/[svn]/web/index.xml
ViewVC logotype

Diff of /web/index.xml

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 257 by abate, Tue Jul 10 17:19:53 2007 UTC revision 258 by abate, Tue Jul 10 17:19:59 2007 UTC
# Line 15  Line 15 
15  <box>  <box>
16  <ul>  <ul>
17  <li><a href="/cgi-bin/cduce">Online demo</a></li>  <li><a href="/cgi-bin/cduce">Online demo</a></li>
18    <li><a href="papers.html">Technical papers and talks</a></li>
19  <li><a href="team.html">CDuce team</a></li>  <li><a href="team.html">CDuce team</a></li>
20  <li><a href="plan.html">Site map</a></li>  <li><a href="plan.html">Site map</a></li>
21  <li><a href="bench.html">Site map</a></li>  <li><a href="bench.html">Site map</a></li>
# Line 24  Line 25 
25  <box>  <box>
26    
27  <p> CDuce ("seduce") is a new typed functional language with  <p> CDuce ("seduce") is a new typed functional language with
28  innovative features: </p>  innovative features.</p>
   
 <ul>  
 <li> a rich <b>type algebra</b>, with recursive types and arbitrary  
 boolean combinations (union, intersection, complement); </li>  
 <li> a natural notion of <b>subtyping</b>, which allows to use a value  
 of a given type where a value of a supertype is expected; </li>  
 <li> <b>overloaded functions</b> with late binding (dynamic dispatch);  
 </li>  
 <li> a powerful <b>pattern matching</b> operation, with dynamic  
 dispatch ontypes and recursive patterns.  </li>  
 </ul>  
   
29    
30  <p> Although CDuce is a general programming language, it features  <p> Although CDuce is a general programming language, it features
31  several characteristics that make it adapted to XML documents  several characteristics that make it adapted to XML documents
# Line 46  Line 35 
35  describe tree-like documents (aka semi-structured data), and XML  describe tree-like documents (aka semi-structured data), and XML
36  documents often come with a description of their type. The type is  documents often come with a description of their type. The type is
37  expressed in a system like DTD, or  expressed in a system like DTD, or
38  <a href="http://www.w3.org/XML/Schema">XML Schema</a>.  </p>  <a href="http://www.w3.org/XML/Schema">XML Schema</a>.
39    XML types play a central role in CDuce.
 <p> Our point of view and our guideline for the design of CDuce is  
 that a programming language for XML should take XML types into account  
 to allow: </p>  
   
 <ul>  
 <li> <b>static verifications</b> of properties for the applications  
 (for instance, ensuring that a transformation produces a document of  
 the expected type);</li>  
 <li> good <b>integration in a general purpose typed programming  
 language</b>;</li>  
 <li> static <b>optimizations</b> of applications and storage (knowing  
 the type of a document seems important to store and extract  
 information efficiently).</li>  
 </ul>  
 </box>  
   
 <box>  
   <p> <b>CDuce</b> is developped by the  
   <a href="http://www.di.ens.fr/~castagna/EQUIPE"><b>Languages</b></a>  
   group of ENS in Paris and the  
   <a href="http://www.lri.fr/bd"><b>Databases</b></a> group of LRI in  
   Orsay, two <a href="http://www.cnrs.fr">CNRS</a> labs.  
40    </p>    </p>
41    
42  </box>  </box>
43    
44  <box>  <box>
45  <p> This page was automatically generated from an XML description of  <p> This site was automatically generated from an XML description of
46  the content by a CDuce program.  the content by a CDuce program.
47  <img src="img/cducepower2.jpg" alt="Powered by CDuce"/></p>  <img src="img/cducepower2.jpg" alt="Powered by CDuce"/></p>
48  </box>  </box>
# Line 88  Line 56 
56    
57  <main>  <main>
58    
59  <box title="Give it a try!" link="proto">  <box title="What is CDuce ?" link="whatis">
60    
61    <p> <b>CDuce</b> is modern programming language, adapted to the
62    manipulation of XML documents.  It is developped by the <a
63    href="http://www.di.ens.fr/~castagna/EQUIPE"><b>Languages</b></a>
64    group of ENS in Paris and the <a
65    href="http://www.lri.fr/bd"><b>Databases</b></a> group of LRI in
66    Orsay, two <a href="http://www.cnrs.fr">CNRS</a> labs.
67    See also the <a href="team.html">CDuce team</a> page,
68    and our <a href="papers.html">technical papers</a>.
69    </p>
70    
71  <section title="Online running prototype">  <section title="Online running prototype">
72  <p> The only available implementation of CDuce is an <a  <p> The only available implementation of CDuce is an <a
# Line 96  Line 74 
74  you can try the examples and play with them, or have a look at this <a  you can try the examples and play with them, or have a look at this <a
75  href="memento.html">memento</a> which briefly explains the syntax of  href="memento.html">memento</a> which briefly explains the syntax of
76  the language.  </p>  the language.  </p>
 </section>  
77    
78    <p>We are planning to distribute a first public release in the next
79    few weeks.</p>
80    </section>
81  </box>  </box>
82    
83  <box title="Papers" link="papers">  <box title="Design and features" link="whatis">
84  <section title="Language Description">  <p> Our point of view and our guideline for the design of CDuce is
85  <ul>  that a programming language for XML should take XML types (
86    DTD, XML Schema, Relax-NG, ...) seriously. We expect the following
87  <li>  benefits:</p>
88  <paper file="papers/cduce-wp.ps.gz">  
89   <title>CDuce: a white paper</title>  <ul>
90   <author>V. Benzaken</author>  <li> <b>static verifications</b>
91   <author>G. Castagna</author>  (e.g.: ensure that a transformation produces a valid document);</li>
92   <author>A. Frisch</author>  <li> in particular, we aim at <b>smooth and safe</b> compositions
93   <comment>  of XML transformations, and incremental programming;</li>
94     A preliminary and abriged version of this work was presented at the workshop  <li> static <b>optimizations</b> and <b>efficient execution model</b>
95     <i>PLAN-X: Programming Language Technologies for XML</i>  (knowing the type of a document is crucial to extract information
96     Pittsburgh PA, Oct. 2002.  efficiently).</li>
  </comment>  
   
  <abstract> <p> In this paper, we present the functional language CDuce,  
  discuss some design issues, and show its adequacy for working with  
  XML documents. Peculiar features of CDuce are a powerful pattern  
  matching, first class functions, overloaded functions, a very rich  
  type system (arrows, sequences, pairs, records, intersections,  
  unions, differences), precise type inference and a natural  
  interpretation of types as sets of values. We also discuss how to add  
  constructs for programming XML queries in a declarative (and, thus,  
  optimizable) way and finally sketch a dispatch algorithm to  
  demonstrate how static type information can be used in efficient  
  compilation schemas.  </p> </abstract>  
 </paper>  
 </li>  
97  </ul>  </ul>
 </section>  
98    
99  <section title="Theoretical foundations">  <p>
100    Some of CDuce peculiar features:
101    </p>
102  <ul>  <ul>
103    <li> XML objects can be manipulated as first-class citizen values:
104  <li>       elements, sequences, tags, characters and strings, attribute
105  <paper file="papers/lics02.ps.gz">       sets; sequences of XML elements can be specified by
106   <title>Semantic subtyping</title>       <b>regular expressions</b>, which also apply to
107   <author>A. Frisch</author>       characters strings; </li>
108   <author>G. Castagna</author>  <li> functions themselves are <b>first-class</b> values, they
109   <author>V. Benzaken</author>       can be manipulated, stored in data structure, returned by
110   <comment> In <i>LICS 2002</i>. </comment>       a function,...</li>
111    <li> a powerful <b>pattern matching</b> operation can perform
112   <abstract> <p> Usually subtyping relations are defined either       complex extractions from sequences of XML elements; </li>
113   syntactically by a formal system or semantically by an interpretation  <li> a rich <b>type algebra</b>, with recursive types and arbitrary
114   of types in an untyped denotational model. In this work we show how       boolean combinations (union, intersection, complement) allows
115   to define a subtyping relation semantically, for a language whose       precise definitions of data structures and XML types;
116   <i>operational</i> semantics is <i>driven by types</i>; we consider a       <b>general purpose types</b> and types constructors are taken seriously
117   rich type algebra, with product, arrow, recursive, intersection,       (products, extensible records, arbitrary precision integers with interval
118   union and complement types. Our approach is to "bootstrap" the       constraints,  Unicode characters);</li>
119   subtyping relation through a notion of set-theoretic model <i>of the  <li> <b>polymorphism</b> through a natural notion of <b>subtyping</b>,
120   type algebra</i>.  </p>       and <b>overloaded functions</b> with dynamic dispatch; </li>
121    <li> an highly-effective <b>type-driven compilation schema</b>. </li>
  <p> The advantages of the semantic approach are manifold. Foremost we  
  get "for free" many properties (e.g., the transitivity of subtyping)  
  that, with axiomatized subtyping, would require tedious and error  
  prone proofs. Equally important is that the semantic approach allows  
  one to <i>derive</i> complete algorithms for the subtyping relation  
  or the propagation of types through patterns.  As the subtyping  
  relation has a natural (inasmuch as semantic) interpretation, the  
  type system can give informative error messages when static  
  type-checking fails. Last but not least the approach has an immediate  
  impact in the definition <i>and the implementation</i> of languages  
  manipulating XML documents, as this was our original motivation.  
  </p> </abstract>  
 </paper>  
 </li>  
   
 <li>  
 <paper file="papers/itrs02.ps.gz">  
  <title>The Relevance of Semantic Subtyping</title>  
  <author>M. Dezani-Ciancaglini</author>  
  <author>A. Frisch</author>  
  <author>E. Giovannetti</author>  
  <author>Y. Motohama</author>  
  <comment>  
    In <i>Electronic Notes in Theoretical Computer Science 70 No.1 (2002)</i>.  
  </comment>  
   
  <abstract> <p> We compare Meyer and Routley's minimal relevant logic  
  B+ with the recent semantics-based approach to subtyping introduced  
  by Frisch, Castagna and Benzaken in the definition of a type system  
  with intersection and union.  We show that - for the functional core  
  of the system - such notion of subtyping, which is defined in purely  
  set-theoretical terms, coincides with the relevant entailment of the  
  logic B+.  </p> </abstract>  
 </paper>  
 </li>  
   
 <li>  
 <paper file="papers/dea.ps.gz">  
  <title>Types récursifs, combinaisons booléennes et fonctions  
 surchargées: application au typage de XML</title>  
  <author>A. Frisch</author>  
  <comment>In french. Mémoire du DEA « Programmation: Sémantique,  
   Preuves et Langages » (Paris VII) </comment>  
   
  <abstract> <p> Nous étudions un lambda-calcul typé avec une opération  
  de filtrage qui permet d'exprimer des fonctions  
  surchargées. L'algèbre de types a des types de base, les types  
  produit et flèche, les types récursifs, les combinaisons booléennes  
  finies arbitraires.  Nous considérons une notion de sous-typage  
  sémantique, issue de l'interprétation des types comme ensembles de  
  valeurs du langage.  </p>  
   
  <p> Les caractéristiques présentes dans le calcul et l'algèbre de  
  types sont motivées par l'utilisation du calcul comme un noyau pour  
  un langage adapté à la manipulation de documents XML.  </p>  
   
  <p> Nous utilisons une approche sémantique pour étudier l'algèbre de  
  types, tout en conservant une preuve syntaxique de préservation du  
  typage par réduction. </p> </abstract>  
 </paper>  
 </li>  
122  </ul>  </ul>
 </section>  
123    
124  <section title="Security">  <p>
125    <a href="bench.html">Preliminary benchmarks</a> suggest that a CDuce
126    program can run faster (30% to 60%) than an equivalent XSLT
127    style-sheet (we performed benchmarks with
128    the xsltproc tools from the Gnome libxslt library).
129    </p>
130    </box>
131    
132    <box title="Research directions" link="research">
133    
134    <p>Our plans concerning the design of the core language
135    include:</p>
136  <ul>  <ul>
137  <li>   <li>a module system to support incremental programming;</li>
138  <paper file="papers/sec.ps.gz">   <li>parametric polymorphism;</li>
139   <title>Security analysis for XML transformations</title>   <li>XML-friendly primitives, to mimic XSLT transformations.</li>
  <author>M. Burelle</author>  
  <author>V. Benzaken</author>  
  <author>G. Castagna</author>  
  <comment> <a href="http://www.myths.eu.org">MyThS</a> technical  
    report. </comment>  
   
  <abstract> <p> In this article we give a formal definition of  
  information flows in the context of XML transformations and, more  
  generally, in the presence of type dependent computations. We  
  formalize a sound technique to detect XML document transformations  
  that may leak private or confidential information. By defining  
  security annotations and by relating various kind of analyses to  
  different query scenarios, we also establish a general framework for  
  checking middleware-located information flows.  </p> </abstract>  
 </paper>  
 </li>  
140  </ul>  </ul>
 </section>  
141    
142  </box>  <p>
143    Apart from the core language design and implementation,
144    our research projects include:
145    </p>
146    <ul>
147    <li> integration of a <b>query sub-language</b> into CDuce, using
148         types as a primary optimization strategy for request evaluation;</li>
149    <li> study of <b>security</b> (confidentiality, ...) properties in the
150         setting of XML transformations.</li>
151    </ul>
152    
153    <p>
154    We wrote several <a href="papers.html">technical papers</a> about
155    the language design and its theoretical foundations.
156    </p>
157    </box>
158    
159  <box title="Slides" link="slides">  <box title="XDuce and CDuce" link="xduce">
160  <ul>  <p>
161    
162  <li>  The starting point of our work on CDuce was the
163  <slides file="slides/lri02.pdf">  <a href="http://xduce.sourceforge.net/">XDuce</a> language developped
164    <title>Semantic Subtyping</title>  at the UPenn DB group. Many of CDuce features originate from XDuce.
165    <author>A. Frisch</author>  Some of our achievements:
   <author>G. Castagna</author>  
   <author>V. Benzaken</author>  
   <comment>Slides of a presentation given at LRI Database Group (2002-12)</comment>  
 </slides>  
 </li>  
   
 <li>  
 <slides file="slides/planx2002.pdf.gz">  
  <title>CDuce: a white paper</title>  
  <author>V. Benzaken</author>  
  <author>G. Castagna</author>  
  <author>A. Frisch</author>  
  <comment>Slides of the presentation given at the PLANX 2002 Workshop  
    in Pittsburgh  (2002-09)</comment>  
 </slides>  
 </li>  
   
 <li>  
 <slides  
  file="ftp://ftp.ens.fr/pub/di/users/castagna/SLIDES/cduce-wp-slides.pdf.gz">  
  <title>CDuce: a white paper</title>  
  <author>V. Benzaken</author>  
  <author>G. Castagna</author>  
  <author>A. Frisch</author>  
  <comment>Slides of the presentation given at MyThS meeting in  
    Bertinoro  (2002-12) </comment>  
 </slides>  
 </li>  
   
 <li>  
 <slides file="ftp://ftp.ens.fr/pub/di/users/castagna/SLIDES/cduce-sec.ps.gz">  
  <title>CDuce: document security</title>  
  <author>V. Benzaken</author>  
  <author>M. Burelle</author>  
  <author>G. Castagna</author>  
  <comment>Slides of the presentation given at FOSAD school (september  
 2002)  </comment>  
 </slides>  
 </li>  
166    
167  </ul>  </p>
168    <ul>
169    <li>integration of first-class and overloaded functions, arbitrary boolean
170        connectives, and extensible (or not) records, to the semantic
171        definition of subtyping;</li>
172    <li>a subtyping algorithm without backtracking;</li>
173    <li>extending pattern matching to capture non consecutive
174        subsequences; removing tail condition for exact matching
175        (they arrived independently to another solution);</li>
176    <li>efficient evaluation model that takes profit of static type information;</li>
177    </ul>
178    <p>
179    Of course, the work on XDuce continued during our, and they
180    developped nice ideas: mixed attribute-element types (same
181    expressive power as our records, but they can sometimes avoid exponential
182    explosion where we cannot); powerful filter operation.
183    </p>
184  </box>  </box>
185    
186  <box title="Related links" link="links">  <box title="Related links" link="links">
187  <ul>  <ul>
   
 <li> <link url="http://xduce.sourceforge.net/" title="XDuce"> The  
 starting point of our work on semantic subtyping and CDuce was XDuce.  
 CDuce extends XDuce with first-class and late-bound overloaded  
 functions, and generalizes the boolean connectives (explicit union,  
 intersection, negation types). </link> </li>  
   
188  <li> <link url="http://www.research.avayalabs.com/user/wadler/xml/"  <li> <link url="http://www.research.avayalabs.com/user/wadler/xml/"
189  title="XML: Some hyperlinks minus the hype"> By Philip Wadler. </link>  title="XML: Some hyperlinks minus the hype"> By Philip Wadler. </link>
190  </li> </ul> </box>  </li> </ul> </box>
Legend:
Removed from v.257  
changed lines
  Added in v.258
CVS Admin">CVS Admin
 ViewVC Help
Powered by ViewVC 1.1.5