Any copy of a file can be defined as an instance of the file entity.
The main idea of Virtual Entity is that, any time a native file is
created and uploaded to the Net, it is possible to initiate a Soul for
this file. The Soul of a file, according to Ve’s imaginary, is the
combination of a set of Metadata plus editable space for information
interchange, not very dissimilar from a wiki. Digital entities are
subdivided is four main substances, that are Text, Audio, Video and
Image. All entities are either natural born analogue or natural born
digital.
An interesting process to be taken into account is ‘transmutation’:
what happens when content is mutating substance, or carrier, from
analogue to digital, and vice-versa? Substances are always immanent,
and an entity can be defined in accordance to more than one substance,
because this identification depends on the specific approach to
content, rather than on a supposed unique matter. A number of critical
examples and ambiguities can be analysed and tested. If software, as
source code, is considered text, an executable is something more
malicious, composed, an ‘organon’ creating a different functioning. Be
software a daemon, or a polumetis spirit, it is not a simple entity.
An important and problematic element of digital content management
is the organisation of Metadata, that is information about information.
Main issue is establishing a compatible form that is firm through time.
One critical aspect is that, if metadata are stored inside a file, when
information is updated on one copy of the file, all other copies are
not accessing the new data. Another problem is the lack of persistence
through successive codification, encoding, editing and transformation.
Storing metadata inside the Soul of a file, that is a database
separated from any instance and independent from a specific file
structure, is a technical proposal to overpass some of the limits the
use of metadata encountered so far. In such structure information would
be accessible from any instance of the file, because the repository of
metadata is always providing the latest version. The idea of dividing
entities in four groups is useful also in this: a core metadata set
listing a certain number of fields that are substance independent, is
combined to other fields that are substance specific. A possible
approach to define these basic elements of description can be the
following: metadata have to be, as much as possible, both machine and
human understandable. To be optimised, only a minimal amount of
necessary information is to be considered relevant; permanent and
global characteristics are preferred to local, non permanent methods of
description. Inspiring researches are the Latent Semantic Analysis on
one side, that focuses on the relationships, in vectorial semantics,
between a set of documents and the terms they contain, and the symbol
abstractor named Singular Value Decomposition, an algebraic method to
describe the peculiarity of an object.
