ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/trunk/tengDissertation/Appendix.tex
(Generate patch)

Comparing trunk/tengDissertation/Appendix.tex (file contents):
Revision 2821 by tim, Thu Jun 8 06:39:37 2006 UTC vs.
Revision 2839 by tim, Fri Jun 9 02:41:58 2006 UTC

# Line 113 | Line 113 | OOPSE}\cite{Meineke05} and PROTOMOL\cite{Matthey05} \t
113   As one of the latest advanced techniques emerged from
114   object-oriented community, design patterns were applied in some of
115   the modern scientific software applications, such as JMol, {\sc
116 < OOPSE}\cite{Meineke05} and PROTOMOL\cite{Matthey05} \textit{etc}.
117 < The following sections enumerates some of the patterns used in {\sc
118 < OOPSE}.
116 > OOPSE}\cite{Meineke2005} and PROTOMOL\cite{Matthey2005}
117 > \textit{etc}. The following sections enumerates some of the patterns
118 > used in {\sc OOPSE}.
119  
120   \subsection{\label{appendixSection:singleton}Singleton}
121 +
122   The Singleton pattern not only provides a mechanism to restrict
123   instantiation of a class to one object, but also provides a global
124   point of access to the object. Currently implemented as a global
# Line 127 | Line 128 | static data approach in {\sc OOPSE}. {\tt IntegratorFa
128   pollution.Although the singleton pattern can be implemented in
129   various ways  to account for different aspects of the software
130   designs, such as lifespan control \textit{etc}, we only use the
131 < static data approach in {\sc OOPSE}. {\tt IntegratorFactory} class
132 < is declared as
132 < \begin{lstlisting}[float,caption={[A classic Singleton design pattern implementation(I)] Declaration of {\tt IntegratorFactory} class.},label={appendixScheme:singletonDeclaration}]
131 > static data approach in {\sc OOPSE}. IntegratorFactory class is
132 > declared as
133  
134 <  class IntegratorFactory {
135 <    public:
136 <      static IntegratorFactory* getInstance();
137 <    protected:
138 <      IntegratorFactory();
139 <    private:
140 <      static IntegratorFactory* instance_;
141 <  };
134 > \begin{lstlisting}[float,caption={[A classic Singleton design pattern implementation(I)] The declaration of of simple Singleton pattern.},label={appendixScheme:singletonDeclaration}]
135 >
136 > class IntegratorFactory {
137 > public:
138 >  static IntegratorFactory*
139 >  getInstance();
140 > protected:
141 >  IntegratorFactory();
142 > private:
143 >  static IntegratorFactory* instance_;
144 > };
145 >
146   \end{lstlisting}
147 +
148   The corresponding implementation is
144 \begin{lstlisting}[float,caption={[A classic Singleton design pattern implementation(II)] Implementation of {\tt IntegratorFactory} class.},label={appendixScheme:singletonImplementation}]
149  
150 + \begin{lstlisting}[float,caption={[A classic implementation of Singleton design pattern (II)] The implementation of simple Singleton pattern.},label={appendixScheme:singletonImplementation}]
151 +
152   IntegratorFactory::instance_ = NULL;
153  
154   IntegratorFactory* getInstance() {
# Line 151 | Line 157 | IntegratorFactory* getInstance() {
157    }
158    return instance_;
159   }
160 +
161   \end{lstlisting}
155 Since constructor is declared as {\tt protected}, a client can not
156 instantiate {\tt IntegratorFactory} directly. Moreover, since the
157 member function {\tt getInstance} serves as the only entry of access
158 to {\tt IntegratorFactory}, this approach fulfills the basic
159 requirement, a single instance. Another consequence of this approach
160 is the automatic destruction since static data are destroyed upon
161 program termination.
162  
163 + Since constructor is declared as protected, a client can not
164 + instantiate IntegratorFactory directly. Moreover, since the member
165 + function getInstance serves as the only entry of access to
166 + IntegratorFactory, this approach fulfills the basic requirement, a
167 + single instance. Another consequence of this approach is the
168 + automatic destruction since static data are destroyed upon program
169 + termination.
170 +
171   \subsection{\label{appendixSection:factoryMethod}Factory Method}
172  
173   Categoried as a creational pattern, the Factory Method pattern deals
174   with the problem of creating objects without specifying the exact
175   class of object that will be created. Factory Method is typically
176   implemented by delegating the creation operation to the subclasses.
177 < \begin{lstlisting}[float,caption={[].},label={appendixScheme:factoryDeclaration}]
178 <  class IntegratorCreator;
179 <  class IntegratorFactory {
180 <    public:
181 <      typedef std::map<std::string, IntegratorCreator*> CreatorMapType;
177 > Parameterized Factory pattern where factory method (
178 > createIntegrator member function) creates products based on the
179 > identifier (see List.~\ref{appendixScheme:factoryDeclaration}). If
180 > the identifier has been already registered, the factory method will
181 > invoke the corresponding creator (see List.~\ref{integratorCreator})
182 > which utilizes the modern C++ template technique to avoid excess
183 > subclassing.
184  
185 <      /**
176 <       * Registers a creator with a type identifier
177 <       * @return true if registration is successful, otherwise return false
178 <       * @id the identification of the concrete object
179 <       * @creator the object responsible to create the concrete object
180 <       */
181 <      bool registerIntegrator(IntegratorCreator* creator);
185 > \begin{lstlisting}[float,caption={[The implementation of Parameterized Factory pattern (I)]Source code of IntegratorFactory class.},label={appendixScheme:factoryDeclaration}]
186  
187 <      /**
188 <       * Looks up the type identifier in the internal map. If it is found, it invokes the
189 <       * corresponding creator for the type identifier and returns its result.
186 <       * @return a pointer of the concrete object, return NULL if no creator is registed for
187 <       * creating this concrete object
188 <       * @param id the identification of the concrete object
189 <       */
190 <      Integrator* createIntegrator(const std::string& id, SimInfo* info);
187 > class IntegratorFactory {
188 > public:
189 >  typedef std::map<string, IntegratorCreator*> CreatorMapType;
190  
191 <    private:
192 <      CreatorMapType creatorMap_;
194 <  };
195 < \end{lstlisting}
196 <
197 < \begin{lstlisting}[float,caption={[].},label={appendixScheme:factoryDeclarationImplementation}]
198 <  bool IntegratorFactory::unregisterIntegrator(const std::string& id) {
199 <    return creatorMap_.erase(id) == 1;
191 >  bool registerIntegrator(IntegratorCreator* creator) {
192 >    return creatorMap_.insert(creator->getIdent(), creator).second;
193    }
194  
195 <  Integrator* IntegratorFactory::createIntegrator(const std::string& id, SimInfo* info) {
195 >  Integrator* createIntegrator(const string& id, SimInfo* info) {
196 >    Integrator* result = NULL;
197      CreatorMapType::iterator i = creatorMap_.find(id);
198      if (i != creatorMap_.end()) {
199 <      //invoke functor to create object
206 <      return (i->second)->create(info);
207 <    } else {
208 <      return NULL;
199 >      result = (i->second)->create(info);
200      }
201 +    return result;
202    }
203 +
204 + private:
205 +  CreatorMapType creatorMap_;
206 + };
207   \end{lstlisting}
208  
209 < \begin{lstlisting}[float,caption={[].},label={appendixScheme:integratorCreator}]
209 > \begin{lstlisting}[float,caption={[The implementation of Parameterized Factory pattern (III)]Source code of creator classes.},label={appendixScheme:integratorCreator}]
210  
211 <  class IntegratorCreator {
212 <  public:
213 <    IntegratorCreator(const std::string& ident) : ident_(ident) {}
218 <    virtual ~IntegratorCreator() {}
219 <    const std::string& getIdent() const { return ident_; }
211 > class IntegratorCreator {
212 > public:
213 >    IntegratorCreator(const string& ident) : ident_(ident) {}
214  
215 +    const string& getIdent() const { return ident_; }
216 +
217      virtual Integrator* create(SimInfo* info) const = 0;
218  
219 <  private:
220 <    std::string ident_;
221 <  };
219 > private:
220 >    string ident_;
221 > };
222  
223 <  template<class ConcreteIntegrator>
224 <  class IntegratorBuilder : public IntegratorCreator {
225 <  public:
226 <    IntegratorBuilder(const std::string& ident) : IntegratorCreator(ident) {}
227 <    virtual  Integrator* create(SimInfo* info) const {return new ConcreteIntegrator(info);}
228 <  };
223 > template<class ConcreteIntegrator>
224 > class IntegratorBuilder : public IntegratorCreator {
225 > public:
226 >  IntegratorBuilder(const string& ident)
227 >                   : IntegratorCreator(ident) {}
228 >  virtual  Integrator* create(SimInfo* info) const {
229 >    return new ConcreteIntegrator(info);
230 >  }
231 > };
232   \end{lstlisting}
233  
234   \subsection{\label{appendixSection:visitorPattern}Visitor}
235  
236 < The purpose of the Visitor Pattern is to encapsulate an operation
237 < that you want to perform on the elements of a data structure. In
238 < this way, you can change the operation being performed on a
239 < structure without the need of changing the class heirarchy of the
240 < elements that you are operating on.
236 > The visitor pattern is designed to decouple the data structure and
237 > algorithms used upon them by collecting related operation from
238 > element classes into other visitor classes, which is equivalent to
239 > adding virtual functions into a set of classes without modifying
240 > their interfaces. Fig.~\ref{appendixFig:visitorUML} demonstrates the
241 > structure of Visitor pattern which is used extensively in {\tt
242 > Dump2XYZ}. In order to convert an OOPSE dump file, a series of
243 > distinct operations are performed on different StuntDoubles (See the
244 > class hierarchy in Fig.~\ref{oopseFig:hierarchy} and the declaration
245 > in List.~\ref{appendixScheme:element}). Since the hierarchies
246 > remains stable, it is easy to define a visit operation (see
247 > List.~\ref{appendixScheme:visitor}) for each class of StuntDouble.
248 > Note that using Composite pattern\cite{Gamma1994}, CompositVisitor
249 > manages a priority visitor list and handles the execution of every
250 > visitor in the priority list on different StuntDoubles.
251  
252 < \begin{lstlisting}[float,caption={[].},label={appendixScheme:visitor}]
253 <  class BaseVisitor{
254 <    public:
255 <      virtual void visit(Atom* atom);
256 <      virtual void visit(DirectionalAtom* datom);
257 <      virtual void visit(RigidBody* rb);
258 <  };
252 > \begin{figure}
253 > \centering
254 > \includegraphics[width=\linewidth]{visitor.eps}
255 > \caption[The UML class diagram of Visitor patten] {The UML class
256 > diagram of Visitor patten.} \label{appendixFig:visitorUML}
257 > \end{figure}
258 >
259 > \begin{figure}
260 > \centering
261 > \includegraphics[width=\linewidth]{hierarchy.eps}
262 > \caption[Class hierarchy for ojects in {\sc OOPSE}]{ A diagram of
263 > the class hierarchy. } \label{oopseFig:hierarchy}
264 > \end{figure}
265 >
266 > \begin{lstlisting}[float,caption={[The implementation of Visitor pattern (II)]Source code of the element classes.},label={appendixScheme:element}]
267 >
268 > class StuntDouble { public:
269 >  virtual void accept(BaseVisitor* v) = 0;
270 > };
271 >
272 > class Atom: public StuntDouble { public:
273 >  virtual void accept{BaseVisitor* v*} {
274 >    v->visit(this);
275 >  }
276 > };
277 >
278 > class DirectionalAtom: public Atom { public:
279 >  virtual void accept{BaseVisitor* v*} {
280 >    v->visit(this);
281 >  }
282 > };
283 >
284 > class RigidBody: public StuntDouble { public:
285 >  virtual void accept{BaseVisitor* v*} {
286 >    v->visit(this);
287 >  }
288 > };
289 >
290   \end{lstlisting}
251 \begin{lstlisting}[float,caption={[].},label={appendixScheme:element}]
252  class StuntDouble {
253    public:
254      virtual void accept(BaseVisitor* v) = 0;
255  };
291  
292 <  class Atom: public StuntDouble {
258 <    public:
259 <      virtual void accept{BaseVisitor* v*} {v->visit(this);}
260 <  };
292 > \begin{lstlisting}[float,caption={[The implementation of Visitor pattern (I)]Source code of the visitor classes.},label={appendixScheme:visitor}]
293  
294 <  class DirectionalAtom: public Atom {
295 <    public:
296 <      virtual void accept{BaseVisitor* v*} {v->visit(this);}
297 <  };
294 > class BaseVisitor{
295 > public:
296 >  virtual void visit(Atom* atom);
297 >  virtual void visit(DirectionalAtom* datom);
298 >  virtual void visit(RigidBody* rb);
299 > };
300  
301 <  class RigidBody: public StuntDouble {
302 <    public:
303 <      virtual void accept{BaseVisitor* v*} {v->visit(this);}
304 <  };
301 > class BaseAtomVisitor:public BaseVisitor{ public:
302 >  virtual void visit(Atom* atom);
303 >  virtual void visit(DirectionalAtom* datom);
304 >  virtual void visit(RigidBody* rb);
305 > };
306 >
307 > class SSDAtomVisitor:public BaseAtomVisitor{ public:
308 >  virtual void visit(Atom* atom);
309 >  virtual void visit(DirectionalAtom* datom);
310 >  virtual void visit(RigidBody* rb);
311 > };
312 >
313 > class CompositeVisitor: public BaseVisitor {
314 > public:
315 >
316 >  typedef list<pair<BaseVisitor*, int> > VistorListType;
317 >  typedef VistorListType::iterator VisitorListIterator;
318 >  virtual void visit(Atom* atom) {
319 >    VisitorListIterator i;
320 >    BaseVisitor* curVisitor;
321 >    for(i = visitorList.begin();i != visitorList.end();++i) {
322 >      atom->accept(*i);
323 >    }
324 >  }
325  
326 +  virtual void visit(DirectionalAtom* datom) {
327 +    VisitorListIterator i;
328 +    BaseVisitor* curVisitor;
329 +    for(i = visitorList.begin();i != visitorList.end();++i) {
330 +      atom->accept(*i);
331 +    }
332 +  }
333 +
334 +  virtual void visit(RigidBody* rb) {
335 +    VisitorListIterator i;
336 +    std::vector<Atom*> myAtoms;
337 +    std::vector<Atom*>::iterator ai;
338 +    myAtoms = rb->getAtoms();
339 +    for(i = visitorList.begin();i != visitorList.end();++i) {{
340 +      rb->accept(*i);
341 +      for(ai = myAtoms.begin(); ai != myAtoms.end(); ++ai){
342 +        (*ai)->accept(*i);
343 +    }
344 +  }
345 +
346 +  void addVisitor(BaseVisitor* v, int priority);
347 +
348 +  protected:
349 +    VistorListType visitorList;
350 + };
351 +
352   \end{lstlisting}
353 +
354   \section{\label{appendixSection:concepts}Concepts}
355  
356   OOPSE manipulates both traditional atoms as well as some objects
357   that {\it behave like atoms}.  These objects can be rigid
358   collections of atoms or atoms which have orientational degrees of
359 < freedom.  Here is a diagram of the class heirarchy:
360 <
361 < %\begin{figure}
362 < %\centering
363 < %\includegraphics[width=3in]{heirarchy.eps}
364 < %\caption[Class heirarchy for StuntDoubles in {\sc oopse}-3.0]{ \\
365 < %The class heirarchy of StuntDoubles in {\sc oopse}-3.0. The
366 < %selection syntax allows the user to select any of the objects that
367 < %are descended from a StuntDouble.} \label{oopseFig:heirarchy}
287 < %\end{figure}
288 <
359 > freedom.  A diagram of the class hierarchy is illustrated in
360 > Fig.~\ref{oopseFig:hierarchy}. Every Molecule, Atom and
361 > DirectionalAtom in {\sc OOPSE} have their own names which are
362 > specified in the {\tt .md} file. In contrast, RigidBodies are
363 > denoted by their membership and index inside a particular molecule:
364 > [MoleculeName]\_RB\_[index] (the contents inside the brackets depend
365 > on the specifics of the simulation). The names of rigid bodies are
366 > generated automatically. For example, the name of the first rigid
367 > body in a DMPC molecule is DMPC\_RB\_0.
368   \begin{itemize}
369   \item A {\bf StuntDouble} is {\it any} object that can be manipulated by the
370   integrators and minimizers.
# Line 294 | Line 373 | DirectionalAtom}s which behaves as a single unit.
373   \item A {\bf RigidBody} is a collection of {\bf Atom}s or {\bf
374   DirectionalAtom}s which behaves as a single unit.
375   \end{itemize}
297
298 Every Molecule, Atom and DirectionalAtom in {\sc OOPSE} have their
299 own names which are specified in the {\tt .md} file. In contrast,
300 RigidBodies are denoted by their membership and index inside a
301 particular molecule: [MoleculeName]\_RB\_[index] (the contents
302 inside the brackets depend on the specifics of the simulation). The
303 names of rigid bodies are generated automatically. For example, the
304 name of the first rigid body in a DMPC molecule is DMPC\_RB\_0.
376  
377   \section{\label{appendixSection:syntax}Syntax of the Select Command}
378  
379 < The most general form of the select command is: {\tt select {\it
380 < expression}}. This expression represents an arbitrary set of
310 < StuntDoubles (Atoms or RigidBodies) in {\sc OOPSE}. Expressions are
311 < composed of either name expressions, index expressions, predefined
312 < sets, user-defined expressions, comparison operators, within
313 < expressions, or logical combinations of the above expression types.
314 < Expressions can be combined using parentheses and the Boolean
315 < operators.
379 > {\sc OOPSE} provides a powerful selection utility to select
380 > StuntDoubles. The most general form of the select command is:
381  
382 + {\tt select {\it expression}}.
383 +
384 + This expression represents an arbitrary set of StuntDoubles (Atoms
385 + or RigidBodies) in {\sc OOPSE}. Expressions are composed of either
386 + name expressions, index expressions, predefined sets, user-defined
387 + expressions, comparison operators, within expressions, or logical
388 + combinations of the above expression types. Expressions can be
389 + combined using parentheses and the Boolean operators.
390 +
391   \subsection{\label{appendixSection:logical}Logical expressions}
392  
393   The logical operators allow complex queries to be constructed out of

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines