STANDARDS CHANGE REQUEST - DVD Case Standards ============================================= Purpose: Establish rules on the use of upper/lower case characters in file and directory names -- first, in a media independent manner, then as used when writing DVD archive volumes. CD (CD-ROM or CD-R) archive volumes will continue to be written according to the current standards. Date: January 21, 2005: revision 3.0 Working Group: J. Wilf (lead), C. Isbell, C. Atkins, M. McAuley, T. King, E. Means and L. Huber Background ========== When Chapter 19 of the PDS Standards, "Volume Organization and Naming," was written, it was intended specifically for archive volumes written to compact optical disc (CD) media. It therefore followed the ISO-9660 standard, which specifies the file and volume structure for CD media and requires UPPER CASE directory and file names. As PDS data volume grew, it became necessary to archive on higher-capacity DVD media. The Data Distribution Laboratory (DDL), which had helped develop the PDS standards for CD media volumes, made the following recommendations for DVD. These were based on extensive media testing, with the purpose of ensuring the most uniform ability to read and write DVD media across all computer platforms: 1) Write DVD media using the UDF-ISO bridge format, where UDF (Universal Disk Format) is a filesystem standard chosen for DVD and higher-capacity media, and where the UDF-ISO bridge format creates both an ISO-9660 and UDF partition on the DVD. 2) Convert all directory and file names to lower case before writing the DVD. This leads to a lower case UDF partition and an upper case ISO-9660 partition, since the DVD-writing software automatically does lower-to-UPPER case conversion for the ISO-9660 partition. Case is preserved within the UDF partition. Although the UDF partition can accept either upper or lower case names, there were several reasons for choosing lower case. DDL testing showed that the lower case UDF partition was read correctly across Windows, Solaris, Linux, and Mac platforms, whereas an upper case UDF partition was read incorrectly on Windows as having an upper case first character, followed by lower case subsequent characters. Also, user interface studies have shown that lower case directory and file names are far more readable. The problem is that the recommendation described above conflicts with the PDS Standards Reference, which always specifies upper case names for files and directories. Data providers are understandably confused when they see upper case names in the Standards but are told to create lower case names on DVD. In addition, it is recognized that the archive should be as independent as possible from the media on which it is currently stored. CD media are already giving way to DVD. Plans are in place to move responsibility for long-term archiving to the NSSDC, where "archive packages" will be stored on the media of NSSDC choice -- and refreshed and migrated to new media as needed. In this environment, it is important to have a file/directory standard that is media independent. And therefore, this SCR establishes a media independent file/directory standard on the use of upper/lower case filenames first, then describes the standards that apply when writing these files to DVD media. NOTE: Recommendations on specific hardware, software, software settings, and media for writing to DVD media will be given in a separate document, rather than the Standards Reference. Current Urgency =============== The MGS-Extended and NEAR missions have already delivered archive data on DVD media, and the Cassini mission is planning to do so in the near future. Recommendations =============== Given the urgency described above, the PDS Project Engineer recommends the following actions: 1. To Management Council: Approve the DVD CASE SCR as soon as possible. 2. To the PDS Standards Lead: Upon Management Council approval, publish the changes in the next version of the PDS Standards Reference, as described in the paragraphs below. *** The Following Issues need to be Resolved in order to complete this SCR *** ============================================================================== These need to be discussed with the entire Tech Session, since all discipline nodes are stake-holders: 1. What upper/lower case requirements should be levied on all files and directories in the archive (independent of media)? A. From comments heard and received to-date, it seems generally agreed that mixed case in a single file/directory name should not be allowed, e.g., HelloWorld.TXT is not an acceptable file name. B. Given A, there have been three proposed standards for further restricting the use of upper/lower case file and directory names. These are listed in order from least to most restrictive: 1) Case uniqueness: File/directory names may be either upper or lower case, provided they are a unique sequence of characters within their parent directory, e.g., there cannot be both helloworld.txt and HELLOWORLD.TXT files in the same directory. 2) Case consistency within a data set archive: All File/directory names within the same data set archive must have the same case, and that case may be either upper case or lower case. (Lower case is recommended, but this is a guideline, not part of the standards.) 3) Case consistency across the entire archive (with lower case chosen as the standard): All File/directory names in the PDS archive must be lower case. 2. What about software files that must have mixed-case names, e.g., Java class files, where the name must match mixed-case class names (HelloWorld)? Suggestion: PDS requires that these software files are distributed within ZIP archives, e.g., helloworld.zip. 3. What are the requirements for file names used in PDS labels (e.g., pointers)? Suggestion: Either upper or lower case file names may be used. (It is recommended that the file name in the label match the case of the actual file name, but this is a guideline, not part of the standards. As stated below, PDS tools are required to handle file/directory names in a case insensitive manner.) Note: Values for ODL identifiers that may use file name information, e.g., VOLUME_ID, will remain upper case. 4. What are the requirements for file/directory names used in PDS index tables? Suggestion: Since index tables may be used by external tools to access files, the file/directory names in them should be made consistent with the actual file names. 5. From comments heard or recieved to-date, it seems generally agreed that PDS Tools and data access middleware are required to be able to access files in a case insensitive manner. 6. Section 11 of the Standards Reference already contains a standard for writing DVD media using ISO9660_UDF bridge. Since the UDF partition accepts upper or lower case file names, it should be written consistent with the media independent case requirement decided in item 1, above. 7. Again, NOTE: Recommendations on specific hardware, software, software settings, and media for writing to DVD media will be given in a separate document, rather than the Standards Reference. Changes to the Standards Reference ================================== [This section will be written when the issues described above are resolved.] Changes to the Data Dictionary - None ============================== Changes to the PDS Tool Suite ============================= All Common PDS Tools will be modified, if needed, in order to access PDS files in a case insensitive manner. Changes to the PDS Online System ================================ All Online System software that accesses PDS data repositories will be modified, if needed, in order to access PDS files in a case insensitive manner. Impact Statement ================ [This section will be written when the issues described above are resolved and an assessment made of needed tool and system changes]. Open Issues =========== [See above] [END OF SCR]