Every IT person alive has fixed something in the "wrong" way, a way that wasn't scalable, secure, or otherwise proper, but a way that worked. It may have been intended to be temporary, but time passes and it becomes permanent. Over time, such fixes are often applied to problems caused by previous bad fixes. But they work, tenuously, for now.
This sort of activity takes place constantly, and indeed, it might not become a huge problem for quite some time. But there will be a point when there's no more room for kludges and poor fixes.
[ Also on InfoWorld.com: Learn how data deduplication can slow the explosive growth of data with Keith Schultz's Deep Dive Report. | Looking to revise your storage strategy? See InfoWorld's iGuide to the Enterprise Data Explosion. ]
Simply put, when you've attached enough Band-Aids to the corpus that it's more bandage than not, isn't it time to start over?
It's one thing to understand that such problems exist (and always will, to some degree) within a corporate IT infrastructure; it's quite another when the problem is extreme and affects products your company produces, sells, and supports.
I've seen some horrifying examples of software development malfeasance recently -- things like a thoroughly perplexing dependency on a specific network interface chipset for a simple client/server database application. I'm still puzzled how a standard Windows TCP/IP-based application developed and sold in 2010 could possibly be Ethernet-interface dependent. The only thing I can think of is that somewhere along the line, the developers "fixed" some problem with some terrifying shim that only works on Intel interfaces, and then wrote the requirement into the spec.
Another software-based example is an application that requires not only local administrator privileges to install, but also to run. In a domain environment, every user must be granted full local admin rights on their workstation to run this one application. It singlehandedly renders most corporate security policies moot. This application costs tens of thousands of dollars, and due to training and existing corporate mindshare, it's apparently indispensable. It's also an affront to all that is holy in IT.
And those are just two small examples. There are thousands more just like them.
One might think that the free market would take care of problems like this, because the company with the inferior product would necessarily be marginalized in favor of the company with the superior one. Unfortunately, that isn't always the case. In fact, I'd say it's becoming somewhat of a rarity, especially in markets with minimal competition.