DIGITIZATION

Ultimately, everything that people are interested in accessing will have to be digitized. The reason is that digital searching will become so easy, inexpensive, fast and ubiquitous that users will not tolerate, or will not access, traditional materials. Capture requires a concerted, shared, worldwide effort. The cost of digitizing is not trivial, so it makes little sense for any work to be digitized more than once. Yet without any registry of digitized works, many books are digitized multiple times, while others are ignored.

Converting text, images and objects to digital form requires much more than digital photography or even high-resolution scanning and requires, instead, the following:

• initial input, either scanning or keyboarding
• conversion to one of a set of standard formats
• optical character recognition (OCR) to capture text characters for searching
• OCR correction (since OCR is inherently error prone)
• creation and input of metadata and cataloging information
• special techniques for non-textual materials, such as music, images, videotape, etc.

Policy issues, discussed below, will determine the resources made available for digitization and how they will be allocated.