Information Technology Reference
In-Depth Information
the dynamic nature of the software's changing mood, it was impossible to know
this mind, and people realised that some software is written not to be a tool, but to
be a creative individual. In fact, when in the most negative of moods, The Painting
Fool refused to paint a portrait and sent the (often shocked) sitter away, citing a
particularly depressing keyphrase in a particularly distressing newspaper article that
it had recently read.
In these cases, The Painting Fool pointed out explicitly: “No random numbers
were used in coming to this decision”. This is because we feel that accountable
unpredictability is important for creative systems. That is, we have found that when
people realise that a certain important event has happened or an important artefact has
been produced because of a random act, any dialogue (perceived or real) comes to an
abrupt halt, and detracts from the creative experience. In contrast, unpredictability
through accountable actions such as reading newspaper articles can add a great deal
to a creative experience, at the very least by providing additional talking points.
Hypothesis 5 Accountable unpredictability enhances the experience people have
when told about software creating an artefact, whereas random number based unpre-
dictability detracts from the experience.
1.5 Formally Capturing Progress in Creative Systems
Naturally, another major set of stakeholders in the notion of software being creative
are the Computational Creativity researchers who aim to write such systems, and use
them to study creativity in people and machines. As they are familiar with the issues
of simplistic arguments for and against creativity in software, these stakeholders
require more formalism in any argumentation put forward to support the hypothesis
of increased creativity in software.
We have focused on formalising the general notion of progress in Computational
Creativity research. To do this, we first introduced the FACE and IDEA descriptive
models in [ 45 , 46 ]. The FACE model categorises generative acts by software into
those at (g)round level, during which base objects are produced, and (p)rocess level,
during which methods for generating base objects are produced. These levels are
sub-divided by the types of objects/processes they produce: F g denotes a generative
act producing some framing information, A g denotes an act producing an aesthetic
measure, C g denotes an act producing a concept and E g denotes an act produc-
ing an example of a concept. Generative acts producing new processes are defined
accordingly as F p , A p , C p and E p . Tuples of generative acts are collated as creative
acts , and various calculations and recommendations are suggested in the model with
which to compare creative systems. We developed the IDEA model so that creative
acts and any impact they might have could be properly separated. We defined various
stages of software development and used an ideal audience notion, where people are
able to quantify changes in well-being and the cognitive work required to appreciate
a creative act and the resulting artefact and/or process.
Search WWH ::




Custom Search