standard) have tackled the issue of transferring the data
schema gs well as the data. If taken up and successfully
implemented, this approach would represent a significant
step forward. In the update context, it would potentially
make detached update more viable, as transfer of the
data model (provided the update process can accept and
use it) should allow the update process to update the
data in terms of the model, performing the necessary
integrity checks before returning the updated data to the
master database.
4.2 Data-centric v. Process-centric
Alongside the progress towards "data-centric" transfer
standards cited above, there is growing move towards
"process-centric" standards. The motivation is to provide
mechanisms for access to data where it sits, in its "local"
format. Powerful stimulus for this approach is coming
from the Internet community and the wider IT industry.
The debate in the formal standardisation bodies (ISO
TC211, CEN TC287) is still in progress, and the outcome
is presently unclear, although the tide is moving in the
process-centric direction.
The same trend is apparent in the Update process, for
the reasons stated in this paper. The need is for the data
model to be accessed via the Application Programme
Interface (API) ofthe spatial database. Object-orientation
can provide a very clean interface at this level,
particularly in terms of maintaining spatial database
integrity, which is the increasingly dominant concern.
The parallels between the activity of receiving and
managing update information and the activity of issuing
data and updates are apparent in the trends in
technology and standardisation.
5. CONCLUSIONS
These trends are rapidly leading to a point at which the
ability of update processes such as soft copy
photogrammetry to deliver coordinate information and
thematic interpretations is increasingly standardised, and
in effect determined by the nature (resolution etc) of the
available imagery rather than the processing available in
the update tools. The utility of the update tools will rather
be determined by their abilities to interact with
increasingly complex data models, either in a detached
mode via transfer formats that carry data model as well
as data or in direct mode via the Open APIs of spatial
databases with powerful data modelling capabilities.
Increasingly these are following the Object-Oriented
paradigm.
Within this framework, progress can also be anticipated
on spatial generalisation, in the short term by
orchestrating the posting of updates between databases
supporting different scale ranges. As the powerful
conceptual framework of Object-Orientation provides
more usable spatial generalisation functionality,
increasingly scale-independent databases will become
possible (e.g. across a band of scales), but update of
958
these will involve interactions with the underlying data
models. These models must include provisions for
metadata, and some appropriate mechanisms for
recovery of history. The issues are trade-offs of system
cost against data utility.
As spatial databases cover wider areas and have richer
content, so the requirement increases to issue data only
to those entitled to it, to issue changes in a timely
manner and to validate both content and updates. There
is a growing need to devise means of managing change
only (or incremental) update. Just as the balance has
shifted from primary data capture to the implementation
of updates to data already held, so it will shift further to
the issuing of updates to the users or customers for the
data. There are underlying common factors to both the
receipt and the issuing of change information. A key
concept is that of unique object-ids. Object-Oriented data
models and database versioning provide powerful
techniques for Spatial Database Update - a Key to
Effective Automation.
References from Books:
Taylor, David A. (1990). Object-Oriented Technology: A
Manager's Guide. Addison-Wesley Publishing Company.
References from Other Sources:
Ibbs, Tony J. (1995). Progress in the standardisation of
geographic data - a U.K., European and International
Perspective. IUSM Working Group on GIS/LIS,
Hannover, September 25-28, 1995.
Woodsford, Peter A. (1995). The Significance of Object-
Orientation for GIS. IUSM Working Group on GIS/LIS,
Hannover, September 25-28, 1995.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996