Today, we are classifying knowledge as we did in 1800: the same structure but with different formats and supports.
At the time when knowledge went from being stored only in our heads to be written, information required physical and localized support. The need to create an organizational system emerged with large collections of data. As the volume and complexity increased, it was necessary to create physical systems of management and classification based on categories and hierarchies that would enable their location. These categories and hierarchies were adopted by libraries and, ultimately, offices.
At the same time, writing allowed giving a structure to the knowledge we had of the world. Categories were created to classify and relate ideas through a hierarchical system that sought to rationally reflect a natural order. Yet, the classification of information is anything but natural, since there are different logical ways of classifying knowledge.
Trying to mimic how offices worked in the 50’s is the root of our today’s information chaos.
When the first computers were created, their operation was based on command-line interfaces that required technical knowledge. To evolve into personal computers, it was necessary to simplify their operation. This is how Alan Kay of Xerox PARC introduced the desktop metaphor in 1970.
Computers adopted the metaphor of an office, with icons, actions, and procedures that mimicked familiar elements for non-expert users. By using analogies, such as creating, saving, or deleting a document, they made accessible abstract and strange concepts that are now all too familiar to us. But for the user, documents and folders stopped being a metaphor to become real things.
However, any shortcut has a price. Just because the system based on files and folders is easier to understand doesn’t mean it’s more practical and efficient. By adopting the desktop metaphor, its limitations were also adopted. This system separated information into individual pieces to identify and locate them, generating a hierarchical system.
It was assumed that this would facilitate the search for information, but cognitive approaches to knowledge organization derived from empirical studies have shown that users are more comfortable with verbal search systems. In turn, experts in knowledge organizations, such as Birger Hjørland, have revealed that different viewpoints require different systems of organization of information, as exemplified by multinationals with employees from different cultures.
Towards an information system based on how our mind works
While computers remain anchored in a limited and unproductive system inherited from physical files, the internet has been organizing knowledge in a much more efficient way for some time already: hashtags, ontologies, semantic search…. These are, in short, management systems closer to how language and our minds work.
What would happen if we “stop using” folders to store our digital assets (from documents to videos) to facilitate the search for information inside our companies? Can ideas be “closed” inside a folder?
But let’s go even further, does it make sense at all to use “files” as they are today, to find the information we need?
According FADGI this is a “digital file” definition:
At a high level of abstraction, a digital or computer file is a stored segment or block of information that is available to a computer program. Files are so named because they are the counterparts of the paper documents traditionally kept in file folders, usually stored in a file cabinet. Computer operating systems consider files as a sequence of bytes, while application software interprets the binary data as, say, text characters, image pixels, or audio samples.
Let’s think: when we are looking for some internal information from our company, we are now forced to guess in what data source it might be, then we have to look for a folder, then we have to find the file (usually a document or a pdf), read it (or shortcut the read by searching inside the document), and then, if we are lucky, we get the information we needed.
But is this really a good approach? For most of our information needs, we only require a very small fraction from the file we are reading to get relevant information: we just need a paragraph or two from the whole file. So, we have been looking in different data sources, opened several folders, read several files, just to get the real unit of knowledge: a paragraph.
This huge inefficiency is reflecting some dramatic numbers:
– According to IDC knowledge workers waste around 8 hours per week looking for internal company’s information.
– According to Iena Sheengar (Columbia Business School) average knowledge worker must process, consciously or subconsciously, the equivalent of 147 newspapers of information every day.
Yes, you read it correctly: we waste 8 hours per week and we must process daily the information contained in 147 newspapers, and we are facing this huge numbers with a methodology used few centuries ago to classify “papers”.
It really seems we are not doing things efficiently today, and the solution can only come from artificial intelligence. AI will (re)evolutionize how we access to our company’s information in a very short time.
As data grows, insight-driven enterprises are reinventing their data supply chains to stay agile and create competitive advantages. Harnessing the power of unstructured data (documents, images, emails, media, etc.) is vital to this process. But as 80% of all data is unstructured, this is not an easy task.
The future of knowledge accessibility will be by using AI and Natural Language processing, to deliver actionable insights derived from the full spectrum of content and data sources. Whether they are internal or external. Some solutions are already extracting valuable information from large volumes of complex, diverse data sources, extracts name entities: name of people, name of companies, dates. etc…to provide deeper insights to users so they can make better decisions and discover knowledge. The unit of knowledge are not “files” anymore, the unit of knowledge providing wisdom are paragraphs from different documents that are put together to have fast and good overview from all the information needed to do our job faster and better and overcome, definitely, information chaos.