This content standard was adapted from "Metadata for Administration and Preservation of Digital Images (MAPDI)" prepared by Dr. Helen R. Tibbo and Claire Eager with assistance from a preservation metadata working group at the School of Information and Library Science for the State Library of North Carolina, Department of Cultural Resources, NC Exploring Cultural Heritage Online (ECHO) program through an LSTA Grant in September 2002. For more information on that project, please see http://www.archimuse.com/mw2003/papers/tibbo/tibbo.html
Additional thanks go to Michael Adamo at the Digital Production Center, Perkins Library, Duke University for his assistance on Checksum utilities and directions.

INTRODUCTION

Preservation metadata can be elusive because it is so often embedded in descriptive structures in such categories as condition or conservation history. The advent of digital surrogates created through digitization projects, though, puts preservation metadata on the front burner in terms of importance. These digital surrogates require a host of preservation information about their creation in order to build in long-term sustainability. With the increasing pace of technological innovation, hard-ware and software obsolescence worries, and storage medium changes, it is essential to maintain accurate and complete data regarding your digital production. This will ensure effective migration and thus underwrite your sustainability efforts. In addition, it ensures that NC ECHO's "Scan Once Methodology" can be used to full effect given the ability to preserve not only your original materials but also the digital surrogates that you create (see NC ECHO's Guidelines for Digitization: 4. Digital Production).

What is Preservation Metadata?

Preservation metadata for digital objects is just one aspect of your metadata program. This content standard only addresses the metadata needs for the preservation of your digital objects. Descriptive, analytical, administrative, and structural metadata are handled through other metadata standards. This should not be considered a viable replacement of such standards as Dublin Core, EAD, or other metadata structures supported by NC ECHO's Metadata Initiatives.

Where Do I Keep My Preservation Metadata?

Preservation metadata is internal information retained by an institution responsible for the maintenance of digital surrogates. This metadata format is not intended to be shared in consortial efforts but to assist in the long-term sustainability of their digital objects. It will also ensure that when a consortial digital repository is created for NC ECHO, each institution will have maintained the appropriate metadata to deposit their digital objects for long term preservation. Because of this, Preservation Metadata for Digital Objects is system-neutral. In other words, preservation metadata may be recorded in a variety of formats, such as an Excel spreadsheet, an Access database, or other similar tools. This standard has been created to ensure that the preservation metadata maintained by an institution will provide information that is essential or useful in the long-term sustainability of their digital objects.

Institutions currently using a collection management system or digital asset management system for digital objects may very well have addressed the issue of preservation metadata within their existing system. If you are using such a system, we recommend that you review the content standard that follows to ensure that your system is recording at least the required data elements specified in the standard.

For those institutions that do not already maintain a system that accounts for preservation metadata, an Access database tool is provided by NC ECHO for the maintenance of the preservation metadata. Adoption of this particular tool is not required by partner institutions. It is provided for those institutions that do not already have a system in place. For information about this tool, including implementation guidelines, please see PMDO Access Tool (http://www.ncecho.org/dig/pmdo.shtml). Note that NC ECHO does not assume responsibility for the maintenance of individual instantiations of this tool. It is being provided for the convenience of institutions but individual implementations are not supported by NC ECHO. While we will provide what guidance we can, institutions will be responsible for maintaining their own databases.

This tool and content standard were initially developed through an NC ECHO grant by Dr. Helen R. Tibbo at the School of Information and Library Science at the University of North Carolina at Chapel Hill. While adjustments have been made to each, the foundation created by Tibbo, graduate assistant Claire Eager, and her advisory team provide the backbone of this standard.

PURPOSE OF DIGITAL OBJECTS

It is expensive for institutions to go back and re-digitize their holdings. Few ever do so. In addition, many originals could suffer from the handling and exposure to bright light required by digitization. Therefore, it is best to simply "scan once," create a master image, and make any future duplicates from it. While using the term "image" here, this would also refer to other kinds of digital objects, such as sound and video files. As with the "scan once" terminology, this theory has grown from the initial attention to the creation of digital images to the creation of digital objects.

The Master image: The highest quality copy of a digital image, often called the master image, is expected to be a quality surrogate of the original. As such, it should represent the un-manipulated original and be created at a high resolution and stored in an uncompressed format (usually TIFF). High resolution equals large amounts of information captured, and large amounts of information captured usually equal a higher quality digital image. Higher quality digital images have a longer life and are more versatile. It is the master image that holds the promise of versatility and longevity. From it, high quality prints or publications might be made as well as derivatives for a variety of uses.

Derivative images: Access images are lower resolution copies are derived from the master by using a "save as" function and changing the storage format and resolution. They may be of varying quality and are generally manipulated for better display upon the screen or page (cropping, re-sizing, etc.) Additional images, such as "thumbnails" (even lower resolution copies) may also be created from the master or access image. These thumbnails allow for even quicker downloads of pages, and faster retrieval of large numbers of images.

Master images must be of the highest quality. Web images need not require such stringent quality controls. But, before compromising on image quality, consider the cost of migrating the image. Because migration is costly, it is far sounder to migrate a high quality (master) image than one of lesser quality. All digital objects will have to be migrated, if kept long enough. While the primary use of images in North Carolina ECHO is focused on Web access, repositories need to be mindful of future use, remembering the fragile nature of the originals. For more information about the creation and storage of these images, please see the NC ECHO Guidelines for Digitization (http://www.ncecho.org/dig/digguidelines.shtml).

The importance of this tripartite approach to digital objects cannot be overstated. In creating preservation metadata for your digital objects, you must keep in mind that you should retain information about each digital image, Master, Access, and Thumbnail. The Purpose element is thus required so that you may distinguish between the different versions of the same image.

PMDO CONTENT STANDARD: ELEMENTS AT AT GLANCE

* Master images have special requirements: They should not be compressed and should not contain a watermark. Therefore the Compression and Watermark fields would have a value of NO for master images, and Compression Type and Compression Degree would not apply to master images. For more information regarding this, please see Master Images in the NC ECHO Guidelines for Digitization.

PMDO CONTENT STANDARD ELEMENTS

APPENDIX A: USING CHECKSUMS TO VERIFY DATA TRANSFERS

1. Launch the checksum program by double clicking the icon. A Windows Explore type window will appear.
2. Navigate to the folder that needs checksums and click the Checksum button.
3. Another window will open and you will see the contents of the folder you have selected. Click Select All.
4. If the folder you selected has folders within it and you would like a checksum produced for each individual file then click the Add recursively button. If you do not click this button the program will generate one checksum value for each folder in the directory.
5. Click OK and the program will produce checksums for the selected files.
6. After the program has generated the checksums a Save As dialogue box will appear.
7. Navigate to where you would like to save the file and give the file a distinct name such as the folders path. Change the File type to Text File and click save.
8. You can now transfer your files to their new location.
9. Once the files are in their new location repeat steps 2-8.
10. Once you have checksum files for both directories you can now compare them to ensure that there is no data loss.
11. Open the checksum text file and copy the contents of the file. Paste this information into an Excel spreadsheet.
12. Delete any rows in the spreadsheet that do not have checksums.
13. Select column A and Click Data > Text to Columns.
14. In the dialogue box that opens click the "Delimited" radio button and then NEXT.
15. Place a check in the "Space" check box, click NEXT and then click finished.
16. Label the columns with the location associated with those checksums.
17. Repeat steps 11-16. In step 13 select column C instead of column A.
18. You should now have 4 columns of information in your spread sheet that looks like this:
    
    
19. Select the cell in column E row 2 and type "=EXACT(A1,C1)" then tab out of the field. This compares the two cells and returns a TRUE or FALSE.
20. The value should say TRUE. If it does not, the transferred file is corrupt and must be retransferred.
21. Select this field again; there will be a small black square that appears in the lower right hand corner of this box.
22. Click and drag this box down the entire E column so that all the checksum values can be compared.