• The southern portion (between -10 and –34 degrees latitude) — data used is from World Vector Shoreline (WVS) data, a digital data file at a nominal scale of 1:250000. The dataset includes shoreline data and political boundaries only.
• The central portion (between +10 and –10degrees latitude) and above the light blue dividing line — data used is from World Data Bank II (WDBII) (designed for 1:2 000 000). You will note that this dataset also includes inland water resources. The dataset is dated (being compiled in 1970), and hence political boundaries are not necessarily correct.

The quickest way to see the differences in resolution between the two datasets will be to zoom deeply into a small portion of the East Coast at the light blue (-10deg lat) dividing line, and move up and down from this line. (You will observe that the end points do not join properly; however, all the borders and the West Coast lines join more accurately).

To appreciate the detail in the WVS-data, take deep zooms into the East coast which has a more ‘colourful’ coastline than West. Also into the Cape Town area (south eastern tip of the continent). I compared the shoreline data with 1:250 000data contained in a high quality Southern African Atlas. The dataset includes minute detail which has been left out of the maps and that I know to be correct; (either from personal knowledge of shorelines or by comparing the data with 1:100 000 maps). Offshore islands, often no greater than rock outcrops, are included in abundance.

The WDBII data appears to be exactly what it purports to be. (Regrettably I have already encountered dated African borders). One plus point is that borders are contained in separate data-files, which opens the possibility of adding borders only to the existing data. Another plus point is that inland water (lakes) are included, and separate river data is also available.

Finally – what shall we say about the original data plotted? (layers 52 & 53) Very rough indeed, but having the advantage of showing shapes in clear lines when the globe is viewed in its entirety. And of course, a fairly very light globe containing the entire world, to which detail can be added as desired.

Started in the same painfully slow manner as used in plotting of the original data. Gradually step by step refinements/shortcuts have been brought into the process. Finally on Tuesday I seriously tackled VBA coding in Excel macros and eventually succeeded in automating the whole process to a fine degree. The final plotting process enables me to plot about 4 lines per minute, and works as follows:

Step 1: Spreadsheet preparation
Copy a block of about 40 000 points into a spreadsheet and prepare the sheet for plotting — takes approx. 10 minutes.

Step 2: Plotting

(The above keystrokes can be performed in fast succession, and the only waiting is for the actual plot to complete).
Whilst human presence remains desirable to monitor the overall process, it is not believed much more can be done to automate and speed up the procedure.

Drawing Size
The original drawing: 2 792 448 bytes
The present drawing, including all additional lines: 9 463 808 bytes

The requirement for additional detail data will vary from individual to individual. I project that satisfying my own level of interest may result in a drawing of approx. 50 Mb. It is further estimated that adding the level of detail presently under consideration on a global scale will result in a drawing size of over 100 Mb.

The total drawing is still loading fast on my machine (which is a bit ancient now). Obviously the possibility always remains open to have separate drawings for great detail on a continent basis

Graphics Display

I used the little macro ‘Seasons.d3m’ to test the effect of the additional data on the globe’s ability to rotate reasonably freely. (My graphics card is very mediocre also!)

Time per revolution (all layers visible) 3 mins 20 secs.
Time per revolution (new layers invisible) 1 min 34 secs.

This (serious) slowing down is not causing too much concern. It is visualized that as the design develops, some macros can be introduced to assist in turning layers on and off as required in order to prevent a graphics display overload.

Expanding City / Country Listing
It is visualized that by introducing appropriate branching into a macro such as Locate.d3m, many separate Dat-files can be handled. In this manner many thousands of entries should be possible.

The data used over the last week is obtained from http://rimmer.ngdc.noaa.gov/coast/.

It is suggested that you visit this page and scroll down to display information about the data used and link to related pages for more detailed descriptions.

A brief extract is appended at the end of this document as Appendix 1.

What I particularly appreciate about this page is that extraction can be done on a small scale basis, and that in each instance an option exists for display of the line about to be extracted. I print this display block, which then acts as a convenient reference during plotting of data. There is nothing worse than plotting data here, there and all over the place without a mental picture of where one is heading! The prospect of downloading a total global data resource of (say) 75 Mb does not appeal to me — the organization required to handle such a vast amount of data is daunting.

It is a pity that data is extracted on a non-continuous basis. Particularly the WVS-data is extracted in a peculiar pattern of setting a starting point (say) 1 degree away from the area previously covered — then drawing little disconnected lines all over the place around this point and only many lines later joining them up to form a continuous line. I tried to capture something of this process in a series of screen images appended as Appendix 2.

With the WDBII-data I found it possible to page down the spreadsheet and spot matching end and starting points. These can then be joined up at that stage, prior to plotting. On average a reduction of approx. 30% in lines to be plotted could be achieved. It was also fairly easy to separate main coastline from islands and inland water resources, which enabled plotting in separate layers and assigning a pale blue colour to the latter. (A dark colour all over a continent appears to be overpowering).

Obviously I will keep my eyes open for other data sources, but have the feeling that they will all have their roots in the same sources.
 

It is believed that this question can only be answered by other users of DesignCAD. One has to establish the level of interest, and analyze the response. To this end, and subject to your permission and own feelings, I would suggest the following:

• Placing a globe together with macros and a text file in a single zip-file on your site. The globe should include a portion of the greater detail included in the present submission. The macros should be refined to some extent. The text file must include a structured request for response from interested parties ….

• The ‘structured’ request must establish level of interest, and in particular whether interested parties are prepared to assist with further development of a globe. I am prepared to act as coordinator, and with your permission the response from interested parties can be directly forwarded to myself.

• Further development of a globe can hopefully thereafter be done as a team effort. If we can only get half a dozen of users willing to assist, the task is not so formidable. More important is that we can get new ideas and fresh approaches, and in particular local knowledge of geographic areas covered.

• Depending on the level of interest, it can be decided whether it may be worth your while to host future extensions and developments, or whether interested parties must just carry on sharing additional information amongst each other by e-mail.

These data sets were converted into the MAPGEN "coast" format, an efficient binary format that compresses data by storing differences between adjacent coastline points. The WVS data was stored with sufficient precision to resolve 50 m changes along the ground, while the NOAA data was stored with sufficient precision to resolve 1 m changes along the ground
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

World Vector Shoreline
Original source of data: Defense Mapping Agency (DMA) (now National Imagery and Mapping Agency (NIMA .
World-wide coverage
Suitable for scales close to 250,000
Accuracy; requirement for this data is that 90% of all identifiable shoreline features be located within 500 meters (2.0mm at 1:250,000) circular error of their true geographic positions with respect to the preferred datum (WGS 84).
Horizontal Datum - World Geodetic System (WGS 84)
Vertical Datum - shoreline based on Mean High Water (MHW)
The WVS coastline data is available on NGDC's "Global Relief" CD-ROM, along with the now-ancient WDBII political bounds and rivers.
Note: The version of WVS used in the coastline extractor does not include all the features (such as lakes, rivers, bradwaters, glaciers, international boundaries, country names, etc.) that DMA has in their data set.

Here's some info about the original data, which explains why Central America yields duplicates:

SUMMARY: The World Vector Shoreline (WVS) is a digital data file at a nominal scale of 1:250000, containing the shorelines, international boundaries and country names of the world. The World Vector Shoreline is a standard US Defense Mapping Agency (DMA) product that has been designed for use in many applications. The WVS is divided into ten ocean basin area files. Together the ten files form a seamless world, with the exception of Central America, where there is an overlap between the Western North Atlantic file and the Eastern North Pacific File.

The main source material for the WVS was the DMA's Digital Landmass Blanking (DLMB) data which was derived primarily from the Joint Operations Graphics and coastal nautical charts produced by DMA. The DLMB data consists of a land/water flag file on a 3 by 3 arc-second interval grid. This raster data set was converted into vector form to create the WVS. For areas of the world not covered by the DLMB data (e.g. the Arctic and Antarctic), the shoreline was taken from the best available hard copy sources at a preferred scale of 1:250000. The WVS data are stored in chain-node format, and include tags to indicate the landside/waterside of the shoreline.

REFERENCE: E.A. Soluri and V.A. Woodson 1990 World Vector Shoreline. International Hydrographic Review, LXVII(1)

The next three pages contain 14 snapshots from the plotting process.

You will be able to spot the plotting area more or less dead centre and slightly to the upper end of the lower portion (below the light blue line) of the African continent.

Each snapshot just shows an additional line plotted, and is representative of what I refer to as a peculiar development of an area around freshly marked points.