Dr. Sumida - Mix N' Match Summary (in construction)

I'm sorry I had a very little time and still don't have good ideas to connect the proposed data sets with each other. However the following (temporal) comments of mine may be a hint to constructing compound data.

*In all the data sets except Geoffreys', data set are composed of 'entities' (such as main stem and branch) that have identification number, size or other numerical variables, properties, etc.

*Data sets seem to be classified into smaller categories.
First, we can classify them into SPATIAL and NON-SPATIAL data
(1) SPATIAL : x-y-z coordinates, branch azimuth, branch inclination, etc.
(2) NON-SPATIAL : properties showing nature of the entity such as species name, AGE, branch status (live/dead), etc. Some NON-SPATIAL data is accompanied by SPATIAL DATA (eg species name of a epiphyte and its position on a branch).

*SPATIAL data (1) can be classified into STRUCTURAL and NON-STRUCTURAL data.
(1-1) STRUCTURAL : a data set that has spatial connection with other data sets (e.g. x-y-z coordinates of branch base and that of branch tip). Hence it needs further description showing connections.
(1-2) NON-STRUCTURAL : a numerical data set that does not have spatial connection with others. Eg, Leaf area in a x-y-z voxel., branch density (# of branches per unit length of a trunk)

*STRUCTURAL data (1-1) can be classified into DIRECTION and MAGNITUDE data
(1-1-1) DIRECTION : a data set that describes a 3-D linear-structure as in Sumida's skeleton tree. Trunks and branches of Sumida's skeleton trees do not have diameter.
(1-1-2) MAGNITUDE : a data set that describes an entity's magnitude such as stem/branch diameter, crown width, crown length, foliage spread of a branch system, etc. which are often measured independently with other structural data but still shows some structures.

Probably it is not so difficult to connect MAGNITUDE data with DIRECTION data under several assumptions. I think a more important issue is how we can relate NON-STRUCTURAL data with STRUCTURAL data.

In my classification above, NON-STRUCTURAL spatial data are, in other words, PROBABILITY data (e.g. the probability of the presence of (the number of) branches at a specific position on a trunk is known, but their exact positions are unknown). In case of Geoffrey s', data, original measurement data may have x-y-z positions of foliage that can be 'visible' from his laser meter. In this sense it is not a probability data. However, another problem is that you cannot say which foliage detected by the laser belongs to which tree or which branch.

So….what should we do with them?