Use Case Extract - Extending the Snapshot Concept .1 Introduction

Copyright © 2011 Open Geospatial Consortium, Inc. All Rights Reserved. 75 the dynamic aspect of feature properties for filtering into account was developed in the SAA Pilot and OWS-8 and is documented in [OGC 11-073]. As the properties of an AIXM feature can change their value during the feature lifetime, a user may also need to consider the feature state over a period of time. The evaluateDuring operation thus allows for both the provision of a time instant and time period as parameter for determining which feature data is relevant for filtering. In case that a time instant is provided the Event Service can retrieve a Snapshot of the relevant AIXM feature. Otherwise, it needs to extract the feature information that is relevant for the given period of time. An “Extract” thus is the equivalent of a Snapshot, just that it provides complete information on the state of a feature for a given period of time, not a point in time. The feature timeslices - Baselines, Permdeltas and Tempdeltas - contain this information. A subset of these timeslices usually already suffices to represent the requested information. Thus it is unnecessary to retrieve the complete feature information from the authoritative data store. In addition to the “Eventing” use case just described, another use of an “Extract” may be caching. A client may want to explore the state of an AIXM feature for given point in time but the application already automatically retrieves feature state information for adjacent periods of time in a background process. This would speed up browsing through the data and is similar to mapping clients which cache map tiles that are adjacent to the current view as well as zoom level. An approach for retrieving an AIXM feature with just the desired subset of timeslices is documented in [OGC 11-073]. The following sections describe which timeslices are needed for an Extract and the options for encoding it.

10.1.3.3 Computation

An Extract is computed by identifying the timeslices of an AIXM feature that are relevant for the time of interest assigned to the Extract, i.e. the time period for which the state of an AIXM feature shall be determined. It is tempting to say that only those timeslices are relevant that are not before and not after the time of interest. In addition, one might think that only Baselines and Tempdeltas are of interest. Although both assumptions cover most cases, they are not entirely sufficient. Consider the following figure. 76 Copyright © 2011 Open Geospatial Consortium, Inc. All Rights Reserved. Figure 26 – State of an AIXM feature defined via its timeslices and time of interest of an Extract of that feature The Temporality Model allows that Permdeltas are not accompanied by an according Baseline - see section 10.1.1. In Figure 26, Permdelta two and four fall into this category. They permanently change the values of properties P2 and P3 respectively. Permdeltas therefore need to be taken into account as well for computing an Extract. The following list defines which timeslices are relevant for an Extract: 1. Baselines which are not before and not after the time of interest of the Extract are relevant. 2. Permdeltas which are not after the time of interest of the Extract are relevant, unless: a. they have a valid time which equals the start of the valid time of one of the Baselines identified in the previous step in that case, the Baseline incorporates the state change that such a Permdelta represents b. a Baseline exists whose valid time is not before or after the start of the time of interest of the Extract and the valid time of the Permdelta is before the valid time of that Baseline the Baseline overrides or incorporates all state changes that such Permdeltas represent