1.1 --- a/doc/WK/content.tex Tue Nov 06 08:56:24 2007 +0100
1.2 +++ b/doc/WK/content.tex Thu Nov 08 10:21:59 2007 +0100
1.3 @@ -23,7 +23,7 @@
1.4 \end{figure}
1.5
1.6 %\end{document}
1.7 -No complex syntax is necessary anymore. This is the way how users can be brought to that site. An further advantage is, that the quality of the articles increases if users find the information in an article which turns out to be incorrect, imprecise or simply lacking depth, they are able to correct these things. Obviously there is also a major disadvantage, that comes along with this kind of dwvelopment since everyone is now able to manipulate articles to the wrong side. You can easily write articles, which are poor quality and contain information that is incorrect. But in the majority of cases, they get overwritten very soon. Another example off web2.0 is \textit{flickr} which is a huge digital photo database.You can upload and manage your photos online. Also Assembla is a great tool, which is a Wiki which facilitates the managing of project. You can split your work in tasks and watch the progress in every phase of the project.%assembla zitieren GENAU !
1.8 +No complex syntax is necessary anymore. This is the way how users can be brought to that site. An further advantage is, that the quality of the articles increases if users find the information in an article which turns out to be incorrect, imprecise or simply lacking depth, they are able to correct these things. Obviously there is also a major disadvantage, that comes along with this kind of dwvelopment since everyone is now able to manipulate articles to the wrong side. You can easily write articles, which are poor quality and contain information that is incorrect. But in the majority of cases, they get overwritten very soon. Another example off web2.0 is \textit{flickr} which is a huge digital photo database.You can upload and manage your photos online. Also Assembla is a great tool, which is a Wiki which facilitates the managing of project. You can split your work in tasks and watch the progress in every phase of the project.\cite{assembla}
1.9 This
1.10 % .. watch the progress in every phaseof the project ..???? WIRKLICH ?
1.11 is the way we should go when we want to bring \sisac{} to the web. The users should be able to do their work interactively and they should like working in the system and learning some mathematics. Also the navigation in the system is an important point. If the user does a calculation and doesn't know why this works like this, \sisac{} have to make it clear to the user, by showing the theorems or rules which were used and also some examples.
1.12 @@ -91,7 +91,7 @@
1.13 % .. die Abkuerzung fuer LearningManagementSystem ist oben schon eingefuehrt !
1.14 works together closely with the LCMS,
1.15 %Learning Content Management System, ...BITTE IM FOLGENDEN IMMER KURZFORM...
1.16 -in this way as the LCMS manages the learning contents for each course and the LMS is responsible for the whole course management, meaning that the system for instance tells you which course the user has to do next. As stated above, the Learning Content Management System is able to make a User Management, which we will need for our school mode. There it must be possible that teachers has more rights than the students and there is one very important requirement for \sisac{} to the eLearning system: the teacher should be able to supervise the student while doing his calculations. This gives the teacher the possibility to see where the student has problems and which things the student is good in. One requirement to ensure the learning progress definitely is to start out with easy examples and make them more challenging after the student has solved them successfuly. An other requirement to the system is that the user should be able to do calculations that he has already done, and had no problems so far, because maybe later in a higher level he can not remember. Moreover it would be necessary to solve the calculation immediately.
1.17 +in this way as the LCMS manages the learning contents for each course and the LMS is responsible for the whole course management, meaning that the system for instance tells you which course the user has to do next. As stated above, the LCMS is able to make a User Management, which we will need for our school mode. There it must be possible that teachers has more rights than the students and there is one very important requirement for \sisac{} to the eLearning system: the teacher should be able to supervise the student while doing his calculations. This gives the teacher the possibility to see where the student has problems and which things the student is good in. One requirement to ensure the learning progress definitely is to start out with easy examples and make them more challenging after the student has solved them successfuly. An other requirement to the system is that the user should be able to do calculations that he has already done, and had no problems so far, because maybe later in a higher level he can not remember. Moreover it would be necessary to solve the calculation immediately.
1.18 If you are in search of a more detailed account on eLearning, simply consult \cite{Mason:E-Learning}.
1.19
1.20
1.21 @@ -167,7 +167,7 @@
1.22 \paragraph{Problem}
1.23 A Problem can be described by finding an output for a given input.
1.24
1.25 -If you are in search of a more detailed information about these terms, simply consult \cite{isac}.
1.26 +If you are in search of a more detailed information about these terms, simply consult \cite{isac-web}.
1.27
1.28 For each of these three dimensions is a button with the caption of the name of the dimension. If the users clicks for example on the \textless methods\textgreater button, \sisac{} opens a new window for this dimension of knowledge, and then the user can get the informations about \textit{methods} that he wants to know exactly. However, also clicking on another button opens a new window for this dimension of knowledge.
1.29 At Fig.\ref{browse-knowledge} on page p.\pageref{browse-knowledge} you can see an example you can get, if you browse through the dimension \textit{methods}.
1.30 @@ -205,6 +205,8 @@
1.31 %What we want to have is a system which brings the modern aspects like interactivity to the user. Therefor it is necessary to have a look at which technologies are useful for us and which concepts should be avoided.
1.32 %WN071031 Versuch, auf gleichem Platz mehr Information unterzubringen ...
1.33
1.34 +What we want is a system with all the modern aspects of Web2.0. Therefor it is necessary to to find some relevant technologies.
1.35 +
1.36 The issue of this thesis is to transfer \sisac s front-end from Java Swing (including Java RMI) to Web2.0 technology. The transfer concerns the presentation of highly structured math knowledge as well as highly dynamic interaction.
1.37
1.38 First we give a survey on relevant Web2.0 technologies and then we relate them to the requirements following from sect.\ref{exist-user-interface}.
1.39 @@ -212,12 +214,19 @@
1.40 \section{Survey on relevant Web2.0 technologies}
1.41
1.42 \subsection{Servlets}
1.43 -Because we use Java as a language on the server side
1.44 +What we have now is a Java Swing application which presents the math data to the user in a comfortable form. However, when transfer to the web we have to choose the language we will use on the server side. Possibilities are:
1.45 +\begin{enumeration}
1.46 +\item Active Server Pages (\textit{ASP})
1.47 +\item Perl
1.48 +\item Hypertext Preprocessor (\textit{PHP})
1.49 +\item JavaServer Pages (\textit{JSP})
1.50 +\end{enumeration}
1.51 +Because the existing front-end is written in Java it is near to use Java and JSP on the server side.
1.52 % WN071031 ... bitte verstaendlich machen und in Teilsaetze zerlegen ...
1.53 -programming the first idea to make it interactive is to write a servlet, process the data the user types in and write the data back in form of html code back to the output stream. This method won't be very nice because in each servlet there is a lots of code only for presenting the data to the user.
1.54 +First idea to make things interactive is to write a servlet, that processes the data the user types in and writes data back in form of html code. This method would not be very nice because in each servlet there is a lots of code only for presenting the data to the user.
1.55
1.56 -\subsection{Java server pages}
1.57 -An other possibility is using JavaServerPages and getting the request parameters inside of the JSP code. For demonstration have a look at the following example \begin{verbatim}
1.58 +\subsection{JSP}
1.59 +An other possibility is using \textit{JSP} and getting the request parameters inside of the \textit{JSP} code. For demonstration have a look at the following example \begin{verbatim}
1.60 1 <%@ page language="java" %>
1.61 2
1.62 3 <html>
1.63 @@ -231,49 +240,44 @@
1.64
1.65 \end{verbatim}
1.66
1.67 -If these parameters exist this example will work fine and would be easier and shorter than the first solution. But there is still a big disadvantage. It is possible to write code inside of a JSP and nowadays the requirements of a web application are huge and there are enough HTML tags inside one file. So imagine you write your additional Java code inside this file and no programmer will be able to overview all this code and the maintenance of such a site would be impossible.
1.68 +If these parameters exist this example will work fine and would be easier and shorter than the first solution. But there is still a big disadvantage. It is possible to write code inside of a \textit{JSP} and nowadays the requirements of a web application are huge and there are enough HTML tags inside one file. So imagine you write your additional Java code inside this file and no programmer will be able to overview all this code and the maintenance of such a site would be impossible.
1.69
1.70 \subsection{Java server faces}
1.71
1.72 -Because of these problems there are some web-frameworks which should make such things easier, for example struts
1.73 +Because of these problems there are some web-frameworks which should make such things easier, for example
1.74 %WN071031 ... bitte eine Referenz (www, article, book) oder weglassen --- die Technik, Hinweise auf genauere Information zu geben, ist eines der Erfalgsfaktoren unseres Wissenschaftsbetriebes !!!
1.75 -or JavaServer Faces
1.76 -(short JSP)
1.77 + JavaServer Faces (\textit{JSF})
1.78 %WN071031 ... Abkuerzungen gleich zu Beginn mit dem Abgekuerzten Begriff verbinden !
1.79 \cite{Marinscheck:jsf}.
1.80 %WN071031 ... Referenzen bitte gleich beim 1.Auftreten !
1.81
1.82 -A main thing of such frameworks is, that they are structured in form of the Model-View-Controller principle (MVC) \cite{Buschmann:Patterns}.
1.83 +A main thing of such a framework is, that they are structured in form of the Model-View-Controller principle (\textit{MVC}) \cite{Buschmann:Patterns}.
1.84 %WN071031 ... Referenz bitte beim 1.Auftreten; alles andere macht den typischen Leser nervoes ;-)
1.85
1.86 -This principle brings the advantage to the developer, to write lots of the layout in such a layout language like JSP and strictly seperates the functionality from the layout things as you can see in Fig.\ref{model2}. %on page p.\pageref{model2}.
1.87 +This principle brings the advantage to the developer, to write lots of the layout in such a layout language like \textit{JSP} and strictly seperates the functionality from the layout things as you can see in Fig.\ref{model2}. %on page p.\pageref{model2}.
1.88
1.89 \begin{figure} [htb]
1.90 \centerline{\psfig{
1.91 figure=myfig/model2.eps,width=8cm}}
1.92 % figure=myfig/model2}}
1.93 -\caption{\textit{MVC} when using JavaServer Faces}
1.94 +\caption{\textit{MVC} when using \textit{JSF}}
1.95 \label{model2}
1.96 \end{figure}
1.97
1.98 - So the result of this model is better for developers to write the code and also to do maintenance work inside of the code because there is less code inside of an JSP file.
1.99 -In \sisac{} project we will use JSP
1.100 + So the result of this model is better for developers to write the code and also to do maintenance work inside of the code because there is less code inside of an \textit{JSP} file.
1.101 +In \sisac{} project we will use \textit{JSF}
1.102 %WN071031 JavaServer Faces ... bei _einer_ Beschreibung bleiben (im Gegensatz zu Dichtungen, wo der Autor meist seine Vielseitigkeit im Worte-Finden beweist)
1.103 -because this framework has three main advantages against struts:
1.104 +because this framework has two big advantages:
1.105 \begin{enumerate}
1.106 \item User can choose the language for describing the site
1.107 \item JSF has some reusable standard components
1.108 -\item JSF are more up to date
1.109 -%WN071031 aber Moden koennen auch schnell wieder vergehen ?!?
1.110 \end{enumerate}
1.111
1.112 -%WN071030 das kommt hier zu spaet...Detailed information to the \textit{MVC} principle, have a look at \cite{Buschmann:Patterns}.
1.113 -
1.114 %WN071031 das ist einer der seltenen Faelle von Formatierung in latex ...
1.115 \noindent
1.116 Now I want to describe the three main parts of this \textit{MVC} on the server side. %WN071030 bitte die Formatierung von Begriffen wie {MVC} _einheitlich_: _immer_ italic, ODER _immer_ '' ODER _immer_ \tt{}, ...
1.117
1.118 -\paragraph{Managed-Beans}
1.119 +\paragraph{Managed-Beans}\ref{ExampleBean}
1.120 As you can see above, the data models are called Java-Beans and such a bean is a class in Java with some special properties:
1.121 \begin{itemize}
1.122 \item the class must be public
1.123 @@ -283,20 +287,20 @@
1.124 \end{itemize}
1.125
1.126 \paragraph{View}
1.127 -The view part holds (as for instance, in case of \sisac) the JavaServerPages but as I wrote before, with JSF the developer can choose the language for describing the sites.
1.128 +The view part holds (as for instance, in case of \sisac) the \textit{JSP}s but as I wrote before, with \textit{JSF} the developer can choose the language for describing the sites.
1.129
1.130 \paragraph{Servlet Controller}
1.131 -The servlet controller is necessary for generating and initializing the managed beans, for generating and activating the JSPs and for holding the servlets.
1.132 +The servlet controller is necessary for generating and initializing the managed beans, for generating and activating the \textit{JSP}s and for holding the servlets.
1.133
1.134 -But there is also a disadvantage when using JavaServer Faces, because when the applications gets more complex, the faces-config.xml increases very fast and it's difficult for the developer to find the things he wants. But for this problem exists a solution, with a so called {\it faces config editor} all the entries are sorted in a way to make the entries easy findable for the user. For example, all navigation rules can be found in one sub menu.
1.135 +But there is also a disadvantage when using \textit{JSF}, because when the applications gets more complex, the \textit{faces-config.xml} file increases very fast and it becomes difficult for the developer to find the things he wants. But for this problem exists a solution, with a so called {\it faces config editor} all the entries are sorted in a way to make the entries easy findable for the user. For example, all navigation rules can be found in one sub menu.
1.136
1.137 \subsection{Tabs and {\it dojo}'s instead of windows}
1.138
1.139 -What we want to have in isac, is one Browser tab
1.140 +What we want to have in \sisac{}, is one Browser Tab
1.141 %warum _B_rowser aber _t_ab (gross / klein-Schreibung) ?
1.142 - seperated into some subparts. One way to get this is to use the standard html frames. You get the look that you want to have, a tree2 component which is described in Sect.\ref{tree} inside one frame. The tree works as usual and everything is ok, but if you click at a leave in the tree, you may get a problem. And in our case of \sisac{} we definitely get this problem, because by clicking on such a leave which can represent a theory, we want to see the explanation in the neighbour frame and this is impossible with frames, because you have to tell the target and this didn't work in combination with the tree2 component.
1.143 + seperated into some subparts. One way to get this is to use the standard html frames. You get the look that you want to have, a tree2 component which is described in Sect.\ref{tree} inside one frame. The tree works as usual and everything is ok, but if you click at a leave in the tree, you may get a problem. And in our case of \sisac{} we definitely get this problem, because by clicking on such a leave which can represent a theory, we want to see the explanation in the neighbour frame and this is impossible with frames, because you have to tell the target and this did not work in combination with the tree2 component.
1.144
1.145 -So with the frames we are not satisfied anymore and we have to find an alternative. I found an other way to separate the window in a toolkit called dojo.You can see dojo in action at \cite{dojo}. Dojo doesn't use frames to separate a window, but it use div's. With this div's everything works fine with the tree2 component, because it isn't necessary to specify a target. This is my solution to the problem of separating a window, sub parts should be resizeable and the tree2 component works fine, as you can see in the \textit{isacweb-preparations} project in the cvs. There is a file named \textit{dojo-test.jsp}. In this file you can see, that in dojo normal div' s and usually used and everything we want works fine.
1.146 +So with the frames we are not satisfied anymore and we have to find an alternative. I found an other way to separate the window in a toolkit called dojo \ref{TreeBacker}.You can see dojo in action at \cite{dojo}. Dojo does not use standard frames to separate a window, but it use {\tt div}s. With this {\tt div}s everything works fine with the tree2 component, because it is not necessary to specify a target. This is my solution to the problem of separating a window, sub parts should be resizeable and the tree2 component works fine, as you can see in the \textit{isacweb-preparations} project in the cvs. There is a file named \textit{dojo-test.jsp}. In this file you can see, that in dojo normal {\tt div}s are used.
1.147
1.148 \section{Stepwise calculation}
1.149
1.150 @@ -330,6 +334,10 @@
1.151 \label{popup}
1.152 \end{figure}
1.153
1.154 +\subparagraph{Display in MyFaces tree2}
1.155 +One other possibility would be to present the calculation with the \textit{tree2} component which is described in more detail in Sect. \ref{tree}. However, when the user wants to get the intermediate steps we could send the id of this node to the server. The math engine then calculates the intermediate steps and sends them back. After that it is possible to insert subnodes after the node the user clicked on.
1.156 +At this point I have to say, that I have not tried out this method in a practical example, but finding one node by the id musst be possible if you can create the nodes with an id.
1.157 +
1.158 %WN071031 das Auf- und Zuklappen von Calculations (wie's in isac derzeit das Swing macht) kann sogar das alte html:
1.159 %Back, Ralph-Johan and Grundy, Jim and von Wright, Joakim
1.160 %Structured Calculational Proof
1.161 @@ -341,12 +349,11 @@
1.162 \section{Browse the mathematics knowledge}\label{tree}
1.163
1.164
1.165 -For the \sisac{} project JSF with it's basic components is not enough for presenting the data in the way to make everything clear to the user. However, an Austrian programmers team is working on a reference implementation since 2001 which has a lot more components than the JSF standard programme. This implementation is called \textit{MyFaces}
1.166 -%WN071031 \cite{}
1.167 +For the \sisac{} project JSF with it's basic components is not enough for presenting the data in the way to make everything clear to the user. However, an Austrian programmers team is working on a reference implementation since 2001 which has a lot more components than the JSF standard programme. This implementation is called \textit{MyFaces} \cite{apache}
1.168 and since 2004 they belong to the \textit{ Apache Software Foundation}.
1.169
1.170 A lot of people are still working on this reference implementation for providing new components and functionality to the developers. An additional advantage is that MyFaces is an open source software, which means that everyone is allowed to see the source, use the components and to expand the functionality in the way the developer needs the components.
1.171 - For browsing the knowledge in \sisac{} we basically need a tree component for presenting the data in a hierarchical form. If you take a look at the \textit{Apache} homepage \cite{apache} you will see, that there are different tree components for selection. For the \sisac{} project we need the \textit{tree2} component because you have more options for giving the tree a nice look, for example with images in front of the node as Fig.\ref{tree2} on page p.\pageref{tree2} shows.
1.172 + For browsing the knowledge in \sisac{} we basically need a tree component for presenting the data in a hierarchical form. If you take a look at the \textit{Apache} homepage you will see, that there are different tree components for selection \cite{apache}. For the \sisac{} project we need the \textit{tree2} component because you have more options for giving the tree a nice look, for example with images in front of the node as Fig.\ref{tree2} on page p.\pageref{tree2} shows.
1.173
1.174 \begin{figure} [htb]
1.175 \centerline{\psfig{
1.176 @@ -355,11 +362,11 @@
1.177 \label{tree2}
1.178 \end{figure}
1.179
1.180 - For a more detailed information on how to use the \textit{tree2} component consult the isacweb-preparations project and the classes \textit{TreeBacker}, which is for loading the data into the tree and the \textit{TreeBean}, which has the get and set methods for the Bean. An important file for the style of the tree is the \textit{tree2.jsp}, which specifies how the nodes look like and which method is called, by clicking on a leave.
1.181 + For a more detailed information on how to use the \textit{tree2} component consult \ref{TreeBacker}, which is for loading the data into the tree and the \ref{TreeBean}, which has the get and set methods for the Bean. An important file for the style of the tree is the \textit{tree2.jsp}, which specifies how the nodes look like and which method is called, by clicking on a leave.
1.182
1.183
1.184
1.185 -An other advantage is that this component gets the data, which is needed to build the tree out of a backing bean. And this fact has the advantage to load the data you want to present with the tree from a database or from a file.
1.186 +An other advantage is that this component gets the data, which is needed to build the tree out of a backing bean. And this fact has the advantage to load the data you want to present with the tree from a database or from a file.
1.187
1.188
1.189 For each dimension of the knowledge in the browser a separate dojo div is used to present it to the user. So the user has the possibility to search in the knowledge in the dimension he wants, and there would not be as many windows as in the existing \sisac{}. In this form I think the informations are more intuitive for the user and it will probably be less confusing to the user.
1.190 @@ -367,7 +374,7 @@
1.191
1.192 \section{Show context between calculation and knowledge}
1.193
1.194 -One problem we have in the browser is to show the context between the calculation and one dimension of knowledge. First of all there must be a button or something comparable, which establishes the context. One idea is to have three buttons-one for each dimension. When the context is enabled for instance to the theories, then the buttons to enable the context have to be invisible for the user. To disable the context, there must be a \textit{ToWorksheet} button in each knowledge dimension, but only the one which has the context must be visible. In the other subwindows the buttons shouldn't be available for the user.
1.195 +One problem we have in the browser is to show the context between the calculation and one dimension of knowledge. First of all there must be a button or something comparable, which establishes the context. One idea is to have three buttons-one for each dimension. When the context is enabled for instance to the theories, then the buttons to enable the context have to be invisible for the user. To disable the context, there must be a \textit{ToWorksheet} button in each knowledge dimension, but only the one which has the context must be visible. In the other subwindows the buttons shouldn not be available for the user.
1.196 Now we have the possibility to enable and disable the context between the calculation and every dimension of knowledge. But how should we make it clear to the user, which dimension the context has?
1.197 One possibility is just to enable the buttons in the dimension which has the context. An other one is to make a cover icon visible in the specified subwindow. But I think these things are not strong enough. So a better solution would be to have the same background colour in the \textit{Worksheet} and in the dimension which has the context. One example how this could look like is demonstrated by Fig.\ref{context2} on page p.\pageref{context2}.
1.198
1.199 @@ -383,13 +390,7 @@
1.200
1.201 \chapter{Conclusion}
1.202
1.203 -... careful selection of software components ...
1.204 -
1.205 -... tests using them showed:
1.206 -
1.207 -... highly structured math data can be represented ...
1.208 -
1.209 -... highly dynamic interaction during construction of calculations can be modelles ...
1.210 +Online learning has become an important aspect of the way education is delivered. So it is necessary to transfer \sisac{} to the web. The most important part by doing this is to select the software components carefully. In this part \textit{JSF} is the technology we want to use. There we have the possibility to present the highly structured math data in a way every user can work with. Also the criterion of dynamic interaction during construction of calculations can be fulfilled. In the appendix you can find some tests on the software components.
1.211
1.212
1.213 \begin{appendix}
1.214 @@ -415,15 +416,15 @@
1.215 These tests are briefly described and related to the discussion in Chap.\ref{isac-gui} as follows:
1.216
1.217 \begin{enumerate}
1.218 -\item \label{AjaxServlet} AjaxServlet
1.219 -\item \label{ExampleBean} ExampleBean
1.220 -\item \label{FacesRedirectFilter} FacesRedirectFilter
1.221 -\item \label{HealthPlanController} HealthPlanController
1.222 -\item \label{RegistrationBean} RegistrationBean
1.223 -\item \label{SuggestionBean} SuggestionBean
1.224 +\item \label{AjaxServlet} AjaxServlet: In the ajax-test.jsp file you can enter your name, which is send to the AjaxServlet and print out asynchron.
1.225 +\item \label{ExampleBean} ExampleBean: Shows a very simple example how Beans work. In example1.jsp the start function of the ExampleBean is called.
1.226 +\item \label{FacesRedirectFilter} FacesRedirectFilter: Redirects all requests sent to xxx.jsp to xxx.faces.
1.227 +\item \label{HealthPlanController} HealthPlanController: example3.jsp calles the function signup from the HealthPlanController. Randomly the user gets accepted or rejected.
1.228 +\item \label{RegistrationBean} RegistrationBean: Is called in the file example2.jsp. There you have to type in your email and password. If booth are correct you get an Ok, otherwise a warning.
1.229 +\item \label{SuggestionBean} SuggestionBean: If the user gets a warning from the example before, then he gets presented an example of a correct input which is generated in the Suggestion Utils file.
1.230 \item \label{SuggestionUtils} SuggestionUtils
1.231 -\item \label{TreeBacker} TreeBacker
1.232 -\item \label{TreeBean} TreeBean
1.233 +\item \label{TreeBacker} TreeBacker: Here you can find some faked data for an example tree2. The layout for the tree can be found in tree2.jsp. Dojo is used for separating the window. How this works can be found in dojo-test.jsp
1.234 +\item \label{TreeBean} TreeBean: Has the methods which are called, when clicking on a leaf.
1.235 \end{enumerate}
1.236
1.237 \end{appendix}
2.1 --- a/doc/bib/WK.bib Tue Nov 06 08:56:24 2007 +0100
2.2 +++ b/doc/bib/WK.bib Thu Nov 08 10:21:59 2007 +0100
2.3 @@ -48,8 +48,8 @@
2.4 publisher = {dpunkt.verlag},
2.5 }
2.6
2.7 -@Misc{isac,
2.8 - author = { {Institute for softwaretechnologie} },
2.9 - title = { The ISAC - Project for Transparent Software in Applied Mathematics},
2.10 - howpublished = {http://www.ist.tugraz.at/projects/isac/},
2.11 -}
2.12 \ No newline at end of file
2.13 +@Misc{assembla,
2.14 + title = {assembla -- Tools, Talent, and How-to for agile teams and rapid app development},
2.15 + howpublished = {http://www.assembla.com/},
2.16 + }
2.17 +