values within a multi-dimensional quantity
	are uniform in terms of data type and units. Accuracy may still vary between 
	values.

	b. Multi-dimensional quantities have axes. Axes are described as a set of
	values with associated units, accuracy.

	c. A set of axes help to define (but may not be a complete description) of
	a "frame of reference" or "frame".

	d. You need to allow for alternative axes, and alternative frames of reference. 
	By extension then of point b, the alternative axes may be specified by a "frame".

	e. You need to have 2 frames inside something like a "frameset" in order to do
	a transform between frames.

	f. You need to have the ability to transform both the values (and units) of
	a quantity as well as the axes values (and units). Sometimes you want to transform
	just the values, sometimes just the axes that describe these values, and sometimes
	you want to transform both the values, and the axes that describe them. 
	  
         g. Transforms should be possible not only between coordinate systems (e.g
	sky, pixels, etc) but also for non-positional systems like magnitude to flux.

	h. Values in the quantity or on the axes can either be actual numbers, or 
	generated by an algorithm. And, I believe, we agree that this algorithm is likely 
	to be described by a frameset, or something similar, that operates on the 
	underlying "list index" which is a simple index of the values in the quantity/axis.

	These are broad points, and there are a number of more specific, smaller, points
	that we also agree on, but I'll leave aside for the time being.

	Ed, David, have you seen any other commonalities between the models? Any
	disagreement on this agreement?


	Regards,

	-b.t.


-- 

  * Dr. Brian Thomas 

  * Code 630.1 
  * Goddard Space Flight Center NASA

  *   fax: (301) 286-1775
  * phone: (301) 286-6128