Ayesha Malik, Object Machines
Date: Thursday, February 06
Time: 10:45am - 11:30am
Location: California Ballroom A & B
XML is the most flexible way of transporting bioinformatics data. XML schemas are the objects necessary to design XML structure and to ensure that it does not violate the essential rules of the domain. A badly-structured XML document can usually be traced to a poorly-designed XML schema, which then affects the efficiency and extensibility of the system.
There is currently no one industry standard in bioinformatics for designing these schemas; some XML protocols, such as BSML (Bioinformatic Sequence Markup Language), are still available in DTD format, which is now held to be not as flexible as XML schemas.
In the absence of industry standards and guidelines, managers and programmers find themselves constructing XML schemas that do not allow their applications to be flexible and cannot integrate with other systems. Malik discusses how to design extensible and object-oriented schemas that can be used internally, or to form a consortium to make an industry standard. Learn how to convert or transform your XML so that it can easily be plugged into the protocol any other institution that your system interacts with is using.
The presentation outlines best practices for designing XML schemas for bioinformatics. Specific topics include:
Following Object Oriented Principles:
1. How to minimize dependencies between programs (i.e., minimize "coupling" between programs)
2. How to group together related pieces of information (i.e., maximize "cohesiveness" of programs)
Choosing between Composition and Inheritance
Achieving modularity by building component libraries with domain-specific data types
Choosing Appropriate Design Patterns:
When to use Russian Doll, Salami Slice, and Venetian Blinds Design Patterns
Integrating With Other Systems:
The use of XSLT stylesheets
Download presentation file