The Flightless Phoenix
Hill Times, February 21, 2018
Garbage fed into the creation of the Phoenix system means plenty of garbage to take out for those left to clean up the mess.
Out of the mists of the ancient Mediterranean comes the story of the birthing of a Phoenix from the ashes of a predecessor; a story that lives on to remind us of the unique myths that helped our ancestors cope with the fates and the gods. The ongoing saga of our modern Phoenix, a seemingly similar mythical bird, brings much angst and no delight to those who toil for the government of Canada.
The modern Phoenix, perhaps with a hubristic Icarus onboard, has two wings. One was to create a new centralize pay system out of a labour-intensive, successful decentralized service based on the knowledge of the experienced. The other wing was to carry a replacement for the 40-year old cheque issuance system used by 101 governmental departments and agencies.
This bird has yet to fly is as evident as that from the old pictures depicting the attempts to get a heavier than air machine of the ground. As with those early machines, their designers observed the birds. It was only when da Vinci-like minds examined how birds flew, the first successful heavier than air machines marched us successfully into the technological marvel of flight.
In the haste associated with our new budding and still building technological marvel of the computer, we often forget that the machine cannot and will not do the thinking for us. The thinking must come first and when that is ignored, the old adage associated with computers rears its ugly head. Even Charles Babbage, the generally accepted originator of the digital programmable computer, in the early 19th century was asked “if you put into the machine wrong figures, will the right answers come out?”
The answer of course is GIGO, or garbage in, garbage out. This is largely forgotten when dreams of large financial savings are combined with promised enormous new efficiencies and sold to uninformed and largely ignorant minds on matters technological.
The thinking for the old systems was done by some 2,000 experienced pay advisors across the 101 departments and agencies. Some 1200 of these pay advisors were eliminated by early 2016 just as the modern Phoenix was to fly and realize dreams of $70 million in annual savings. Today, no one mentions those savings, instead the emphasis is on the conclusion that the total cost of Phoenix, assuming those involved in its implementation persevere, will come close to a billion dollars
The experienced pay advisors were replaced by 460 newly hired pay advisors in a new Pay Centre in Miramichi, in northeastern New Brunswick. This cadre of pay advisors along with software adapted out of a commercial 1987-era product, PeopleSoft, owned by Oracle, had to deal with 80,000 pay rules flowing out of more than 105 collective, changeable agreements in addition to individual employment contracts. To make this work, Phoenix also included more than 200-custom built programs.
The Auditor General in a report to Parliament in November provided many of the above numbers but also noted that as of last June the number of outstanding pay change requests continued to grow to nearly half a million (494,500). Periodic updating of these numbers since gives no confidence that the modern Phoenix will outdistance these problems in any meaningful way and at some point, soar with the eagles.
Sadly, the Phoenix debacle is not unique. Since the introduction of the Internet some twenty-five years ago and development of a variety of associated speciality applications meant to speed and create efficiencies – sold largely on the basis of large financial savings they represented – governments everywhere, and the private sector as well, is littered with failures. Every area of human activity is included ranging from space to hospitals to passports to social services to financial management and even the management of weapon systems and the avoidance of war.
In foreign operations alone, COSICS (an early several hundreds of millions of dollars failure for secure communications at Foreign Affairs), passport issuance and immigration processing all had significant failures. Even a government-wide financial system (FINEX) ran into problem when it was introduced at Foreign Affairs and it was discovered that it did not have provision for dealing with currency exchange.
It does not have to be this way. The most fundamental of errors is the distance between those who have done the work and know its intricacies and those who seek to replicate it in a digital format. More often than not the latter dominate within this process and those who should provide an overview are blinded by possibilities that some magical system will be produced leading to great success.
Some twenty-five years ago I was directly involved in the development of a complex system designed to support the delivery of services to Canadians overseas. The first contract for the system was signed in February 1993 and within a few years the system – now called COSMOS – was implemented and deployed to over two hundred diplomatic offices overseas.
Today the COSMOS system continues in full operation and, giving value to its effectiveness, the technology has been purchased by seven other countries to assist in the delivery of their consular services. The cost of the system over the last quarter of a century is less than fifty million dollars. (Details on the system can be found at WorldReach Software, the Ottawa based developer.)
The debate still rages as to whether the government should walk away from Phoenix or persevere in the hope that the bird will eventually fly. There is no answer to the question except that if those involved does not include those with expertise of the work that has to be done by the system, then there is no hope for a fix. Garbage in will always result in garbage out in the world of computers.
Gar Pardy is retired from the Canadian foreign service and throughout his career was involved in the application of computers to support the work of the Department.
– Published with the permission of the author