Firegoose for data providers
Why make data available
Making structured data available on the web allows it to be recombined and visualized in unexpected and valuable ways. Allowing data to be mashed-up with that of other providers multiplies its value.
Formats / Protocols
Microformats are an easy way for existing web sites to serve structured data. The structure comes from CSS tags used as markup to label the fields of your data. The Firegoose, as well as many other tools, can read microformats. The advantage is that no extra software infrastructure is needed beyond that which is already in place to serve web pages.
A website that serves data in the Gaggle Microformat needs no further effort to integrate with the Gaggle. This level of integration can be acheived in a matter of hours and can provide real benefit to users.
Embedded XML provides similar functionality along with the advantage of a well defined syntax.
Linkage between web sites and web services
It is increasingly common for a data provider to maintain two external interfaces, a web site serving HTML to web browsers and a web service serving XML (or an alternative like JSON). It is less common for any linkage to exist between the two. A web service by itself allows scripts and workflow tools access to the data. But these scripts and workflow tools sacrifice the simplicitly and interactivity of the browser and can be difficult for non-technical end-users.
Providing a machine-readable link from the HTML to the web service allows the web site to serve as a friendly user interface which, when augmented with tools like Firegoose, can also provide programmatic access to the underlying data. Users can then browse the database using the familiar point-and-click web interface. When they find relevant data, they can, using the Firegoose or other similar tools, acquire the data of interest in structured form. The data can then be brought into powerful visualization and analysis tools (like Cytoscape or R).
Data integration can take place in the browser and process only the records of immediate interest on-demand. This process can replace other more cumbersome forms of data integration such as maintaining local copies coerced to a common schema.
Custom handler scripts
Broadcasting data from Firegoose to a website requires that a custom handler be written. A custom handler is also necessary to parse nonstandard formats into Gaggle data types. Examples of these handlers can be seen in the Firegoose source in the folder /firefox/chrome/content/handlers/.
Getting HelpData providers interested in enabling data exchange through the Firegoose are encouraged to contact us for help through the Gaggle discussion groups or by emailing the authors of the Firegoose paper.
|© 2006, Institute for Systems Biology, All Rights Reserved|