In his bestselling book, “You Are Not a Gadget”, Jaron Lanier talks about our understanding of the humble computer “file” as an entrenched software philosophy becoming invisible through ubiquity: “files are now part of life”:
The file is a set of philosophical ideas made into eternal flesh. The ideas expressed by the file include the notion that human expression comes in severable chunks that can be organized…and need to be matched to compatible applications.
What do files mean to the future of human expression?
What indeed. We treat files as an inherent part of our daily computing reality. But a proper understand of metaphors shows us that the notion of ‘files’ is nothing but a useful construct – there is nothing ‘real’ about them. So the question is: is it time to get rid of the file as the principal – and limiting – building block of computer-based human expression?
Aside from the wider debates that Lanier raises, this issue alone warrants some attention and response and there is the seed of a response further in the book where he states that “Instead of collections of bits being offered as a product, they would be rendered as a service”. After last Friday’s Semantic Link podcast recording, I re-found that quote – it was obviously haunting my sub-conscious when I stated, about OpenData initiatives, that we should worry less about datasets being offered up as a product and concentrate more effort on providing public information as a service.
This is certainly the attraction of the Open Data and Open Gov movements – public information is ‘trapped’ in files, where data is difficult to access and reposition, analyse and reuse, and providing access directly to the source will liberate data from that straightjacket. Or so the story goes.
The file can certainly seem like a straightjacket and, with the advent of versioning, backups, synchronisation between different devices and other functions that lead to the plethora of different versions of any piece of work (see my post on the whole idea of versions as a social construct), files can be increasingly unmanageable and even inappropriate to the way in which we, as humans, collect our thoughts and ideas and present them for sharing with others.
Files do, however represent a very explicit, user defined context – a flow of words, images and ideas are put together in a particular way for a particular reason and at a particular time.
Some, less responsible elements within the OpenData movement, would seem to be arguing that it is both desirable and possible to simply unpackage all the boxes and strew the content on the floor, leaving anyone to pick up arbitrary pieces and put them back together as they see fit. Most would sensibly argue that this is a step in the wrong direction but if the tools and the exposed content allow it, who is to stop you? What guarantees are there that the data is always going to be understood in some context, whether as intended or implied initially or as explicitly claimed by a future masher-uper?
It would seem that this remains the single biggest challenge for the OpenData movement: to guarantee the persistence of relevant context when data is re-purposed, including the data’s provenance, reliability, relationship to a particular event, geographical scope, etc.