Sometimes I think the community of Javascript libraries is like a high school popularity contest, with crowds swarming one cool thing before dropping it and moving on to the next. In a project last year, we started with a little hand-coded JS to spice up the site a little bit, then started dipping into the bigger UI libraries of Dojo, script.aculo.us, Prototype and more as the clients expectations went through the roof. I still have deep misgivings on making a web site work like a rich client application using HTML, CSS Javascript and/or AJAX. It’s still a web page, not an RIA, and concerns over web-scale scalability, responsiveness, variation in the client machine (six different browsers, JS on or off, Flash on/off, various readers for accessibility to the visually impaired, rendering on small devices, etc.) make rolling-your-own a fool’s mission.
We’d stabilized on a set of tools long enough for me to start to dig deep into its capabilities, when along comes a suggestion we look at Yet Another JS library, Ext. It is impressive at first glance, no doubt, and the demos have the requisite whiz, bang, oooh and aaah. But concerns over maturity, licensing, suitability to task, cost of retooling existing pages make it a questionable switch. At some time you have to commit, people. Live with what you’ve got or face the costs of serious rewrite. The tenets of RAD, XP and “Agile” as promoted by Scott Adams in DIlbert, not true practitioners, have given folks in charge the impression that they can just chase after the next shiny thing and follow the fashions without concern to the engineering implications of launching a world-class web site. Ah, well, my consultant friends tell me. This is why we get the big bucks.
Now I hear tell jQuery is where its at…