2D scans

[bruceravel]

In #16, a conversation happened about extending XDI to 2D scans. There was some agreement that it would be a good idea to define a 2D extension, but the consensus was that it requires discussion and possibly different versioning.

To summarize:

1. An ascii representation of a 2D scan (a simple map, a sequence of XAS scans, etc) is of value
2. We must be mindful that ascii is not really appropriate for a very large volume measurement or for high dimensionality. Other, non-ascii formats are available for that.
3. XDI can be readily extended to accommodate 2D data with a few additional defined headers and some way of denoting breaks in the data table. In Target programming languages #16, @newville suggested one possible way of denoting breaks.
4. Documentation and clear test cases are required.
5. In Target programming languages #16, @Yohko pointed out that a 2D scan might not have some of the headers specified as required by the XDI specification.

In the end, it was suggested that 2D scans be tabled for the XDI 1.0 specification pending further discussion.