Geoscience Reference
In-Depth Information
Fig. 5
Initial dataset
download results
4.4 Comparing Two Approaches
Using the above approach to rendering and the detailed and generalised data-
sets, two separate mobile Apps were created, both of which have the entire city
of Sheffield as their starting viewpoint. The first App queries and downloaded
the detailed dataset into memory on the test device (known as the “Full Dataset”
test) and times both the download and interaction. The second App queries a gen-
eralised version of the dataset at the startup, and re-queries the data when the
user zooms and pans (i.e. requests data from the server each time, known as the
“Dynamic Dataset” test). The generalised dataset is queried for situations where
the horizontal extent on screen is greater than 250 m, with the detailed dataset
being requested for screen extents of 250 m or less. In both cases, time measure-
ments are taken at various stages through the process and written to the device as
a comma-separated values (CSV) file in order to permit performance comparisons.
5 Initial Results
In total ten time measurements were carried out for each of the two test datasets
and for each of the tasks (download of the initial dataset, zoom and pan). The
results were then averaged for each task. Figure
5
shows the resulting initial down-
load times and Fig.
6
showing the comparative zoom and pan times, with time
(in milli-seconds). Table
2
gives the results and Fig.
7
shows extracts rendered
datasets.
In Table
2
, the “Import XML” time refers to the time to convert the dataset into
XML from the database using PHP and download the dataset to the mobile device.
