INTEROPERABILITY

Given that numerous libraries around the world will be developing digital collections, how will it be possible for a user of library A to access and view material housed in library B? The existence of numerous digital libraries will make it essential to share digitized items and ensure that cataloging, searching and retrieval tools at each one can be used readily with materials from others.

Formats and Standards

While standards may have the effect of inhibiting innovation, they are essential to interoperability. Agreement must be achieved on such fundamental issues as how text is to be stored. Is it straight ASCII, Microsoft Word, HTML, SGML, XML or something else? What kind of compression will be used? If text is compressed, how will searching be done? How are images, music and videotape to be represented? If agreement is not reached, at least the number of different ways in which works are digitized should be reduced to a number small enough to allow each library to support them.

Digital libraries must also have a second set of intake standards, going not to technology but to quality and reliability, which are discussed later on. Archivists question the permanence of digital materials since they note that electronic documents can be modified readily and the media on which they reside become obsolete at least once each decade. The question then is how an ever-expanding corpus of information will be converted to new media and formats as these evolve.

Metadata

This term is often used to mean information about an item, rather than the information in the item itself. Examples include the author, title, date of acquisition, price paid, donor, etc. It is particularly critical to capture metadata that is not present in or derivable from the item. For example, the author's date of birth is often not printed in a book but can be important in distinguishing among authors with similar names (particularly parents and children). Libraries may "share" content by simply providing links, but uniform access to the content requires uniform metadata and a procedure for generating and storing it economically. It is of little avail to exchange documents at light speed if they must be held up for months until a human cataloger can prepare metadata.

Character Set Representations

This is not merely a question of different alphabets and writing systems, a major hurdle in itself, but also an issue of how characters are represented. For example, there are several widely differing mappings of Chinese characters into ASCII. There is some appeal to having a worldwide universal standard, such as Unicode, but the notion of attempting to list all of the world's glyphs and freeze them in a standard reduces flexibility and tends to overlook obscure or variant writing systems and restrict the development of new ones. Possibly a standard should be developed that permits new character sets so long as the definition of the glyphs and the representation mapping is maintained in an accessible location.