82
P. Wright
for analysis it had to be recoded as a series of real world objects. Linework broken for
text and other features had to be made up and polygons constructed. Tables of
attributes had to be assembled from the captions, legends and notes shown on the
chart.
The topographic data used in the trial was derived from a more diverse range
of sources. The map outline and woodlands were digitised from repromat. Contours,
rivers, roads and urban areas were derived from existing digital data. The topographic
data required less reprocessing than the hydrographic data, because in effect it was
already separated into layers. But it was deficient in various other respects. The
rivers, woodlands and outline datasets contained no feature codes or attributes.
Linework in the rivers dataset was discontinuous, and contours were broken where
form lines and other features were present on the paper map from which the data was
derived.
The geometry of some of the road junctions and roundabouts was too
complex to be displayed at a scale of 1:25,000, the scale of the prototype digital map.
Having created or reprocessed the data in each of the eight individual sets of
data, these then had to be merged to form a single multi-layered dataset.
This
involved clipping and erasing surplus data from each of the datasets. Data landward
of the Mean High Water Springs mark was taken from the topographic datasets of the
OS and data seaward of the Mean High Water Springs mark from the hydrographic
datasets of the HO. All positions were transformed to a common reference system, in
this case the UK National Grid. The datasets were registered one with another and
any apparent mismatches in the data were investigated. Although it was recognised
that there would be benefits in referencing all heights and depths to a single datum,
this was not done during the trial. In accordance with the standard practices of the OS
and HO, all heights above Mean High Water Springs were referenced to Mean Sea
Level (Ordnance Datum Newlyn) and all depths and drying heights below Mean High
Water Springs to Chart Datum, the level of the Lowest Astronomical Tide. One of the
biggest difficulties arose because the datasets were not based on a common data model.
Some of the feature codes were not unique and the datasets could only be merged if
the tables of attributes were identical.
Although raster data was not used in the trial, some consideration was given
to the processes that would be involved in combining this type of data for use in GIS.
These processes included clipping and erasing any surplus data, transforming the data
to a common projection, spheroid, datum and scale, edge matching adjoining datasets
and converting all the data to a common format. Apart from the limited use that can
be made of raster data in GIS, one of the main drawbacks of this type of data is that if
it is necessary to enlarge or reduce the scale of adjoining datasets in order to combine
these into a single map or chart, all text, numerals and symbols are enlarged or
reduced in the same proportions. A user panning about the resulting image of the
map or chart may easily gain the false impression that the scale of the data varies from
one part of the map to another.
P. Wright
for analysis it had to be recoded as a series of real world objects. Linework broken for
text and other features had to be made up and polygons constructed. Tables of
attributes had to be assembled from the captions, legends and notes shown on the
chart.
The topographic data used in the trial was derived from a more diverse range
of sources. The map outline and woodlands were digitised from repromat. Contours,
rivers, roads and urban areas were derived from existing digital data. The topographic
data required less reprocessing than the hydrographic data, because in effect it was
already separated into layers. But it was deficient in various other respects. The
rivers, woodlands and outline datasets contained no feature codes or attributes.
Linework in the rivers dataset was discontinuous, and contours were broken where
form lines and other features were present on the paper map from which the data was
derived.
The geometry of some of the road junctions and roundabouts was too
complex to be displayed at a scale of 1:25,000, the scale of the prototype digital map.
Having created or reprocessed the data in each of the eight individual sets of
data, these then had to be merged to form a single multi-layered dataset.
This
involved clipping and erasing surplus data from each of the datasets. Data landward
of the Mean High Water Springs mark was taken from the topographic datasets of the
OS and data seaward of the Mean High Water Springs mark from the hydrographic
datasets of the HO. All positions were transformed to a common reference system, in
this case the UK National Grid. The datasets were registered one with another and
any apparent mismatches in the data were investigated. Although it was recognised
that there would be benefits in referencing all heights and depths to a single datum,
this was not done during the trial. In accordance with the standard practices of the OS
and HO, all heights above Mean High Water Springs were referenced to Mean Sea
Level (Ordnance Datum Newlyn) and all depths and drying heights below Mean High
Water Springs to Chart Datum, the level of the Lowest Astronomical Tide. One of the
biggest difficulties arose because the datasets were not based on a common data model.
Some of the feature codes were not unique and the datasets could only be merged if
the tables of attributes were identical.
Although raster data was not used in the trial, some consideration was given
to the processes that would be involved in combining this type of data for use in GIS.
These processes included clipping and erasing any surplus data, transforming the data
to a common projection, spheroid, datum and scale, edge matching adjoining datasets
and converting all the data to a common format. Apart from the limited use that can
be made of raster data in GIS, one of the main drawbacks of this type of data is that if
it is necessary to enlarge or reduce the scale of adjoining datasets in order to combine
these into a single map or chart, all text, numerals and symbols are enlarged or
reduced in the same proportions. A user panning about the resulting image of the
map or chart may easily gain the false impression that the scale of the data varies from
one part of the map to another.