TCRS – A Methodology and Tool Set for Specifying Data Content Jesse Campos, Greg Hull Science Applications International Corporation (SAIC) 12901 Science Drive, Orlando, Florida 32826 greg.a.hull@saic.com jesse.j.campos@saic.com Farid Mamaghani SEDRIS Organization 19223 SE 45 th Ct Issaquah, Washington 98027 farid@sedris.org Keywords: Environmental Data, Interoperability, TCRS, SEDRIS, XML ABSTRACT : Specifying the requirements for environmental data at the input and output of systems or applications, and the ability to automatically evaluate and validate the data based on such requirements, is a key ingredient to successful data interoperability. This paper will describe an innovative and ongoing development, the Transmittal Content Requirements Specification (TCRS), which addresses this challenge. TCRS leverages the SEDRIS technologies of DRM, EDCS, and SRM to provide a formal methodology for the expression, and the subsequent evaluation, of environmental data requirements. TCRS is composed of several key technical components, including a process and methodology for articulating and capturing the requirements; a complete syntax for expressing the requirements; an XML encoding of the syntax; and automated tools, such as a parser and evaluator, for validating transmittals that claim conformance to a given set of requirements. This paper will provide an overview of the various technical components, and will highlight examples in the application of TCRS to real world data sets. 1. Introduction capability, but do not directly implement the verification of data content or the notion of “fit for use”. The SEDRIS technologies provide a robust mechanism for the SEDRIS DRM and EDCS provide the framework to representation of environmental data across all domains conduct content and applicability analysis. But in order to and applications. In order to provide these broad specify environmental data requirements, a capabilities, five technology components have been complimentary technology, the Transmittal Content developed. These five technology components are: Requirement Specification (TCRS), has been developed to augment the DRM and EDCS in order to capture such • requirements. The capabilities presented in this paper The SEDRIS Data Representation Model (DRM) describe the methodology for the expression of that supports the full representation of any environmental data requirements, an XML encoding for environmental data. • the specification of such requirements, and a toolset for The Spatial Reference Model (SRM) that data verification. supports an extensive definition of Spatial Reference Frames (SRF). • This paper begins by describing the role that TCRS plays The Environmental Data Coding Specification in the interchange of environmental data. This will (EDCS) that provides complete classification include inherent problems with current specification and attribution for environmental data. • techniques and examples of initial TCRS solutions. The An interface specification, the SEDRIS API, paper then introduces the current TCRS methodology and which allows the user to develop applications to tools for improving interchange. The XML encoding of consume and produce environmental data. • TCRS and examples using the syntax will be highlighted. A binary file format, the SEDRIS Transmittal The paper will include several examples of TCRS Format (STF), specifically designed for the requirements. The future development and enhancement efficient storage and access of SEDRIS data. plans will conclude the paper. These SEDRIS technologies provide for a complete and unambiguous data representation and interchange
2. Need for TCRS requirements. A TCRS requirement , in this sense is composed of two parts: a domain and a condition. The 2.1 Data requirements domain is the set of all objects for which a requirement is evaluated. The condition is the set of criteria applied to A TCRS document is a specification of the environmental each object in the domain . In the previous example, the data “requirements” expressed using the SEDRIS DRM, domain consists of all (software) objects that represent EDCS, and SRM terminology. A requirement may be trees. The condition is that the (software) object about the semantics of the data, such as which representing the tree also be attributed with a height as environmental objects must or must not be present. Or a well as a stem diameter. Note that what is missing here is requirement may be about the structure or form of information about how the trees may be represented and representation that the data may take, such as whether how they may be attributed. data is in gridded or point sample form. Specifying data requirements allows the evaluation of data based on “fit Consider two other requirements, again expressed in for use” criteria for a particular application or set of English, but using DRM terms: applications. The specification provides an “Acceptance Test” for data, much like acceptance tests for software or “Buildings must be represented by <Point_Feature>s with hardware products. However, a common methodology <Property_Value> components for height and width.” and a common terminology are required to accomplish this. “Buildings represented as <Point_Feature>s must be given <Property_Value> components for height and 2.2 Specification solutions width.” Traditionally two mechanisms have been used to capture At a glance these two requirements may seem the same, requirements. The first is simply to express the but the difference lies in the distinct domains to which requirements in English, usually through Word their respective conditions are applied. This is what is documents. The second is to develop highly unique, often not clear in English requirements and is the reason tailored verification software. The Close Combat Tactical that TCRS syntax must have an explicit distinction Trainer (CCTT) Correlated DataBase (CDB) TCRS is an between its domain and condition. The domain of the example of expressing requirements through a Word first example is all DRM objects which represent document [1]. Environmental data model compliancy Buildings, while the domain of the second is only those checkers are an example of the TCRS method of writing <Point_Feature> objects representing buildings. tailored software. Therefore a transmittal that contains buildings represented as <Polygon>s with a height and width will fail the first 2.3 Short-comings of specification solutions requirement. Note that the example is still missing an important piece; namely which EDCS entries will be used The problem with requirements embedded in unique to identify the concepts of building, width, and height. software is that they are single-point solutions. The TCRS syntax must also use the syntax of EDCS, and, Requirements expressed in Word documents can be similarly, of the SRM. This requires the TCRS syntax to problematic since they are often unclear and imprecise be flexible and expressive, allowing a user to define many due to human interpretation, and are not machine different object sets for domains and conditions. This is parsable. For example, the following could be stated as a done by introducing Object Matching Expressions requirement: “All trees must have a height and a stem (OME). diameter”. In order to determine if a data set meets this requirement, it is necessary to specify how a tree is 3.2 Object Matching Expressions represented, and whether “tree” refers to a single tree, a bush, a treeline, etc. These types of ambiguities can be An OME is an expression composed of a set of individual avoided through the use of a formal syntax that utilizes criteria, where each criteria is a pass/fail test for a given terms from the DRM and the EDCS. object. An OME is evaluated against a SEDRIS object and matches the object if the evaluation is determined to 3. Using TCRS be true. In the process of matching one object, other selected objects (components for example) may be 3.1 TCRS Concepts involved. This set of objects satisfies the OME if the evaluation is determined to be true. There are 10 different By relying on the SEDRIS DRM and the EDCS, the criteria that can be used to build OMEs and are defined TCRS concepts provide the capability to specify the below: broadest possible range of environmental data
Recommend
More recommend
