Version 1.1 of the NOST Definition of FITS was recently completed and approved by the NOST FITS Technical Panel under the chairmanship of Bob Hanisch, Space Telescope Science Institute (STScI). Version 1.0 was originally accredited in June 1993 and subsequently forwarded to the International Astronomical Union (IAU) for accreditation as the official FITS standard recognized internationally. When that version was advanced to the IAU, a question arose regarding recommendations for units to be used when measuring angles. In response to the concerns of the IAU, the new version contains recommendations for the use of decimal degrees for angular measurements.
This updated recommendation should clear up some "units" ambiguities in the original FITS papers. Version 1.1 of the NOST FITS standard is now awaiting final approval of the NOST FITS Accreditation Panel. When final approval is obtained, the new standard will be made available in both on-line and hard copy formats. NOST intends to submit again the NOST-approved version to the IAU for its international approval.
NOST has been pushing the development of a standard means of referencing files (or objects), on the same underlying physical media, from within Standard Formatted Data Units (SFDUs) on that media. An annex of the May 1992 SFDU - Structures and Construction Rules Recommendation contained proposed canonical Consultative Committee for Space Data Systems (CCSDS) Referencing Environment Specifications. Those specifications were included to allow testing and prototyping of systems using SFDUs and were stated not to be the final solutions. Since that time several other proposed specifications have also made their way into use by several projects.
The current efforts seek to document formally these early efforts and to finalize these specifications so that data systems can truly interoperate using these referenced objects. The current proposals recognize four referencing environments.
CCSDS1 is a DOS-compatible and CD-ROM-compatible directory/file naming specification. CCSDS2 is a long directory/file naming specification that, except for allowing a length of 255 characters rather than 14 characters, complies with the portable POSIX specification. CCSDS3 provides for absolutely or relatively referenced data objects by file position on sequential media such as tape. Finally, CCSDS0 provides a means to specify equivalent names in two or more of the environments mentioned above when the product creator expects the data to be transferred across systems or media that have different file referencing needs.
NOST hopes to reach consensus with the other CCSDS representatives at an upcoming international workshop this fall. Therefore, NOST readers should expect to comment on a draft Recommendation (Red Book) version early in the next calendar year. In the meantime, NOST would like to remind all projects that NOST personnel are always available to discuss the current proposals and to advance suggestions to the international community. NOST expertise is always available to projects to help them prototype the most up-to-date drafts of these standards so any future changes will have the minimum amount of impact.
CCSDS has been addressing the need for additional formalism in how the meaning of data entities/data structures, found in NOST science data products, can be specified so as to increase automation in the handling of newly received data products. This is a complex and far ranging problem, and there are many different approaches that have been tried by various physical and computer science communities. Recently, CCSDS has taken the approach of providing a minimal but user extendible set of standards for expressing simple semantics needed for the interpretation of data products. These conventions are organized in two levels so as to work with existing, somewhat self-describing data formats such as FITS, Common Data Format (CDF), or Hierarchical Data Format (HDF) as well as those without self-description. Currently, the draft provides minimal support for the specification of data type domains because this opens up a broad set of issues and the need for much greater formalization.
NOST needs to determine the usefulness of this approach. The current draft, entitled "Data Entity Dictionary Specification Language," is expected to be ready for review and testing by mid-October 1995. It will be accessible from the NOST Home Page. Readers can look for it then and send NOST their comments.