The SEDRIS Data Representation Model
APPENDIX A - Classes
Model

Class Name: Model

Superclass - <SEDRIS Abstract Base>

Subclasses

This DRM class is concrete and has no subclasses.

Definition

An instance of this DRM class specifies a representation of some environmental entity as a feature representation, a geometric representation, or both. This representation is usually a "generic" representation that can be referenced many times in a transmittal to create many instances of representations of similar environmental entities.

The special case of the "null model" is the case in which both the feature and the geometric representation of the <Model> are empty - that is, they contain no primitives. This is instanced in cases where some state or condition of a representation exists but has no primitives, such as a representation of an environmental entity that has been completely destroyed, or that is out of viewing range.

Primary Page in DRM Diagram:

Secondary Pages in DRM Diagram:

Example

  1. The lowest level of detail of a tank's turret.
  2. A 1 degree by 1 degree tile of terrain containing thousands of instances to other <Models>.
  3. An aircraft carrier that has both a geometric representation and a feature representation.
  4. A data provider has an overall model (call it "car") made up of several components: top, two sides, four tires, back, front, underneath. When put into this data provider's < Model Library>, each of these components (as well as the overall "car" placeholder) is represented as an instance of <Model>. The data provider's organization has a place in its database where "car" is instanced, so that at an IG the "car" appears. This is represented in the resulting SEDRIS transmittal by a < Geometry Model Instance> of "car" appearing in the scope of an <Environment Root>. No other <Models> in this data provider's mapping to SEDRIS can reference the "car" <Model>.

    In this case, the "car" will have model_reference_type set to SE_MDL_REF_TYP_ROOT, since "car" can be instanced outside the scope of the <Model Library>, and in fact has a <Geometry Model Instance> under an <Environment Root>.

    On the other hand, the "top" <Model>, representing the top of the car, cannot be instanced outside the <Model Library> in the data provider's scheme of things, but only as part of more complex <Models>; consequently, the "top" <Model> has model_reference_type SE_MDL_REF_TYP_COMPONENT.

  5. A producer has a <Model> "plane" that has several components (two wings, tail, fuselage, etc). However, the producer has a <Model> "ship" that instances "plane" to identify a ship with planes on its deck. The transmittal will instance "ship", which has model_reference_type = SE_MDL_REF_TYP_ROOT. Since the <Model> "plane" could be used elsewhere in the transmittal, its instance under "ship" has model_reference_type = SE_MDL_REF_TYP_ROOT_AND_COMPONENT.

FAQs

Why do I need a dynamic_model_processing flag? When I retrieve data from a SEDRIS transmittal, won't the dynamic <Models> also be processed?
That depends entirely on how the data processing code was written. If the code starts at <Environment Root> and processes everything that the components of <Environment Root> eventually refer to, then dynamic <Models> will not be processed unless the dynamic <Model>(s) have a connection to the database (unless at least one reference is made to the dynamic <Model>(s) from somewhere within <Environment Root>). In some Interchange systems, the way around this has been to "connect" the dynamic <Model>(s) to something like the SW corner of the database. However, this requires "adding" information, which is not the purpose of SEDRIS.

As a consumer, if I am given a transmittal containing a <Model Library>, I keep track of which <Models> are instanced by some <Environment Root>. Does this information correlate at all with the dynamic_model_processing information of each <Model>?
Not necessarily.

As <Model Libraries> are passed around projects and re-used for purposes other than that for which they were built, many <Models> in any given <Model Library> may not be instanced in a given <Environment Root> that references that <Model Library>. For instance, a <Model Library> may contain a <Model> of a free-standing brick wall that by chance is not instanced by any <Environment Root>, but this does not make the brick wall dynamic.

On the other hand, consider a <Model> representing a biplane, in which the propellers are able to rotate. An <Environment Root> representing the Smithsonian Air and Space Museum might contain an instance of this <Model> representing one of the exhibits.

Can't any <Model> really be dynamic in the database? All I have to do is put it through my special processing, and it moves. Couldn't all the flags for dynamic model processing therefore be set to SE_TRUE?
Although the consumer can determine additional processing for any data read from SEDRIS, these flags are to be set by the data provider. If a <Model> instance is dynamic or has moving parts, then the data provider is required to provide this information.

Relating to the has_moving_parts flag, can't this information be derived by searching the entire <Model> for <LSR Transformation> instances that have <Control Links> attached to them?
Yes. The information is provided to allow consumers to detect at the 'top' level of the <Model> whether it has moving parts anywhere within its scope, rather than forcing them to (potentially) search the entire <Model> to derive this information.

Why can <Model> have at most 2 <Hierarchy Summary Item> components?
A <Model> is not required to have a hierarchy summary at all, but if a data provider wants to provide a summary of hierarchy, then a <Model> may have one summary for <Geometry>, one summary for <Features>, or a summary of each.

Can <Model> have both <Hierarchy Summary Item> and <Primitive Summary Item> components (as opposed to either/or)?
Yes.

Constraints

Composed of (two-way)

Composed of (two-way metadata)

Component of (two-way)

Inherited Field Elements

This class has no inherited field elements.

Field Elements

SE_String name; (notes)
SRM_SRF_Parameters srf_parameters;
SE_Model_Reference_Type model_reference_type;
SE_Boolean dynamic_model_processing; (notes)
SE_Boolean has_units; (notes)
SE_Boolean has_moving_parts; (notes)

Notes

Fields Notes


name

 A meaningful short name.  A full description will be in
 the <Description> component. This name is unique in a transmittal
 depending on the model_reference_type enumeration.

dynamic_model_processing

 Set to SE_TRUE only if this <Model> is used by the data provider to
 represent something that moves throughout the environment, such as a
 vehicle.
 This flag is used to identify information at the top level of model
 data, so it can only be set at the level where "model_reference_type"
 is not SE_MDL_REF_TYP_COMPONENT.

has_units

 Only takes effect if the srf_parameters are LSR.

 This flag allows a producer to say "This LSR Model is in metres" vs.
 "This LSR Model is unitless (it has no units)".

 Sometimes we want LSR models in metres, when we are modeling real-world
 things, e.g. tanks, ships, trees. Sometimes those models are scaled when
 instanced (trees are commonly scaled, but ships and tanks aren't.)

 Sometimes we want an LSR model to be purely unitless (e.g. a logo
 model).

has_moving_parts

 Set to SE_TRUE only if this model contains <Control Links> attached to
 <Transformations>, which allow or limit motion.

Prev: Mesh Face Table. Next: Model Instance Template Index. Up:Index.

Last updated: October 1, 2002 Copyright © 2002 SEDRIS™