By John Seely Brown and Paul Duguid; Published 2000 by Harvard Business School Press; ISBN 1-57851-708-7
This is a great book that highlights the errors that occur when we take too much of an info-centric view and forget the social aspects that influence how information technology is used. Short summary/snippets here:
The industrial revolution didn't lead to 'atom overload', why do we perceive information differently? It exists in stories, pictures, documents, conversations...
Information systems can suffer from 'tunnel design' - a one-dimensional info-centric view that over simplifies forces at work. A viable system must embrace not just technology, but the social - the people, organisations and institutions involved.
Automated agents (bots) are increasingly being used as brokers - of information, products and as merchants (comparison). But their uses are limited. Finding information in a disorderly and fast-changing web is difficult; Who controls the product recommendations for your basket - your preferences or what the supplier wants to sell; How do price comparisons compare to quality and service?
Work can rarely be completely automated - 'work to rule' often makes work impossible, human delegation relies on sympathetic disobedience (strict obedience can lead to disaster)
Designing bots to imitate or replicate human activities raises moral and social-institutional questions. But such issues have not typically been on the critical path of computer science
Office work is too easily painted as information handling. Good office design can produce powerful learning environments, random hot desk allocation makes incidental learning much harder to come by. People often find what they need to know by virtue of where they sit and who they see. Invisible peer support keeps many systems ticking over, undocumented and unrecognised but invaluable
Process re-engineering tends to focus on inputs and outputs rather than the meaning of the process, and people are inserted as needed. Support engineers may operate alone when fixing products, but they will often meet their peers informally, over breakfast/coffee/etc. - that majority of that time is often spent discussing work, sharing war stories, swapping notes, coming up with new ideas.
Process automation risks demanding that people forget what they know. Practice isn't perfect either - if it evolves too independently it becomes decoupled from the organisation's purpose.
Shouldn't a shift towards knowledge represent a shift towards people? To find information, we typically ask 'Where is it?' To find knowledge, we typically ask 'Who knows ....?'
To understand learning, explore the conditions in which knowledge and best practice thrive. People learn best when it is 'on demand' - in response to need, supported by regular use. Mandatory training does not equate to learning. Learning is social, it is often found in clusters (e.g. Silicon Valley)
There is a huge difference between 'Learning about...' and 'Learning to be...' Accumulating information is not enough to be considered an expert practitioner.
The use of documents often demonstrates the relationship between atoms and bits, but the relationship is often ignored - many people assume they are just carriers of information and therefore easy to transfer online. Documents can provide warranty and validity to information. The well-worn book, with comments in the margins, is different to the pristine new copy. A 'Take This' Post-it note on the back of a chair is easier to use (and more reliable) than a sheet of paper containing descriptions of items.