Following is the first in a periodic series of opinion columns that IT veteran Rob Klopp is writing for Techwire. The views are his own.
For two years and some odd months, I was an appointee of the Obama administration, working as the CIO of the Social Security Administration. There I found a time-bomb of 40-year-old assembler-language and COBOL systems driving a system that is responsible for 5 percent of the GDP of the U.S. (60M lines of COBOL, 7M lines of Assembler). These systems are expensive to operate, difficult to impossible to extend, and maintained by baby boomer staff that has retired or is retiring. Sound familiar?
While there I published a series of popular blogs on IT modernization on the Federal CIO site, cio.gov, and this article is a continuation of that thread.
Since leaving federal service, I have been working to help to modernize IT systems in California state government. Many of the issues that troubled the SSA and other federal departments are issues in the state. Throughout this series, I'll try to identify common government and state-specific issues and suggest fixes.
In this first post, I will introduce the several objectives, issues and ideas that will be elucidated in the posts to follow.
Let me start with the most critical objective: We need to invest in modernizing the skills and experience of state IT employees. IT modernization cannot be accomplished by staff with dated skills and experience. Government IT modernization cannot be accomplished by relying on commercial contractors. When we engage with commercial technology vendors, they need skilled state staff to support them as partners.
We need to build that experience by funding and taking on real modernization programs. We do not need policies that prod staff to use modern technology like the cloud. We need product development programs to use the cloud. We do not need policies describing software engineering best practices. We need product development programs that use best practices.
Only by developing modern infrastructure and using modern software engineering techniques will we accomplish the staff upgrade objective. Note that this objective suits the state as well as the contractors who support the state. Contractors desperately want to exercise these modern capabilities.
Technology is changing at a very rapid pace. We need to recognize that in today's environment, we need to build or buy systems capable of change. Managing technological change is the primary requirement for modern systems. We need to build or buy systems that are optimized to quickly utilize the technologies that have just become or are now becoming mainstream: machine learning, wearable technology, the Internet of Things (IoT) and autonomous things, virtual reality interfaces, ever more mobile computing. Silicon Valley has developed new architectures and approaches that are optimized for rapid change and are no longer optimized around interconnected and tightly integrated enterprise systems. We need to utilize these new architectures.
Not least, we need to engage our government program owners as product owners of these new systems. The days when IT built requirements or user stories on behalf of the business have passed. IT and the programs must partner together and join with vendor partners to develop these new products and to grow the products as technology advances. We cannot overlook this significant organizational change.
Finally, we need to tolerate some risk, knowing that we might fail here and there. There are specific techniques available to allow product programs to fail fast, in two-week increments or, worst case, in two-month increments. We need the courage both to start risky programs and to stop programs early when they get off track.
I know that all of this sounds like a Pollyanna story, but in the posts that follow I will provide real-life examples of how this has been done. I will provide stories that relate the modern way to do things with the ongoing issues we have when we use the legacy technology, architecture, and approaches that brought us to this point. I will suggest how the state can make real progress in a two-year time frame. I may even suggest some quick wins; but they will be in the context of a broad modernization program, not as showy features.
I think that you will be surprised to see how easy it is to get started and how obvious it is that progress can be had. I think that you will appreciate the importance of rebuilding the skills of the state IT staff and providing a path for very smart state workers to make a career in IT.
In the next post, I’ll tell a story that helps you to understand better where we stand with regard to experience and to set some objectives to help us know when we are back on track.