slides: write conclusion

doc/slides/slides.tex

@@ -39,7 +39,7 @@
         \item Narrative Knowledge
         \item Organizational Knowledge
     \end{itemize}
-  \item Each component should be formated, stored and accessible in a
+  \item Each component should be formatted, stored and accessible in a
     format optimized for the given kind of knowledge.
   \end{itemize}
 \end{frame}
@@ -82,7 +82,7 @@
 
   \vspace{2mm}
 
-  While the explorative part is probably the more interesting to talk
+  While the exploring part is probably the more interesting to talk
   about, the implementation probably has more practical use.
 \end{frame}
 
@@ -104,7 +104,7 @@
 \begin{frame}{Implementation: Components}
   \begin{itemize}
     \item \emph{TODO}: Add pretty picture.
-    \item Involved in the implemtation for \emph{ulo-storage} are the
+    \item Involved in the implementation for \emph{ulo-storage} are the
       following components.
       \begin{itemize}
         \item \emph{Collector:} Fetch XML~files from Git~repositories
@@ -122,7 +122,7 @@
 \begin{frame}{Implementation: Choice of Database}
   \begin{itemize}
     \item choice of db: triplet store for triplets is the only reasonable choice
-    \item fun fact: graphdb is more picky than other DBs
+    \item fun fact: GraphDB is more picky than other DBs
   \end{itemize}
 \end{frame}
 
@@ -177,11 +177,19 @@
 
 \subsection{How to Represent Algorithms}
 
-\begin{frame}{Applications and Questions: How to Represent Algorithms}
+\begin{frame}{Applications and Questions: Algorithms that Solve Problems}
   \begin{itemize}
-    \item query that wants to find algorithms that solve $NP$-complete
-      graph problems
-    \item how to represent algorithms and problems and programs; oh my
+    \item One query for a tetrapodal search system is the following:
+      ``Find algorithms that solve $NP$-hard graph
+      problems''~\cite{tetra}.
+      \begin{itemize}
+        \item Exploit \texttt{ulo:theorem} and \texttt{ulo:proof}?
+          Tempting but potentially very complicated.
+        \item Algorithms aren't programs! Programs implement
+          algorithms that solve problems.
+      \end{itemize}
+    \item This illustrates the difficulty in designing an ontology
+      (schema) that is both expressive and concise.
   \end{itemize}
 \end{frame}
 
@@ -191,9 +199,16 @@
 
 \begin{frame}{Summary}
   \begin{itemize}
-    \item {???}
-    \item {???}
-    \item {???}
+    \item Importing existing XML~exports into a queryable database is
+      pretty straightforward. It does introduce some interesting
+      questions.
+    \item \emph{Versioning} of data sets requires us to re-create
+      databases. This does introduce latency but is probably the only
+      way out as diffing huge data sets is not feasible.
+    \item Existing exports have ``holes'', they only use small-ish
+      subsets of~{ULO}. On the other hand, representing algorithms and
+      algorithmic problems might require us to extend~{ULO}? Maybe instead
+      of tetrapodal search we need $n$-podal search.
   \end{itemize}
 \end{frame}