120 likes | 251 Views
The Strudel Web Site Management System. Mary Fernandez, Daniela Florescu, Jaewoo Kang, Alon Levy, Dan Suciu. What is Strudel?. A yummy pastry made of a thin sheet of dough filled with fruit or cheese.
E N D
The Strudel Web Site Management System Mary Fernandez, Daniela Florescu, Jaewoo Kang, Alon Levy, Dan Suciu
What is Strudel? • A yummy pastry made of a thin sheet of dough filled with fruit or cheese. • A web site management system that allows a site’s content, structure and appearance to be managed independently.
Strudel in a Nutshell • Semi-structured data • Keep information in a database (data graph) • Write queries to create structure of Web site (site graph) • Same representation used for both graphs • For each object in site graph, use template to generate HTML
Why Do We Need Strudel? • HTML sucks
Site restructuring • Tailored pages • Data integration (sort of)
Hasn’t this been done before? • Front Page • Cascaded Style Sheets • XML • CGI-bin scripts • commercial database systems
The Data Graph • Contains collections of objects with labeled edges • Objects are either nodes with id’sor atomic Faculty fac4 fac15 “name” “research” “name” “research” “research” “Oren Etzioni” “info retrieval” “Alon Levy” “databases” “info retrieval”
StruQL - The Query Language • Site TRansformation Und Query Language • Used to query and create new graphs (data graphs or site graphs) • Flexible (regexp’s, nested clauses)
Building a Site Graph WHERE Faculty(x), x -> “research” -> r, x -> “name” -> n CREATE New(r) LINK New(r) -> “fieldName” -> r New(r) -> “researcher” -> n COLLECT ResearchPage(New(r)) ResearchPage r1 r2 “fieldName” “researcher” “fieldName” “researcher” “researcher” “databases” “Alon Levy” “info retrieval” “Oren Etzioni” “Alon Levy”
HTML Templates • HTML with extensions for formatting, conditionals, and enumeration • Associated with specific object or collection or is specified in object’s definition A template for the ‘r’ objects: <$fieldName>: <UL><SFOR n in researcher> <LI><SFMT @n EMBED> </SFOR></UL>
Data Integration • Strudel allows integration of information from heterogeneous sources using a virtual architecture and a global-as-view approach... • … or at least it will. • Currently relies on wrappers
Results and Future Work • Tested on AT&T and CNN sites • Stuff to worry about • Data integration • Ordering • Incremental updates • Conversion of existing sites