Its good to review the concepts.
1) Definition of Feature: "abstraction of real world phenomena"
Reference: several OGC and ISO standards.
So Jon's comment is very accurate:
I think a "feature" is just a geographic "thing".
2) Definition of Coverage: "feature that acts as a function to return
values from its range for any direct position within its spatial,
temporal or spatiotemporal domain "
Reference: OGC Abstract Spec Topic 6 which is now also ISO 19123
You can get the coverage spec here:
http://portal.opengeospatial.org/files/?artifact_id=19820
So Jon's comment is loosely correct, depending upon what is meant by
data structure.
A "coverage" is a data structure.
But this statement is not correct:
So the relationship is that a Feature "has a" Coverage.
3) A coverage is a feature with some additional descriptions on the
way the feature attributes are modeled.
Which is why the WCS is valuable beyond what WFS provides.
WCS defines operations that take advantage of the descriptions of the
feature attributes in a coverage.
While recognizing that WFS can return features including coverages,
because a coverage is a feature.
For an excellent informal discussion of Coverages, see the
"introduction" in the Coverages spec (Topic 6 - URL above.)
George
