Software lemmings head for the platform cliff--a provocative post from Greg Gianforte
How would you feel if your mechanic handed you a 125-piece wrench set rather than actually fixing your car? What if another mechanic then walked up to you with his tools and started arguing with the first guy about whose tools were better? You sure wouldn't feel like either of them was going to help you with your problem, would you?
Yet that's exactly what software vendors are doing today as they engage in platform wars – to the detriment of their customers and the industry.
-Oracle announces its Fusion platform, revealing a grand architectural chart that shows how they will exercise total control over their customers' computing environments.
-SalesForce.com unveils a platform dubbed "AppExchange" that enables customers to engage with unstable, third-rate software vendors more rapidly and conveniently than ever.
-SAP touts NetWeaver, which grafts a proprietary platform onto a collection of otherwise practical standards to render them impractical and non-standard.
-And, of course, Microsoft seeks to achieve global hegemony with a platform of its own: .NET.
Corporate customers, however, are not out shopping for new computing platforms or a better IT stack. They need business solutions that actually help them compete and succeed in the real world.
They also want something that no platform-obsessed vendor seems able to provide: a technology partner who can actually be held accountable for promised business results – such as increased revenue, reduced expenses, faster time-to-market, or improved customer satisfaction.
Yet, for some reason, software vendors continue with their platform fetish – despite the fact that platforms don't actually provide any tangible value.
AppExchange is a perfect example. It's a great way for a team of clever, untrained coders to package their software for mass consumption, despite the fact that they may lack a professional services organization, financial stability, or a credible long-term technology roadmap. But does any CIO really want to make a strategic purchase from a vendor like that?
Platforms usually aren't platforms anyway. They're "marketectures" that exist purely to rationalize bad acquisitions.
Is "Fusion" really the right name for the duct tape and rubber bands that will supposedly hold together Peoplesoft, JD Edwards, Siebel, Retek and Oracle applications? "Con-Fusion" is probably more apt.
One reason that vendors do the platform dance is that it generates hype. It also creates the fear, uncertainty and doubt that are always useful when you lack a solid value proposition.
Plus, when you architect a platform, you make the third-party developers and integrators responsible for actually doing something with that platform. So you can collect a lot of cash and still have absolutely zero accountability for what happens to your customers afterwards.
So what's the alternative to platform madness?
The answer to that question has two parts. First is the on demand delivery model. On demand eliminates the need to create a proprietary technology platform as a competitive differentiator. The IT stack only exists to support applications.
By taking the entire IT stack (including operating system, database, app server, web server) off the customer's hands, the on demand model renders identity of the stack's individual components meaningless.
The applications themselves must be capable of being integrated, customized and scaled as required – but the underlying "platform" should not be the customer's headache.
The second part of the answer is open source. Open source commoditizes the stack. MySQL replaces Oracle. Linux replaces Windows. TomCat and JBOSS replace Websphere and NetWeaver.
Vendors that are still trying to differentiate themselves in these commodity businesses are clearly headed in the wrong direction. Yet that is exactly what platform vendors continue to do.
So next time a software vendor makes an overblown platform announcement, someone should ask them some hard questions:
Why are you trying to co-opt open source instead of using it to deliver real value?
Have you forgotten how to create applications that actually solve problems, or is that just too much work?
How much will this platform ultimately cost me to adopt, and what will my quantifiable gains be once I've done so?
Software vendors should listen to their customers, deliver results, and charge a fair price for their products. Anything else is just a smokescreen for failure.