Birds of Prey
Informatics planning limits deadly crashes
Preparation for informatics projects is at least half the battle, if not 80 percent of it. The landscape is littered with the remains of failed projects that could have been saved if just a few key elements were addressed before fingers hit the keyboards. To establish a frame of reference, I take you back to last winter at Beaver Creek, CO.
For many years, this posh little community was known for two things:
a) It has been the residence of the late President Gerald Ford who, in his younger years, put many miles of Beaver Creek snow under his skis, and
b) The Birds of Prey downhill ski course.
The only stop on the Men’s World Cup Ski Tour in the United States is at Beaver Creek, and Colorado papers report on the races and training runs daily.
Now, I have always thought that one of the two craziest sports in the world is free soloing, a form of rock climbing whereby insane climbers scale astonishing mountains
and rock walls without any ropes or safety equipment. There is a one-fall-per-customer policy here, so all the successful climbs in a climber’s history are permanently negated by just one mistake. The second craziest sport in the world, in my book, is the World Cup downhill ski race.
Most of us only see World Cup skiers during the Winter Olympics. However, if you search deep within the winter cable or satellite listings, you can find several broadcasts of the World Cup downhill circuit, most of which takes place in Europe. The problem with watching World Cup skiing on television is that the speed of the racers, the steepness of the pitch, and boundary between control and high-velocity crashes cannot be reasonably assessed.
I have always wanted to get a first-hand look at a World Cup downhill course, and Beaver Creek offered me a chance. Atop Beaver Creek, a few signs point to the Birds of Prey course, but access to the actual start house, and indeed the first hundred or so meters of the course are off-limits. Mortals can access the course further down the run, and it is all black diamond or double-black diamond. Now, ski resorts use a few different criteria to define single- and double-diamond runs.
• It could be the run is an intense mogul field closely bordered by trees or rocks, or both. Thus, skiers who start down these runs and figure out they can’t gently traverse across the moguls for fear of smacking into wood or granite are in for a very dangerous and miserable experience.
• Second, single- or double-diamond runs might be couloirs; these are usually 45-degree or greater notches between rocks that give a skier almost no room to turn. For these runs, falling is not an appealing option, as many have rock jumps awaiting the tumbling, screaming victims as they pick up steam.
• Double diamonds also may be found in the back bowls of many resorts, and typically involve insane drops between many unforgiving trees.
• Last, there are steep, icy runs with turns requiring the shoulders to sometimes come within inches of the run in order to navigate the hills. World Cup downhill courses belong to this category.
Somewhat innocently, I skied through a wooded area obviously traveled by others, only to find myself just down the run from the start house. As soon as I cleared the trees, I saw the pitch of the run, and noted that the surface was very hard snow packed down to a glistening icy sheen. Absolutely no one was on the course, and it became obvious as to why this was the case. As I dug my edges into the side of the hill, I snapped a few digital photos, but realized they wouldn’t capture what I was seeing: a smooth, mind-boggling plunge down a sheet of ice whereby just pointing your skis straight downhill guaranteed 60 miles per hour in moments, on the way to nearly 80 mph if you were a World Cup racer bombing down this thing in competition. There’s no way to describe the feeling of struggling just to stay upright on the hill, and imagining someone actually screaming down this run in a full aerodynamic tuck.
According to Alison Berkley, writing for the Denver Post, the course is 1.7 miles long with a vertical drop of 2,483 feet. Racers start at an elevation of 11,427 feet, and sail down to 8,943 feet in one minute, 39 seconds (course record by Daron Ralves). The average speed is a blood-draining 60.98 mph, and skiers hit 77.13 mph on some sections.
I don’t know about you, but 40 mph on skis is cooking, nearly twice that would be beyond all reasonable comprehension … and clipping along at 77 mph on ice just causes you to shake your head in utter astonishment. There are deaths and serious injury in this sport, and these skiers are the best in the world.
The point is that the Birds of Prey course doesn’t prepare itself for World Cup competition. Ridges or ruts in the snow, the tiniest pebbles, or any irregularity can send a competitor on the ballistic trajectory of a cruise missile toward the nets (if they’re lucky), or the tall pines (if they’re not). Even not hitting anything at all can mean snapped ligaments, broken bones and racing suits melted by the friction. Thus, planning is everything. Twenty-five thousand person-hours are required to ready the slopes for that 1:39 plunge, and those hours are expended over three month’s time. It is a major project management effort.
When it comes to informatics, a multi-site project with 100 or more concurrent users could consume 6,000 vendor hours, and many unmetered hours of internal resources. Projects spanning multiple sites might cost several million dollars in time, software, hardware and associated annuity support costs. Successful projects require up-front planning prior to the rubber meeting the road.
It is often surprising to learn that such planning has not been done, other than management memos that have been circulated promising that a system will be in by a certain date. These projects are often considered to be straight contract labor efforts with a fixed deliverable, and that is a devastating misconception. Major projects are customer-vendor partnerships that will have a lot of peaks and valleys, and a misunderstanding that the customer is at least half of that partnership is an all-too-common mistake.
Yet it happens. Projects begin with no documented user requirements, at least not those that are of material value to the developers. It’s debatable whether a functional specification needs to be completed up front; I believe it should be, but at the very least, a plan for documenting the functional and design specifications during the development is critical. In FDA-regulated accounts, such documents comprise part of the total validation picture. In non-regulated sites, they just make good common sense. The successful vendors are militant about laying out the roles and responsibilities of all parties right up front. ‘We do this, you do that, and we will need short turn-around of decisions at these stages’, and so on.
Ah, let’s take a look at my favorite scenario, the hard-core purchasing agent who starts every negotiation with a demand for a fixed-price contract. Let me tell you why these are a bad deal for customers and vendors alike. Purchasing agents love fixed-price contracts because the perception is that they limit the risk of the customer, and hence put all of the risk on the backs of the vendors. Scott Adams and Dilbert have gone a long way toward vilifying consultants (I guess that’s actually Dogbert), and place blame for cost overruns squarely on their backs. However, it’s not usually that cut and dried. So, the fixed-price contract says that X functionality will be delivered for Y dollars, and usually by Z date.
I have contractors working on my house all the time on fixed-price contracts. Painters, for example, quote a price for the job, and you know the job is done when the paint is on the house. This is often the mentality of purchasing agents in many companies because they do negotiate cut-and-dried deliverables such as building additions and paint jobs. But complex software projects are not in the same universe, and savvy vendors know that fixed-price contracts are a huge risk. Even if every minute detail is spelled out, and contingencies are built in for change requests, nobody is happy when a major change order comes through for $80,000 because there was a misunderstanding on a requirement.
Thus, vendors have to be almost litigious in their delineation of items to be delivered in a fixed-price contract, and they are working in an environment whereby they don’t control customer resource allocation, hardware availability or many other factors. To quote a fixed-price contract means they must invest a lot of time detailing the deliverables, and the cost for such planning is often passed on to the customer. There is also the matter of pricing tasks, and vendors must build risk into that pricing. Like the gas option on rental cars; you pay for the gas whether you use it or not, and most people don’t use all the gas.
This is bad for the customer because, if the requirements were not exhaustively spelled out and, thus, quoted by the vendor, if a new “gotcha” requirement turns up, there’s time to scope it, price it and deliver it, and the customer pays for that, as well as any gutting out of the existing system if it is affected. So, meticulous planning is the rule even though most vendors will push back mightily on fixed-price projects.
Even in time and materials projects, planning is required to prevent the Dogbert stereotype from running the meter up as changes in course and scope pop up. If the user requirements and, to the extent possible, functional requirements are detailed up front, then the number of unpleasant surprises is limited.
For the record, the user requirement that a World Cup downhill course must be a sheet of ice didn’t work to my favor. I can promise you I didn’t point my All Mountain skis down the course and snap into a tuck. If I had, they wouldn’t have found my body until spring, high up in an Austrian fir. I hit more mph than I wanted to, and my skis chattered on the ice like a pack of baboons startled by a lion. I may have eaten more shaved ice than a hungry kid at the state fair, but I made it down with my limbs and Daron Ralves’ course record intact, though I doubt Daron needed two fingers of Laphroaig afterwards.
I was lucky, but you can make your own luck with extensive planning, and you can get down the mountain in style.
Randy Hice is president of the Laboratory Expertise Center. He may be contacted at editor@ScientificComputing.com.