The objective of this aspect of the TRS program is to address the areas of: burn profile definitions, data reduction methodology, data reduction automation, data format hardware independence, and the potential for future expansion. Over the years, the hardware configuration of the TRS acquisition and control system has evolved. Unfortunately, the current acquisition data formats are not compatible with the earlier formats, and conversion from one format to another requires some special effort and a knowledge of the past and present formats. A data base format needed to be developed to avoid this problem in the future. Dr. Nelson developed the THRIFT format based on the concept of the TIFF (Tagged Image File Format) standard which is used to store application independent representations of gray scale images. THRIFT is both machine and version independent, so it is completely expandable into the future. The THRIFT format is an expandable, machine independent data information format based on the concept of linked lists. All information within the file is stored as either two byte integers (short - 16 bit), four byte integers (long - 32 bit) or single byte characters. There are only two possible interpretations for an integer. The byte order is either low byte/high byte format (the Intel standard) or high byte/low byte format (the Motorola standard), so the only machine specific information in the data file is the integer standard being used. The THRIFT format is very simple. First there is an eight byte header. The first two bytes contain the ASCII values for either II or MM to indicate an Intel or Motorola integer format. The next two bytes are a short integer representing the version for the data format in BCD where the low byte is the minor version number and the high byte is the major number. The last four bytes are a long integer that points to the first data directory. This pointer is NULL if there are no data. A data directory consists of a tag counter and a list of tags. The tag counter is a short integer that indicates the number of tags to follow. A tag is an identifier that contains the information about the data specific to the THRIFT format. There are currently 54 tag types defined for THRIFT files. The THRIFT format may be extended by the judicious addition of tags and data types. Once a new tag and/or data type is created, the file format version number must be changed by incrementing the minor and/or major version values. Additions to the format should be performed on an "as required" bases and must be centrally controlled by DNA. Failure to properly control and coordinate format revisions would effectively nullify the usefulness of the THRIFT format. It is also important to note that once a format revision has been released, it cannot be "unreleased". Old format structures cannot be removed, only new structures can be added. A THRIFT file API and Software Development Kit (SDK) have been created and were used to develop the Windows based thermal radiation and air blast data analysis software as well as a series of THRIFT file utilities for the DOS, Windows, and UNIX environments. DNA users are encouraged to retreive the most recent version of the Windows based software (currently 3.0 beta R7). The documentation, and the complete API specification will be available on line at a later time. Source code is availble on authorized request. Home | Products
| Services | Software 06/20/00 ern |