Birkel, Paul A.
pbirkel at mitre.org
Sat Jul 24 08:36:47 EDT 2010
Comments regarding OGC 10-034r3 (http://www.opengeospatial.org/standards/requests/69):
In Clause 3.1 the concept "local tangent plane" is ill-defined as it depends on the concept "Earth's surface" which is undefined. Recommend that the concept be tied to "a reference model of the shape of the Earth, such as an ellipsoid of revolution". Given how the term is used in this proposed standard the concept would more accurately be something like "a plane that is parallel to the local tangent plane to the WGS84 ellipsoid at a location determined by a WGS84 ellipsoid surface normal passing through that position". Given the intended scope of application of this proposed standard it would be appropriate to not assume that the position of the moving object is "close" to the reference ellipsoid. If that assumption is being made, and thus the scope of the "proper" application of this proposed standard is limited, then that scope should be clearly stated.
In Clause 5.3 the specification mixes the direction of the normal to the LTP with the direction of the "plumb line". By definition, on an ellipsoid of revolution while the "vertical axis [may be] taken to be straight down" this is still a different direction than the local gravity gradient. While this difference is small and may be relatively safely ignored with respect to the intended uses of this proposed standard it is important that the difference be made clear (and accurate) so that users of this proposed standard may, someday, be able to interoperate with more robust uses of GML and CRS's. A diagram would be useful in clarifying these differences. Judicious use of the terminology of ISO 19111 (and thus GML) such as "compound coordinate reference system" may be appropriate so that users of this proposed standard understand the relationship of its non-standard "coordinate system" specification to more rigorous systems (or at least that there is a relationship that is not trivial).
In Clause 5.3 the velocity is defined with respect to the LTP. The proposed standard, however, uses the concept of "speed" - which would be the projection of the instantaneous velocity onto a reference plane or axis. This is presumably intended to be the reference LTP, however this is not explicitly stated. The speed could equally validly be determined with respect to the vertical axis (e.g., the speed at which a person moves while riding in an elevator; in conventional usage would this be understood to be zero?). The intended reference surface or reference axis along which the speed is determined must be clearly specified in the proposed standard.
In Clause 5.4 the acceleration is defined as being in the direction of the velocity. While this is true for the acceleration or an object moving in a straight line it is not true in other cases (e.g., a car travelling along a highway curve at increasingly high velocity, which is necessarily being increasingly accelerated towards the center of the curve and not along the instantaneous direction of travel). Is this proposed standard specifically defining the scalarAcceleration (Clause 7) to be the instantaneous acceleration taken along the direction of travel (and thus misleading in the case of non-linear travel)? What is the intended reference surface or axis along which this scalarAcceleration is determined (see preceding comment regarding Clause 5.3)?
In Clause 7 the fragmentary expression "The supporting use case is a moving vehicle or person on a local tangent plane" appears in the documentation. As this is rarely true it might be better worded as "with respect to" rather than "on"; this revision would also be consistent with the definitions of the various measures.
In Clause 7, why is it necessary to restrict the uom for spend and heading to fixed values (e.g., a speed in miles-per-hour or feet-per-second is disallowed) whereas the uom for scalarAcceleration is left completely unconstrained (e.g., furlongs-per-hour-per-hour would be allowable)? Recommend that the proposed standard not constrain the uom for spend and heading to fixed values; this will allow for a wider range of practical application of the proposed standard. If it is intended to tie this proposed standard to fixed uom values then it should be done consistently for all measures, including the scalarAcceleration, which would logically (and consistently) be meter-per-second-per-second, and the elevation, which would logically (and consistently) be meter.
Dr. Paul A. Birkel (Rm. 5H229) 1-703-983-6399
The MITRE Corporation (M/S H505) FAX-1279
7515 Colshire Drive or-6435
McLean, VA 22102-7539 mailto:PBIRKEL at MITRE.ORG
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Requests