Some Layers will already have established zoom levels (like google or ve). Instead of trying to determine them and populate a resolutions Array with those values, we will hijack the resolution functionality here.
When you subclass FixedZoomLevels
The initResolutions() call gets nullified, meaning no resolutions array is set up. Which would be a big problem getResolution() in Layer, since it merely takes map.zoom and indexes into resolutions... but....
The getResolution() call is also overridden. Instead of using the resolutions array, we simply calculate the current resolution based on the current extent and the current map size. But how will we be able to calculate the current extent without knowing the resolution...?
The getExtent() function is also overridden. Instead of calculating extent based on the center point and the current resolution, we instead calculate the extent by getting the lonlats at the top-left and bottom-right by using the getLonLatFromViewPortPx() translation function, taken from the pixel locations (0,0) and the size of the map. But how will we be able to do lonlat-px translation without resolution....?
The getZoomForResolution() method is overridden. Instead of indexing into the resolutions array, we call OpenLayers.Layer.getExent(), passing in the desired resolution. With this extent, we then call getZoomForExtent()
Whenever you implement a layer using OpenLayers.Layer.FixedZoomLevels, it is your responsibility to provide the following three functions:
...those three functions should generally be provided by any reasonable API that you might be working from.