| Information Structures and Indexing
An often overlooked aspect of information management -- particularly when that information is being created or maintained in electronic form -- is the development of file plans and indexing schemes for the information under management.
In any information system, the 'garbage in, garbage out' rule prevails. Even the most comprehensive information set is valueless if the quantum of information that desired at any given moment could not be located. And location of the correct information -- whether to bill a customer, design a widget, or assert a claim in court -- is vital to any organization.
Unfortunately, however, the organization of information is frequently an afterthought, or is not addressed at all. Therefore, as the information set under management grows, search and retrieval becomes more and more difficult and time-consuming and costly, and eventually, effective search and retrieval becomes impossible. Although software tools of increasing complexity and power are available, the 'garbage rule' applies to them as well. No such tool can take poorly organized or unorganized information and magically create order without human intervention.
Consider, for example, imaged documents. For imaged documents, OCR (optical character recognition), combined with full-text searches, is often touted as an automated alternative to proper indexing of the image set.
Full-text searching never provide an adequate alternative to a proper file plan/indexing system. OCR software introduces a significant error rate into the document set, and users have quickly discovered that guessing how a precise topic is mentioned in the body of a large set of documents that may have many authors is a formidable problem. The vagaries of language, and lack of standardization in usage make full text searching a poor substitute for proper indexing, even though it can be a very powerful tool indeed when used in conjunction with a suitable indexing scheme.
Other tools, such as 'fuzzy logic' programs that attempt to analyze and index electronic documents without human intervention are beset with similar problems. None has yet shown itself to be an adequate substitute for a well-done index.
A properly done file plan/index must address several important considerations:
- It must use categories and terms that correspond to the queries actually used by searchers
- It must correspond to the actual record types in use
- It must correspond to the legal requirements imposed upon the records and information in the set
Automated tools cannot give adequate consideration to these things. Although PelliGroup strongly believes in technology as a solution to information management issues, we feel that it is only a partial tool for indexing and file plan development. We do these things the old-fashioned way -- we interview staff, review inventories, ask questions, analyze and revise. Only when we are satisfied that we understand the nature of the system, the kind of queries that staff will really make, the kinds of information they expect to find, and all of the other aspects of an organization's information processes. Do we then consider importing our findings -- indexing structures, file plans, keywords, and so on -- into information management or document management software or other technology tools. | | |