Release of 6 February 1997
Questions and Answers
1. EBCDIC in tape labels
[Reference: COF vs 1.1, Sections 8.1 and 14.3.4]
Q. The document describes the image file name as an ASCII string, but section 14.3.4 says that tapes will have a standard EBCDIC label. Does the file name show up somewhere else on the tape, or does Section 14.3.4 supercede Section 8.1 for tape media?
A. All character strings in the tape labels, including file names, are encoded in upper-case EBCDIC.
2. Image Format Subsets
[Reference: COF vs 1.1, Section 13.2]
Q. Without a copy of the X9.46 documentation I had a hard time understanding this section. Is it merely restating information provided elsewhere in the COF spec, or do I need to review the subsets identified in X9.46?
A. All information required to make COF-compliant image files is included in either the COF document or the underlying TIFF and IOCA specification documents.
3. ImageData to DBF Consistency on CD-ROM
[Reference: COF vs 1.1, Section 13.4.1]
Q. This section refers to information in the DBF file, which is not included on tapes, but I infer from the context that the ItemData is included in the Images File even on tapes. Is this the correct interpretation?
A. Yes. The Images File has the identical format on both tape and CD- ROM media. For the CD-ROM case, that language is intended to enforce consistency between the information in the ItemData field in the Images File and the corresponding information in the DBF file.
4. Notes under IOCA Tables
[Reference: COF vs 1.1, Sections 13.6.1 and 13.6.2]
Q. The notes prohibit JPEG compression for IOCA binary files and G4 compression for IOCA grayscale. JPEG is a grayscale compression algorithm and G4 is for black-and-white, so the stated restrictions are obvious. Was the intent really to limit IOCA files to ABIC compression, or will IOCA/G4 for binary images and IOCA/JPEG for grayscale meet the COF spec? (I realize the issue at hand is limited to IOCA/ABIC, but I wonder if the COF spec is).
A. This is an error; the notes under the two tables are reversed. The restrictions should instead proscribe Group 4 compression from IOCA binary files and JPEG compression from IOCA grayscale files. TIFF is the file format used for Group 4 and JPEG images. The reason IOCA is used for the ABIC binary and concatenated ABIC grayscale compression types is that these compression types are not available in TIFF.
5. Pre-Mastering Tapes
Q. The Media Header File contains information about how many file suites are on the tape. If we always write a single file suite, this is simple. However, if we really give the user flexibility as is suggested, that forces us to write everything to DASD then when we know how many file suites are needed, rewrite it to tape.
A. This has been changed. Such fields have been moved to the Media Trailer File to facilitate a dynamic choice of the number of media sets.
6. Character Set
Q. Numerous references are made to 'cr' 'lf', 'sp', and 'tab'. I assume that these are the hex values 0D, 0A, 20, and 09 respectively, although I can't find this explicitly defined in the COF spec.
A. That is correct.
7. Format of .ini Files
[Reference: COF vs 1.1, Section 8.3.1]
The section cited above makes reference to lines and columns. Am I correct in assuming that the end of a line is delimited by x0D0A?
A. Since the COF was not specific about this, good cross-platform design practice would argue for a reader to also be tolerant of files having only 0x0A at the end of a line of a .ini file.
8. Record Structure
Q. When writing these files to a tape with variable length records,  is any correlation implied or required between "lines" and records?  Specifically, do we need to write one line per record, or  may we write many lines in a single record?
A.  No.  No.  Yes.
The author of this FAQ is Louis H. Sharpe II (firstname.lastname@example.org).
Return to the top of page, to the COF section or to our home page.
Last revised March 25, 1997.
This page is
Send comments to the Federal Reserve Bank of Boston.