Mashups, in my opinion can be considered a synthesis of the existing data on the web. Limitless applications can be thought of when dealing with Mashups. From artistic applications focusing on a concept and using bits of data, displacing and combining them to make a statement, to practical service-oriented mashups such as data layers on maps in an effort to facilitate navigation, mashups introduce a whole new range of possibilities.
Much effort is being put towards creating web based services unique to their own field, and from time to time “inventing the wheel” again; while if we look at the current world wide web as our toolbox and building blocks, one can create a synthetic service taking advantage of semantic data.
In case of AJAX, as an increasingly popular tool not only in mashups but in the creation of web in general, I believe the tendency towards creating desktop-application-looking websites introduces a trend, and a vision. The text point out that the AJAX experience can be inconvenient to navigate back and forth since the bit of information loaded in different parts of the page do not result in loading of a different file form the server. I believe the point is valid. But the back and forth navigation are constrained dictated to us as users and designers by web browsers. Browsers that evolved from their early 90’s versions serving HTML4 pages to now asynchronous data loading through XMLHttpRequest.
The vision, and the trend proposed by technologies such as AJAX, might be a calling towards revolutionizing the way we browse the web. If navigation becomes and issue in AJAX-loaded pages in our browsers, I believe rethinking should be done in the design of browsers. In my opinion, the time when local computers will serve as mere operating-system-as-browsers is not that far ahead, taking off load from local computing and putting it all on servers, with our operating systems acting just as browsers and nothing more. In this rather futuristic context, navigating back and forth triggered by loading the complete page DOM will not be efficient anymore.
If we think of API’s as interfaces between web applications and services, semantic web can be considered the spoken language. The grammar in which countless parts and nodes of this increasingly growing web fall into their own place, making it possible for us to navigate through and benefit from data. This is where closed architectures become unavailable to other services, making it difficult to interface with them. Datascraping in my opinion is the fruit of non-semantic closed architectures. Personally I do not approve of the closed mentality of systems as such.
But this openness does not come at no cost. Obviously issues of intellectual property and privacy introduce challenges here. A coherent policy might be a workaround to the intellectual property challenges.
To sum up, I believe an effort towards making the web a semantic structure, offering easy-to-use API’s accompanied by appropriate licensing to protect intellectual rights can pave the path for a robust web. A web in which applications will act as building blocks of a well structured accessible structure that can be reshaped, re-purposed, re-used and improved.