One of the other biggest problems in distributing digital imagery has been how to make the images easily accessible through descriptive words. How does one describe a photograph? How can photographs be keyworded for easy retrieval? Computers are wonderful for organising and storing information, but how do you find something once it's stored in invisible digital form? This is one of the biggest problems with digital storage today, and it has proved to be the biggest surprise obstacle in the use of digital photography retrieval technology (Russell, 1993).
Keywording has proved to be a very complex, time consuming, and expensive — but a very critical — way of finding an image. It averages around ten dollars in costs and labor to describe an image. Most keyworded images have ended up with descriptions totalling over one- hundred words. This has largely been the result of the need for synonyms, since not everyone uses the same words to describe a situation (Russell, 1993). Thesauruses have been introduced as a tool to get around this problem, but they have first needed to be modified for use with photographic images, and even then have not proven to be very successful (Foss, 1994b).
There are two popular methods for keywording descriptions of photographs. Keywords can be entered into single fields, which allow for the use of a built in thesaurus for analysing each word independently. This method is more successful in finding synonyms, but usually creates problems when people are looking for titles, phrases of words, or celebrity names such as 'The Statue of Liberty', or 'River Phoenix' (Pickerell and Child, 1994).
The alternative method for describing photography simply uses a string of words in paragraph form. While this makes the ability to find 'The Statue of Liberty' and 'River Phoenix' easier, it creates problems with synonyms like 'creek' and 'stream', or interpretive differences (Pickerell and Child, 1994).
Traditionally, photo researchers in stock photography libraries have been able to mentally compensate for these discrepancies. Keywording techniques, therefore, must be written with a vocabulary that a photo researcher or photo-buyer would use. They must also contain singular as well as plural forms of nouns, be systematic, consistent, and address whether the image has a horizontal or vertical orientation. There should also be a statement as to whether the image is model or property released. Keyworders must also be very careful about punctuation, and the use of technical words (Pickerell and Child, 1994).