BoundingBox use in WCS

Hello All,

A collegue and I are trying to clarify the use of the BBOX in both 1.0 and 1.1.

Given an image with the following spatial information:
Origin = (-180.004167,65.004167)
Lower Right ( 180.0041667, -65.0041667)
Pixel Size = (0.00833333,-0.00833333)
If I make a request for a bbox=-180,50,-160,65 and resx=0.00833333 and resy=0.00833333 and CRS=EPSG:4326
What should the spatial information for the returned image be and which pixels should be included?

Since -180 is partway into the far left pixel should the entire pixel be returned or should I only get the next to far left pixel?
Specifically, should the returned coverage have a left coordinate of -180.004167, -180.0, or 179.99583367?

Under the terms of the spec should this or should it not lead to resampling?


-- john