One of my co-workers sent me this article from the folks over at 37signals (makers of Ruby on Rails) on why enterprise software sucks. The hypothesis is that it sucks because the decision makers aren’t the ones who end up being the end users. While that may be true, I think this is an oversimplification.
In a large company, there is bound to be a myriad of reasons why certain software is chosen; some of them political, some of them financial, and yes, sadly placed last, some of them technical. The thesis falls down if you consider products like Novell Novell Groupwise. It’s a terrible product, but we keep using it, and that includes the executives and other decision makers that decided to buy it in the first place. Sometimes in the enterprise space, there’s just nothing better that you can purchase off-the-shelf; at least we’re not stuck with Lotus Notes.
Of course, some of you may comment that non-off-the-shelf-solutions are frequently far better, and I would tend to agree. But implementing such solutions requires a high level of technical skill, and believe it or not, the concept of an enterprise hiring and retaining talented, responsible, and highly intelligent system administrators to implement and maintain these solutions for the next ten years is a fairly new concept. In a way, this is how open source is going to turn management methodologies on their heads; all these cheap tools with great feature sets that rival commercial software are just within the grasp of IT executives, if only they could get the talent to implement it. And is that a risk they’re willing to take yet? I guess it depends on the organization.
On an unrelated note, I’m pleased to reveal that today we finally powered off our old EMC IP4700 storage device, the same one that crashed last November and brought CBC.ca down for several days. While writing this article I had to Google for “EMC IP4700” and remind myself that back in 2001, this thing was considered a top-of-the-line NAS device, and in many ways, it was. Storage technologies evolve so quickly, though, that the device was obsolete in 2004, and the bottom line is, we pushed that obsolescence envelope too far. Why that happened is a long story, but I’m glad that we’ve closed yet another chapter in the history of CBC.ca infrastructure.