1. TABLE CAPACITY REQUIREMENTS.
If, as I have suggested, we set the size of our basic land units (BLUs) at about 1 sq. km., we will require about 8 million records. The actual size of Australia is about 7,700,000 sq. km. (including Tasmania) according to the SBS World Guide. The additional records allow for the fact that some of the BLUs will extend over the sea and the amount of land area covered will be somewhat less than the sum of the BLUs.
For the child records, we could guess at about 20 structures of interest in each BLU, with each structure being defined by 4 chains and each chain containing 5 data points (i.e. 20 data points per structure). This would require 160 million Structure records, 640 million Chain records and about 3.2 billion Data Point records. The bulkiest data items are the GPS co-ordinates. These will be discussed in more detail later, but one would imagine that about 128 characters each would suffice for a first estimate. The BLU table would require 4 sets of GPS co-ordinates, or 0.5 Kbyte in total. Allowing another 0.5 Kbyte (512 characters) for descriptive data, each BLU record will be 1 Kbyte, giving a table size of 8 Gbytes. The Structure records only contain pointers to data points, but there could be more elaborate descriptive data. If we allow 0.5 Kbytes per record, say, the size of the Structures table would approximate to 80 Gbytes. The same arguments apply to the Chain and Data Point tables, so allowing 0.5 kbytes per record in each, the storage requirements would be respectively 320 Gbytes and 1.6 TBytes respectively.
Indexing needs are very hard to evaluate and often add up to several times the amount of data actually stored, but when one notes that many household PCs have internal disk capacities of the order of 1 Tbyte (i.e. 1,000 Gbytes), to which can be added multiple external disks of 1 Tbyte each, it can be seen that the core database is of very modest proportions in the context of the work to be done. This is actually quite important, because when searches are conducted for project materials stored in a multiplicity of other databases and all the links are managed through the core database, its’ compactness and efficiency will have a significant impact on the analytical work in hand.
In the post on database design, some other tables were mentioned, listing the users and projects which were utilizing our database. It is suggested for reasons of security and confidentiality, this information should be stored in a separate database, linked to the core database through the various RID keys.
2. GPS RECORDS.
There are a number of GPS formats in use. Two of them use variants of the usual degree/minute/second co-ordinate system. A third, the Universal Transverse Mercator format, uses the grid lines drawn on maps. Being dependent upon external data, it is not appropriate for the current proposal. The two degree/minute/second formats vary only in that the first uses three separate numbers for each co-ordinate, while the second uses only two. These are degrees and minutes, with the seconds being incorporated as a decimal fraction in the minute value.
Because it is intended that the core database will be accessible to the maximum number of users, the version to be used should be that mounted in the largest number of GPS locator types and mobile telephone brands.
A GPS record consists of three data segments. The first consists of a number and a character string. These are respectively, the Waypoint Number and the Waypoint Name, which are identifiers for the GPS point. The second segment contains the letter N or S, indicating whether the point under consideration is north or south of the equator. It is followed by the latitude in degrees, minutes and seconds or degrees and decimal minutes. The third contains the letter W, indicating that the point is somewhere to the west of the datum line which runs north and south through the observatory in Greenwich, England. It is likewise followed by the longitude in the same format used for the latitude. The limits on the latitudes are 0 and 90 degrees, respectively north and south. The limits on the longitude are 0 and 360 degrees.
3. CORE DATA CAPTURE.
It is anticipated that the detailed physical topography will initially be obtained by processing satellite imagery. Such power of procedures may in the near future be significantly enhanced by the use of circular polarized light. This possibility was raised in a recent Catalyst programme on the ABC, which was examining the eyesight of various creatures which used this technology to view their surroundings in quite extraordinary detail. The commentator suggested that it should be possible to construct cameras which in effect mimic the actions of their eyes.
Places of particular interest or which are accorded a marginal status could be examined by flying over them with cameras and measuring instruments or by site visits. Back in the 1960s, I was employed on the design of a section of the W.A. Standard Gauge Railway Project and needed to identify catchment areas for drainage purposes. At that time, the area of interest had not been covered by the Ordnance Survey authorities, so we had to do our own assessment. We hired a helicopter and flew low over the ground until we could identify a high point as the boundary of the catchment area. We would land and determine the point’s height above sea level (in order to calculate the gradient of the water flow) from the helicopter’s altimeter. Once we had identified one point, we could follow the ridge around the perimeter of the basin, taking measurements as required. While crude, this procedure was perfectly adequate for our purpose and took surprisingly little time.