Recognizing Dead Ends By Their Neighborhoods
The WWW is still a Wild West show, with an ever increasing number of moving parts. With virtually non existent barriers to entry and proliferating architectures, languages and tools, the field is littered with surprises for the unwary and deadly embraces for even experienced practitioners.
Writing software is much like writing a book, in that there is no one right way, but there are certainly a lot of wrong ones. Why this is so is the subject of a debate that has been going on for some time now. In both cases, one contributor to this is the amazing flexibility of programming, at least the mundane portion of it involved with cutting code, all too often before any integrating structure is in place.
More particularly, there is more than one way to express something, regardless of any language, human or artificial. And that multiplicity leads to confusion. With multicore processors, ever cheaper memory and smarter compilers, most of this could be hidden away, with the bowels of the machines themselves and out of sight of those who sign the checks that keep it all running.
The WWW is more sensitive to these issues, particularly where style imagined for customer facing appearances trump the boring business process implementations buried in the plumbing that no one can really see. The biggest indicator of this is answered by a simple question: What is the web site trying to sell?
More than regular boring business implementations, the customer facing skins need continual maintenance. The bright lights that built them originally, or made the most recent changes, almost always move on to bigger and better things.
So how to manage this inevitable turmoil, pay the rent and meet payroll at the same time? By simple tools and metrics that can be run against any web site to identify issues that inform the next changes and choices. Simple things like checking tags, dangling structures, broken/missing links, folder structures, javascript and css file depths, to name a few. Like lighthouses on rocky coasts, using these tools on a consistent basis, before during and especially after changes are applied, will, not may, highlight potential problems before they become actual ones.
© Copyright 2009 Chuck Brooks for FutureWare SCG
A Word From Our Sponser
Stop Toxic Email From The Bad Guys Before It Can Get Close To Your PC! Our 
ePAL for email security is a barrier between the outside world and your regular email program