“How Washington State handled a flood of applications to be its ‘pot czar’” is the most important information technology (IT) story of the year for me. It’s not the titillation of marijuana decrimininalization that deserves our attention: it’s the bottom-line story of delivery of results in a tenth the time, at less than a tenth the cost, that demands deeper understanding.
Shadrach White of cloudPWR is the overt hero of this story. He delivered an application in under three weeks for $7800 that he estimated “would have cost $150,000 and taken months to implement” if he’d done it more traditionally. His approach also made mobile use possible, which multiplied the practical usability of the RFP application.
We in IT often come across extravagant claims like this in white papers, and have learned to be skeptical. Happy customer stories nearly always leave out weeks of training, or proprietary hardware, or one of the other compensating inputs required to complete a project.
As little programming as possible
This Request-For-Proposal (RFP) review system in Washington is different not because it appears as a news story, but because, to the extent I’ve been able to confirm on my own, White truly did beat “traditional enterprise tools” by the ‘game-changing’ factors mentioned above. How did he do it? With as little programming as possible: in his words, “I took three tools and put them together …” Crucial in this is not just that Box, Google Docs, and DocuSign have liberal and low-cost licensing, or that PDF is adequately standardized, but that the tools all have useful, adequately simple application programming interfaces (APIs). Those APIs make construction of “mash-up” workflows practical.
There are limits to reworking consumer-grade tools in this way. I have delivered review systems for many years, and recognize that we still struggle to provide functionality that was available in specialized products from the late ’80s and early ’90s. To a large extent, though, it doesn’t matter: as soon as mash-up approaches give results three or ten times as fast as more traditional ones, as they are doing now, mash-ups will win, and security-consistency-performance loose ends will be tied up after the fact. When Forrester vice president Jeffrey Hammond writes about continuous delivery, modularity, and agility, or Justin James describes “Web services” as a top-priority development trend, the Washington-state RFP review application is the sort of success that makes these abstractions real. It’s not just that “… the risk aversion of I&O [infrastructure and operations] has become problematic“, as Gartner analyst George Spafford puts it; slow IT departments will simply be left behind to clean up the holes in mashed-up solutions.
We might be “… on the Verge of a New IT Landscape … too early to make any solid predictions …” at a business level, as Arthur Cole has it; from everything I know, though, the premium at the implementation level will be to do whatever it takes to escape from old silos, and ensure that service-oriented architectures (SOA) help data reach the places where they can be useful. Mash-ups like White’s are an excellent model to follow.