Kepler

1)When you drag an EML 2 Dataset actor to canvas and Configure Actor, you're presented with different parameters than if you drag onto canvas an EML 2 Dataset actor from a Data search result. The actor from the search has recordid, endpoint and namespace. The actor from the tree does not, and it has 4 semanticType fields. I think it's confusing to have the same actor present different params. I propose at least adding the missing 3 fields as options for the actor pulled dragged from the tree so a user may configure this manually to point at their remote dataset.

2) If you drag two datasources to canvas, both found in a search, and then try to change (eg swap) the recordid, endpoint, and namespace values, some issues occur:
2a) Selected Entity drop down does not refresh its values until you close and re-Configure Actor, initially giving the user the impression nothing has happened, as the output ports do not change.
2b) Selected Entity drop down maintains the last datasource's options (actually it looks like it duplicate the old option once too).

3) If you choose your own EML File and Data File on the EML 2 Dataset actor resulting from a search, the Selected Entity dropdown only contains options from the local data file (not the recordid still pointed at). If this is how we want the actor to work, when an EML File and Data File are configured recordid, endpoint and namespace params should be grayed out. Or Selected Entity should contain options from both local and remote datasets.
3a) If you then remove the EML FIle and Data File params, the actor does not fall back to the remote dataset, you are stuck with the ports and Selected Entity from the local dataset.