More work on language CPs

choke-points.tex

@@ -333,9 +333,9 @@ \section{Language Features}
 
 \subsubsection{Description}
 
-\todo{description}
+\paragraph{Transitive edges.} TODO bound (3..5-style )or unbound
 
-For example, the condition that a certain \texttt{message} does not have a certain \texttt{tag} is represented in the graph as the absence of a \texttt{hasTag} edge between the two nodes. Thus, queries looking for cases where this condition is satisfied check for the (also known as negative application conditions or NACs in graph transformation literature).
+\paragraph{Negative edge conditions.} Some queries define \emph{negative edge conditions}. For example, the condition that a certain \texttt{message} does not have a certain \texttt{tag} is represented in the graph as the absence of a \texttt{hasTag} edge between the two nodes. Thus, queries looking for cases where this condition is satisfied check for the (also known as negative application conditions or NACs in graph transformation literature~\cite{DBLP:journals/fuin/HabelHT96}).
 
 \subsubsection{Language-specific notes}
 
@@ -345,58 +345,69 @@ \subsubsection{Language-specific notes}
 \item[Cypher.]
 Prior to Neo4j version 2.0, Cypher queries used a syntax such as the following:
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=cypher]
 MATCH (source)-[r?:someType]->(target)
 WHERE r IS NULL
 RETURN source
 \end{lstlisting}
+\end{minipage}
 
 In Neo4j 2.0, the \lstinline[language=cypher]{OPTIONAL MATCH} clause was introduced:\footnote{\url{https://dzone.com/articles/new-neo4j-optional}}
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=cypher]
 MATCH (source)
 OPTIONAL MATCH (source)-[r:someType]->(target)
 WHERE r IS NULL
 RETURN source
 \end{lstlisting}
+\end{minipage}
 
 However, the preferred method is to use a pattern condition in the \lstinline[language=cypher]{WHERE} clause:
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=cypher]
 MATCH (source)
 WHERE NOT (source)-[:someType]->(target)
 RETURN source
 \end{lstlisting}
+\end{minipage}
 
 \item[G-CORE.] TODO
 
 \item[SPARQL.]
 Prior to SPARQL 1.1, queries with negative edge conditions used the following syntax:
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=sparql]
 OPTIONAL {
   ?xRoute ?routeDefinition ?xSensor .
   FILTER (sameTerm(base:routeDefinition, routeDefinition))
 } .
 FILTER (!bound(?routeDefinition))
 \end{lstlisting}
+\end{minipage}
 
-Since SPARQL 1.1, the preferred method is using the \lstinline{NOT EXISTS} construct.
-
+Since SPARQL 1.1, the preferred method is using the \lstinline[language=sparql]{NOT EXISTS} construct.
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=sparql]
 ?xSensor rdf:type base:Sensor .
 ?xRoute base:switchPosition ?xSwitchPosition .
 FILTER NOT EXISTS { xRoute ?routeDefinition ?xSensor } .
 \end{lstlisting}
+\end{minipage}
 
 The \lstinline{MINUS} construct can also be used for defining a negative condition for a single edge.\footnote{For details, see the ``Relationship and differences between NOT EXISTS and MINUS'' section in the SPARQL 1.1 specification: \url{https://www.w3.org/TR/sparql11-query/\#neg-notexists-minus}}
 
+\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=sparql]
 ?xSensor rdf:type base:Sensor .
 ?xRoute base:switchPosition ?xSwitchPosition .
 MINUS { xRoute ?routeDefinition ?xSensor } .
 \end{lstlisting}
+\end{minipage}
 
 \end{description}
 
@@ -412,7 +423,7 @@ \subsubsection{Description}
 
 Aggregations on aggregations, filtering on aggregation results (similarly to the \lstinline{HAVING} keyword of SQL).
 
-
+TODO
 
 \subsubsection{Language-specific notes}
 
@@ -432,13 +443,19 @@ \subsubsection{Language-specific notes}
 
 \subsubsection{Description}
 
-Selecting top-k with other graph elements/attributes.
+Selecting top-k with other graph elements/attributes, also known as \emph{windowing} or \emph{ranking}.
 
+PostgreSQL defines the \lstinline[language=sql]{rank()} function for this.\footnote{\url{https://www.postgresql.org/docs/9.1/static/tutorial-window.html}}
+
+TODO
 
 \subsubsection{Language-specific notes}
 
-\begin{description}
-\item[Cypher.] Ranking can be expressed in Cypher as a sequence of ordering, collecting results and taking the top-k values of the result list.
+%\begin{description}
+%\item[Cypher.]
+Ranking can be expressed in Cypher as a sequence of ordering, collecting results and taking the top-k values of the result list.
+
+\noindent\begin{minipage}{\linewidth}
 \begin{lstlisting}[language=cypher]
 ...
 WITH x, ...
@@ -447,11 +464,12 @@ \subsubsection{Language-specific notes}
 WITH xs[0] AS top, xs[0..5] AS top5
 ...
 \end{lstlisting}
+\end{minipage}
 
-\item[G-CORE.] TODO
-
-\item[SPARQL.] TODO
-\end{description}
+%\item[G-CORE.] TODO
+%
+%\item[SPARQL.] TODO
+%\end{description}
 
 \input{query-cards/cp-8-3}
 

ldbc-snb.bib

@@ -298,3 +298,20 @@ @inproceedings{GCORE_SIGMOD
         adsurl = {http://adsabs.harvard.edu/abs/2017arXiv171201550A},
         adsnote = {Provided by the SAO/NASA Astrophysics Data System}
 }
+
+@article{DBLP:journals/fuin/HabelHT96,
+  author    = {Annegret Habel and
+               Reiko Heckel and
+               Gabriele Taentzer},
+  title     = {Graph Grammars with Negative Application Conditions},
+  journal   = {Fundam. Inform.},
+  volume    = {26},
+  number    = {3/4},
+  pages     = {287--313},
+  year      = {1996},
+  url       = {https://doi.org/10.3233/FI-1996-263404},
+  doi       = {10.3233/FI-1996-263404},
+  timestamp = {Sat, 20 May 2017 00:23:06 +0200},
+  biburl    = {https://dblp.org/rec/bib/journals/fuin/HabelHT96},
+  bibsource = {dblp computer science bibliography, https://dblp.org}
+}

listings.tex

@@ -17,8 +17,6 @@
 	stringstyle=\scriptsize,
 	%
 	showstringspaces=false,
-	aboveskip=3pt,
-	belowskip=3pt,
 	columns=flexible,
 	keepspaces=true,
 	breaklines=true,