In his recent book, "Mashup Patterns: Designs and Examples for the Modern Enterprise," (Addison-Wesley, 2009) author Michael Ogrinz indicates that 'screen scraping' in the PC terminal era gave a bad name to any method to extract raw data from a user interface. This is just one hurdle that champions of mashups may face in the enterprise.
But the mashup sponsor should remember, suggests Ogrinz, that the underlying Document Object Model (DOM) can enforce data structures, no matter how scattered the individual HTML elements may be. 'Harvesting' is the operative term these days – not 'screen scraping.'
In his book, Orginz discusses Alerter, Leading Indicator, API Enabler, Infinite Monkeys and Time Series patterns, among others. Together these can be viewed as a set of Harvesting patterns. Looking at mashup projects with such patterns in mind can cut mashup development time on both the front end land back end. There is a heavy helping of so-called REST design patterns throughout.
Much mashup design requires a 'chaining' of APIs, notes Orginz. He shows how a chain of variables may be involved in, for example, establishing the current conditions in exotic and remote travel packages using Web interfaces.
You know you have an Alerter pattern, for example, when you are tracking only two data points – points that may compare a current data reading, while observing for deviation. Originz notes that care must be taken with Alerter implementations to ensure that they do not affect the stability of systems being monitored.
Ogrinz's book offers other various design patterns for enterprise mashup development. Moreover, SearchSOA.com has just published an excerpt from "Mashup Patterns: Designs and Examples for the Modern Enterprise." The excerpt covers issues that set the stage for enterprise mashup planning efforts.