Whither XML ca' 2005

1
Whither XML ca. 2005?

• DBPL/XSym Panel
• 29 August 2005
• Jayavel Shanmugasundaram
• Cornell University

2
A Not-So-Uncommon Dialog
People Non-XML Database Researcher (NXDBR) and me
Location SIGMOD/VLDB/ICDE Conference
NXDBR What are you working on these days?
Jai XML data management, among other
things
NXDBR Really? You are pretty young, but there
used to be something called the
hierarchical data model
Jai Well, XML does have a declarative
query language
NXDBR You are pretty young, but there used to
be something called the
relational data model
Jai Well, XML has other advantages
such as
NXDBR But dont you also do joins in XML?
Besides I dont like XQuery it
is too complex.
3
Are there any fundamentally new data management
challenges enabled by XML?
Yes!
4
XML Features/Challenges

• Standard format for data exchange
• Data publishing
• Filtering/stream processing
• Large-scale data integration
• Optional/Partial schema (e.g., catalogs, schema
  chaos)
• Managing data with partial or no schema
• Text as a first-class citizen (e.g., IEEE papers,
  Shakespeares plays)
• Unify management of structured data and text
• Common ground with the IR community

5
Where are we today?
Polishing a round ball
Barely scratched surface
Missed the boat
Progress but work left
Will we ever scratch the surface?

• Standard format for data exchange
• Data publishing
• Filtering/stream processing
• Large-scale data integration
• Optional/Partial schema (e.g., catalogs, schema
  chaos)
• Managing data with partial/no schema
• Text as a first-class citizen (e.g., IEEE papers,
  Shakespeares plays)
• Unify management of structured data and text
• Common ground with the IR community

6
Towards Unifying DB IR

• IR Community
• INEX initiative
• Keyword Search Structure
• DB Community
• Keyword search over databases
• XQuery Full-Text

7
Where are we today?
Polishing a round ball
Barely scratched surface
Missed the boat
Progress but work left
Will we ever scratch the surface?

• Standard format for data exchange
• Data publishing
• Filtering/stream processing
• Large-scale data integration
• Optional/Partial schema (e.g., catalogs, schema
  chaos)
• Managing data with partial/no schema
• Text as a first-class citizen (e.g., IEEE papers,
  Shakespeares plays)
• Unify management of structured data and text
• Common ground with the IR community