Ajax and the Impact on Caché and Similar Technologies

In an incredibly short period of time, Ajax has undisputably become the dominant method of crafting and designing web applications.

An inevitable consequence of Ajax-based web applications is the single-page web application, where what's happening during the user's session is the swapping out of parts of an otherwise persistent "container" page.

The single-page web application is a highly significant tipping point, heralding a profound change in the design and architecture of web applications. The reason is that Javascript, once regarded as an interesting but limited and flawed language, becomes all important, not only for application logic but also for many aspects of session persistence. Javascript's lean-and-mean objects suddenly become highly relevant and come to the fore. The result is a shift in the balance of power from the back-end to the browser.

This shift leads to a difficult question: what does the back end have to do in such an environment apart from simply save data and generate data-driven page fragments?

Even though Javascript isn't the most performant of languages, the average user's PC is more than capable of handling the persistence and application logic of that individual user. The massive processing power at the back end that was needed to handle potentially thousands of concurrent users in the "classic" web model is radically reduced in the Ajax model.

And so the inevitable question: what unique benefits, if any, does Caché or similar technologies bring to the Ajax-oriented world? Caché ObjectScript, Caché Objects and the MUMPS language have a greatly reduced role, substituted by the more than capable alternative of Javascript. Is what's left - the performance and scalability for which Caché and similar technologies are renowned - enough of a persuasive benefit for these technologies to survive? Will MySQL prevail as "good enough", particularly given that most Internet servers come with MySQL ready for use and without any fuss or further cost. Does it mean that Caché and similar technologies will be relegated to the high-end in the Ajax world? And if so, how can that transition be facilitated?

Difficult questions and issues that need to be confronted and faced by anyone in the Caché, MUMPS and related communities.