Servizio Comunitario di Informazione in materia di Ricerca e Sviluppo - CORDIS

Technology and infrastructures policy in the knowledge-based economy - Solving the paradox of digital memory

There is a range of policy issues to be addressed around the paradox of “digital memory”. This is a paradox because on the one hand many public and private organizations are keen to implement procedures and mechanisms for building “organizational memories” (with appropriate incentive structures and coordination devices); while on the other hand there is a general concern about the “memory capacity” of societies facing the challenges of storing and archiving huge amount of digital data and information.

The problems and difficulties related to digital memory have three sources: Firstly, we should note that with information technologies, what we record are not documents but sets of instructions (Cowan, Foray, 1997). Because paper has a low acid content, printed books and journals can last for centuries. Physically, it deteriorates slowly and because the language in which it is written evolves very slowly, the interpretation of even very old texts remains relatively easy. This is not necessarily true for documents stored on magnetic or optical media. Magnetic tapes deteriorate fast from a physical point of view and, given the rapid changes in software technology and computer hardware, languages can disappear just as fast. The problem is that with information technologies, what we record is not documents but sets of instructions that have to be interpreted and managed by appropriate equipment and software before the information they contain can be used. Thus, although short-term storage and data retrieval costs have decreased, long-term storage (i.e. archiving) and access to old documents clearly remain a problem.

Secondly, the rapidity of knowledge production and codification processes and the low and decreasing costs of storing codified knowledge make the problem of attention more acute (Simon, 1982). The classical formulation of that problem is that, more than ever, it is attention rather than information that is becoming a rare resource as screening and selection of information become important functions. Information abundance is generating a problem for agents to discriminate between information which are important to store and memorize and information that can be simply put in the basket.

One should note that if you look at the web today, people are storing too much, and putting too little in the basket. However, this second problem can be formulated in a slightly different way: given the reasonable assumption that the attention management problem has always existed in some degrees, if we can make the claim that more knowledge is being stored and made available in codified form, the knowledge management problem changes shape. To use the Lundvall taxonomy, when the knowledge we seek (to understand) is tacit, "know-who" is extremely important. When it is stored in codified form, "know-where" becomes important. A new skill must be cultivated, namely how to find things using, for example, the search tools on the web. This has two aspects: the first is how to find likely documents. The second tool that has to be developed is how to judge quality. Because the cost of codifying (and publishing) has fallen so much, it is "too easy" to diffuse knowledge. There is a lot of codified "knowledge" available on the web that is misleading or just plain wrong. How to teach people to filter the good from the bad may be an important issue.

In a sense, know-where and know-who should be combined to develop the real problematic of “software agents”: the best agents would not only be efficient in finding all the information corresponding to a certain question but they will take into account the peculiarities of the user and the situation.

Thirdly, knowledge is divided and dispersed (Machlup, 1984). The division of knowledge is a result of the division of labour in the field of knowledge production. The dispersion of knowledge is related to local situations in which knowledge is produced (a site, a workshop, a laboratory). An increase in the division and dispersion of knowledge makes it more and more difficult for economic agents to locate and retrieve elements of knowledge that would be particularly useful to them. It is probably not disputable that the division of knowledge is increasing over time (specialization), raising the marginal cost of knowledge integration. The dispersion trend is less clear, but, as a law, one can expect a higher dispersion as knowledge production becomes more collectively distributed (located in many places). And this increasing tendency of knowledge division and dispersion again makes the problem of memory very difficult. How to build storage processes that are integrative; that is to say which are not just reproducing the state of division and dispersion of knowledge as it is at the moment of its creation.

Facing those challenges – problems posed by recording the set of instructions, by information abundance and by the division and dispersion of knowledge – an important policy objective should be to reach a higher level of organizational procedures and capabilities in a way that could allow organizations and agents to overcome this digital memory challenge.

Reported by

University Louis Pasteur
Avenue de la Foret Noire, 61
67000 Strasbourg
See on map


Evaluation - Policies