deletions | additions       diff --git a/discussion.tex b/discussion.tex  index ed2e17f..60e5827 100644  --- a/discussion.tex  +++ b/discussion.tex 
...
\end{quote}  \subsection{Future Work} \label{sec:futurework}  \input{future-work}  diff --git a/future-work.tex b/future-work.tex  index a19ab8c..84417d9 100644  --- a/future-work.tex  +++ b/future-work.tex 
...
\section{Future work} \subsubsection{``Hatching'' new designs: Which came first?}  Whereas van Andel speaks of ``patterns of serendipity'' in a  relatively informal way, this paper will rely on the somewhat more  formal theory of \emph{design patterns} \cite{alexander1999origins},  to which it makes several additions and alterations. This theory is  by no means limited to computing, and indeed, has its origins in  architecture and urban planning. Our approach to ``designing for  serendipity'' \cite{andre2009discovery} centres on the use of design  patterns to capture the dynamic aspects of serendipitous situations.  The typical use of design patterns, since they were introduced by  Christopher Alexander  \cite{alexander1979timeless,alexander1977pattern}, is to prescribe as  well as to describe. Design patterns provide models \emph{for} as  well as models \emph{of} \cite[p. 93]{geertz1973interpretation}.  Thus, when Alexander describes the pattern \emph{A place to wait}, he  is telling readers that it is a good idea to consider building such  places when designing living spaces. In connection with our  understanding of serendipity as closely associated with deviations  from familiar patterns, the central concern in this paper is the way  in which \emph{new} patterns are formed.  For example, when Poincar\'e \citeyear{poincare1910creation} describes his  discovery of the existence of Fuchsian functions, he includes the  detail: ``contrary to my habit I took black coffee, I could not  sleep.'' This is much more interesting as part of a story about an  exceptional case of productive insomnia than it is as the broad  characterisation of a typical nightly sleep schedule. It might best  be described as a part of a ``situational pattern,'' with a title like  \emph{Change of pace}, rather than a ``behaviour pattern''; indeed, at  the level of behaviour, a \emph{Change of pace} is the exception to a  pattern! Nevertheless, along with Poincar\'e, we can recognize a  pattern at another level.  \subsubsection{Computational poetry}  Naturally, we hope to realise the Writers Workshop in one or more  suitable formats. Initial experiments with {\sf FloWr} are underway. 
...
$T^{\star}$. Here, an important guidepost for implementation is that  many outcomes will result in new patterns of behaviour that the system  can draw on in subsequent interactions.  diff --git a/introduction.tex b/introduction.tex  index bcf8f6d..22b0698 100644  --- a/introduction.tex  +++ b/introduction.tex 
...
on. In practice, we are likely to attribute \emph{value} to materials that  are useful, and \emph{creativity} to a person who puts materials to use in a  novel way.  % Although computational creativity is well studied in both theory and  practice, the role of \emph{serendipity} has often not been discussed  in this field -- even though serendipity has often played a role in  historical instances of scientific and technical creativity.  Many instances of \emph{serendipity} serendipity  centre on reevaluation. For example, a non-sticky ``superglue'' that no one was quite sure how to use turned out to be just the right ingredient for 3M's Post-it\texttrademark\ notes.  %  Serendipity is related, firstly, to deviations from familiar patterns, 
...
the role it has to play in the future development of computational  creativity.  Whereas van Andel speaks of ``patterns of serendipity'' in a  relatively informal way, this paper will rely on the somewhat more  formal theory of \emph{design patterns} \cite{alexander1999origins},  to which it makes several additions and alterations. This theory is  by no means limited to computing, and indeed, has its origins in  architecture and urban planning. Our approach to ``designing for  serendipity'' \cite{andre2009discovery} centres on the use of design  patterns to capture the dynamic aspects of serendipitous situations.  The typical use of design patterns, since they were introduced by  Christopher Alexander  \cite{alexander1979timeless,alexander1977pattern}, is to prescribe as  well as to describe. Design patterns provide models \emph{for} as  well as models \emph{of} \cite[p. 93]{geertz1973interpretation}.  Thus, when Alexander describes the pattern \emph{A place to wait}, he  is telling readers that it is a good idea to consider building such  places when designing living spaces. In connection with our  understanding of serendipity as closely associated with deviations  from familiar patterns, the central concern in this paper is the way  in which \emph{new} patterns are formed.  For example, when Poincar\'e \citeyear{poincare1910creation} describes his  discovery of the existence of Fuchsian functions, he includes the  detail: ``contrary to my habit I took black coffee, I could not  sleep.'' This is much more interesting as part of a story about an  exceptional case of productive insomnia than it is as the broad  characterisation of a typical nightly sleep schedule. It might best  be described as a part of a ``situational pattern,'' with a title like  \emph{Change of pace}, rather than a ``behaviour pattern''; indeed, at  the level of behaviour, a \emph{Change of pace} is the exception to a  pattern! Nevertheless, along with Poincar\'e, we can recognize a  pattern at another level.  The key idea in this paper is to computationally model situations  where emergence of this particular sort can happen.  %  It will take some work to get there, however. Section  \ref{sec:literature-review} develops 13 key criteria for the  evaluation of serendipity based on a review of several well-known  examples of serendipitous discoveries from human history. Section  \ref{sec:foundations} describes a working testbed for exploring  serendipitous computational discovery. In Section  \ref{sec:patterns-of-serendipity}, we apply our 13 criteria to analyse  several narrative ``patterns of serendipity'' collected by van Andel  \cite{van1994anatomy}. Section \ref{sec:patterns-of-serendipity} is  the theoretical core of the paper; here we give our interpretation of  the design pattern methodology. In Section  \ref{sec:computational-serendipity}, we focus on serendipity in a  computational context, condensing our criteria into an operational  definition, making our treatment of design patterns more concrete, and  proposing an experimental setup that we think will exhibit many of the  relevant features. In Section \ref{sec:related}, we examine related  work, and in Section \ref{sec:recommendations}, we advance our  recommendations for researchers working on computational creativity  (and serendipity).  diff --git a/literature.tex b/literature.tex  index d344836..d77a01c 100644  --- a/literature.tex  +++ b/literature.tex 
...
the genre of detective fiction, and arguably even on detective  work.]}  \subsection{Connections with our formal definition}  \subsubsection*{Key condition for serendipity}  diff --git a/outtakes/basic-examples.tex b/outtakes/basic-examples.tex  new file mode 100644  index 0000000..7aaad76  --- /dev/null  +++ b/outtakes/basic-examples.tex 
...
\begin{itemize}  \item The 17\textsuperscript{th} Century discovery that \emph{quinine} extracted from  the bark of South American cinchona trees could be used to treat and  prevent malaria -- building on a much earlier indigenous Quechua  discovery that the extract stops shivering.  \item Fleming's discovery of {\em penicillin}.\footnote{Merton and  Barber \citeyear{merton} state that the description of this discovery  was the first time that the word \emph{serendipity} was used without  inverted commas or accompanying definition.}  \item de Mestral's invention of {\em Velcro}\texttrademark\ following  the model presented by cockle-burs that stuck to his jacket while  out walking \cite[pp 220-222]{roberts}.  \item Arthur Fry's invention of sticky bookmarks (the prototype for  {\em Post-it}\texttrademark\ notes), using a weak glue developed by  his colleague, Spencer Silver \cite[p. 224]{roberts}.  \item Penzias and Wilson's discovery of the {\em echoes of the Big  Bang} \cite{singh2004big}.  \item Kekul\'e's dream-inspired discovery of the {\em structure of the  benzine ring} \cite[p. 21]{benfey}, cf. \cite[p. 77]{roberts}.  \item Charles Goodyear's invention of {\em vulcanised rubber}  \cite{goodyear1855gum}.  \item The {\em Rosetta Stone} was found by a soldier who was  demolishing a wall in order to clear ground for what was to be Fort  St. Julien \cite[pp. 109 - 111]{roberts}.  \end{itemize}  %% with similar characteristics: e.g. the invention of dry cleaning by  %% a professional dye-maker after his maid spilled kerosene on the  %% tablecloth, or the discovery of a marketable use for sildenafil  %% citrate (better known as {\em Viagra}\texttrademark) which had been  %% trialled as a heart medicine.  diff --git a/outtakes/caution.tex b/outtakes/caution.tex  new file mode 100644  index 0000000..5005fc9  --- /dev/null  +++ b/outtakes/caution.tex 
...
We begin with some words of caution.  %  Note that the classic examples of human serendipity tend to focus on  ground-breaking discoveries. In computational creativity, we have  learned that we must not aim to build systems which perform  domain-changing acts of creativity before we can build systems which  can perform everyday, mundane creativity (distinguished as ``big C''  and ``little c'' creativity.) Similarly, we should be prepared to  model ``little s'' serendipity before we are able to model ``big S''  serendipity. Furthermore, attempts to introduce serendipity into  computer systems may initially diminish artefact value.  %  A system which allowed itself to be derailed from a task at hand might  not achieve as much as one which maintains focus. A system that uses  a random search or that has its behaviour determined by environmental  conditions may be deemed less intelligent than one which follows  detailed, explicit, pre-programming.  %  To such arguments, we would respond that serendipity is not ``mere  chance'' -- the axes of sagacity (skills) and useful results  (recognised as such at least by the discoverer) are equally important.  diff --git a/outtakes/old-outline.tex b/outtakes/old-outline.tex  new file mode 100644  index 0000000..a1e1610  --- /dev/null  +++ b/outtakes/old-outline.tex 
...
The key idea in this paper is to computationally model situations  where emergence of this particular sort can happen.  %  It will take some work to get there, however. Section  \ref{sec:literature-review} develops 13 key criteria for the  evaluation of serendipity based on a review of several well-known  examples of serendipitous discoveries from human history. Section  \ref{sec:foundations} describes a working testbed for exploring  serendipitous computational discovery. In Section  \ref{sec:patterns-of-serendipity}, we apply our 13 criteria to analyse  several narrative ``patterns of serendipity'' collected by van Andel  \cite{van1994anatomy}. Section \ref{sec:patterns-of-serendipity} is  the theoretical core of the paper; here we give our interpretation of  the design pattern methodology. In Section  \ref{sec:computational-serendipity}, we focus on serendipity in a  computational context, condensing our criteria into an operational  definition, making our treatment of design patterns more concrete, and  proposing an experimental setup that we think will exhibit many of the  relevant features. In Section \ref{sec:related}, we examine related  work, and in Section \ref{sec:recommendations}, we advance our  recommendations for researchers working on computational creativity  (and serendipity).  diff --git a/related-work.tex b/related-work.tex  index 811579b..d78ee33 100644  --- a/related-work.tex  +++ b/related-work.tex 
...
\subsection{Related work} \label{sec:related}  Paul Andr{\'e} et al.~\cite{andre2009discovery} al.~\citeyear{andre2009discovery}  look at serendipity from a design point of view. They propose a two-part model, in which  what we might call chance+curiosity exposes the unexpected, and  sagacity+value is determined by another subsystem. This corresponds 
...
The issue of designing for serendipity has also been taken up by  Deborah Maxwell et al.~\cite{maxwell2012designing}, in their  description of a prototype of the SerenA {\sf SerenA}  system. This system is designed to support serendipitous discovery for its \emph{users}  \cite{forth2013serena}. The authors rely on a process-based model of  serendipity \cite{Makri2012,Makri2012a} that is derived from user 
...
reflection we may well ask ``what, exactly, are we looking for as  evidence of reflection?'' \cite{rodgers2002defining}. A detailed  answer derived from the classic work of John Dewey  \cite{dewey1997we} \citeyear{dewey1997we}  is explored in \cite{rodgers2002defining}.} SerenA {\sf SerenA}  is a system like the ones described by Andr{\'e} et al.~\cite{andre2009discovery}, in which the user is expected to have  the ``aha'' moment, and take the creative steps. The computer is  mainly used to facilitate this; and as indicated above this is usually  done by searching outside of the normal search parameters to engineer  potentially serendipitous (or at least pseudo-serendipitous)  encounters. Another earlier example of this sort of system is {\sf  Max}, created by Figueiredo and Campos \cite{Campos2002}. \citeyear{Campos2002}.  The user emailed {\sf Max} with a list of interests and {\sf Max} would find a  webpage that may be of interest to the user. Other search-related  examples support searching for analogies (\cite{Donoghue2002} and  \cite{Donoghue2012}) and content \cite{Iaquinta2008}.  In earlier joint work \cite{colton-assessingprogress},mentioned in  Section \ref{sec:foundations},  we presented a diagrammatic formalism for evaluating progress in computational creativity. It is useful to ask what serendipity would add to this formalism, and how the result compares with other attempts to formalise serendipity, notably Figueiredo and Campos's \citeyear{Figueiredo2001}  `Serendipity Equations' \cite{Figueiredo2001}. Equations'.  %  In \cite{stakeholder-groups-bookchapter}, we advanced several  hypotheses related to the development of the computational creativity  diff --git a/serendipity-in-computational-context.tex b/serendipity-in-computational-context.tex  index 1abfd73..7d163c0 100644  --- a/serendipity-in-computational-context.tex  +++ b/serendipity-in-computational-context.tex 
...
\section{Serendipity in a computational context} \label{sec:computational-serendipity}  We begin with some words of caution.  %  Note that the classic examples of human serendipity tend to focus on  ground-breaking discoveries. In computational creativity, we have  learned that we must not aim to build systems which perform  domain-changing acts of creativity before we can build systems which  can perform everyday, mundane creativity (distinguished as ``big C''  and ``little c'' creativity.) Similarly, we should be prepared to  model ``little s'' serendipity before we are able to model ``big S''  serendipity. Furthermore, attempts to introduce serendipity into  computer systems may initially diminish artefact value.  %  A system which allowed itself to be derailed from a task at hand might  not achieve as much as one which maintains focus. A system that uses  a random search or that has its behaviour determined by environmental  conditions may be deemed less intelligent than one which follows  detailed, explicit, pre-programming.  %  To such arguments, we would respond that serendipity is not ``mere  chance'' -- the axes of sagacity (skills) and useful results  (recognised as such at least by the discoverer) are equally important.  As Campbell says: ``Chance is fundamentally inimical to rationality,  whereas serendipity presupposes a smart mind'' \cite{campbell}. While it might not enhance, or may even diminish, results from a  computationally creative system which has been constructed with other  goals in mind, we believe that serendipity is both possible and useful  to model in future systems.  \subsection{Evaluation criteria} \label{sec:evaluation-criteria}  The 13 criteria from Section \ref{sec:characteristics}  specify the conditions and preconditions that are conducive to 
...
significant advance in the field of computational creativity. Since  the experience is about \emph{learning} rather than winning, there is  little motivation to ``game the system''  (cf. \cite{lenat1983eurisko}). \cite{lenat1983eurisko}.  \paragraph{Writers Workshop: Serendipity triggers.}