Difference between revisions of "Datatypes R2 Issue 20"
| (2 intermediate revisions by 2 users not shown) | |||
| Line 3: | Line 3: | ||
== Introduction == | == Introduction == | ||
| + | The cardinality of a Set is the number of items in the set. | ||
| + | While it's not clear exactly what this might mean with a | ||
| + | SET<REAL> or SET<TS>, it's clear what it means for SET<INT>, | ||
| + | SET<AD>, etc. | ||
| + | |||
| + | IVL is a specialisation of a set. So an interval of 3..5 is | ||
| + | a Set of 3,4,5. Therefore the cardinality of the interval | ||
| + | 3..5 is 3. And the cardinality of an unspecified interval | ||
| + | is unknown. | ||
| + | |||
| + | However in the Static Models, the cardinality of interval | ||
| + | is marked as 0..1. | ||
? backward compatible. | ? backward compatible. | ||
| Line 8: | Line 20: | ||
== Discussion == | == Discussion == | ||
| + | So, something is wrong here. The cardinality value for | ||
| + | interval should be 0..* everywhere, rather than 0..1. | ||
| + | |||
| + | :Yes, the MDF use of "cardinality" is wrong. [[User:Gschadow|Gschadow]] 18:14, 5 May 2007 (CDT) | ||
| + | |||
| + | An alternative view is that IVL is not a specialisation | ||
| + | of SET ;-) | ||
| + | |||
| + | :You can't make one wrong right with another wrong. [[User:Gschadow|Gschadow]] 18:14, 5 May 2007 (CDT) | ||
| + | |||
| + | This problem would be overtaken by the proposal from | ||
| + | Lloyd and Grahame to change the way cardinality is | ||
| + | assigned to attributes. | ||
| + | |||
| + | :Yes, that's the issue indeed. [[User:Gschadow|Gschadow]] 18:14, 5 May 2007 (CDT) | ||
| + | |||
| + | Furthermore, cardinality of IVL<REAL> is well understood. It is the infinity aleph-1. Conversely the "*" in "0..*" is the infinity aleph-0 (of countable items) ..., nay, rather "*" means "unbounded" (no set bound), never actually infinite. But IVL<REAL> where high > low is always infinity aleph-1. [[User:Gschadow|Gschadow]] 18:14, 5 May 2007 (CDT) | ||
| + | |||
| + | == Disposition == | ||
| + | |||
| + | This is closed; you can't fix a problem in the HDF by mucking around with the datatypes. | ||
== Links == | == Links == | ||
| + | |||
Back to [[Data Types R2 issues]] | Back to [[Data Types R2 issues]] | ||
Latest revision as of 04:58, 15 June 2007
Contents
Data Types Issue 20: Cardinality of IVL<T>
Introduction
The cardinality of a Set is the number of items in the set. While it's not clear exactly what this might mean with a SET<REAL> or SET<TS>, it's clear what it means for SET<INT>, SET<AD>, etc.
IVL is a specialisation of a set. So an interval of 3..5 is a Set of 3,4,5. Therefore the cardinality of the interval 3..5 is 3. And the cardinality of an unspecified interval is unknown.
However in the Static Models, the cardinality of interval is marked as 0..1.
? backward compatible.
Discussion
So, something is wrong here. The cardinality value for interval should be 0..* everywhere, rather than 0..1.
- Yes, the MDF use of "cardinality" is wrong. Gschadow 18:14, 5 May 2007 (CDT)
An alternative view is that IVL is not a specialisation of SET ;-)
- You can't make one wrong right with another wrong. Gschadow 18:14, 5 May 2007 (CDT)
This problem would be overtaken by the proposal from Lloyd and Grahame to change the way cardinality is assigned to attributes.
- Yes, that's the issue indeed. Gschadow 18:14, 5 May 2007 (CDT)
Furthermore, cardinality of IVL<REAL> is well understood. It is the infinity aleph-1. Conversely the "*" in "0..*" is the infinity aleph-0 (of countable items) ..., nay, rather "*" means "unbounded" (no set bound), never actually infinite. But IVL<REAL> where high > low is always infinity aleph-1. Gschadow 18:14, 5 May 2007 (CDT)
Disposition
This is closed; you can't fix a problem in the HDF by mucking around with the datatypes.
Links
Back to Data Types R2 issues
