Scheduling is an art as much as a science
Scheduling is an art as much as a science
Our Scheduling Background
Bill Pepoon, Managing Partner at Construction Science, prepared his first CPM schedule at Kansas State University in 1980. The entire logic diagram was sketched by Bill and two classmates on a continuous roll of plotter paper approximately 30 feet long after many discussions on logic and review of the construction documents. The project was a small ambulance facility in Manhattan, Kansas. As part of the final exam, Bill’s instructor inputted the logic into a computer program running on a mainframe computer to confirm the project duration and critical path manually calculated by Bill’s team. He still has the original mainframe printout in his office. By the way, he got an A on his schedule.
In 1983 Bill responded to an ad for a “CPM Scheduler”, which in those days was like seeing an ad for lifeguards on Mars. There are actually people who do this for a living? Keep in mind, there were no commercially available CPM software programs until Primavera Systems opened its doors for business – also in 1983. The consulting firm that hired Bill had been running their own proprietary software on a mainframe computer since the early 1960s. Bill stumbled into an opportunity to learn scheduling from people who had already been scheduling for two decades.
Bill’s first employer also believed strongly in first drafting each schedule on paper. The thought process becomes more deliberate when it is harder to make changes. Even today, some therapists helping writer’s overcome writer’s block recommend going back to pen and paper, or even a typewriter. Another reason for this approach in scheduling is that “computer time” really meant something back in 1983. The cost of running mainframe computers was substantial. Mistakes on paper resulted in additional runs on the computer – and a stern lecture from Helen, the woman in charge of the computers. You did not want Helen calling you for any reason!
Up until the mid-1990s the dominant way of scheduling involved the Activity on Arrow (AOA) Diagram Method. What this means is that each activity has two “nodes” – the “I Node” and the “J Node”. So you would refer to an activity by two numbers. This was also commonly referred to as the “I-J” method. A simple example of an Activity on Arrow diagram appears below:
Notice that durations appear on the arrows between the nodes. The nodes themselves do not carry the durations. To find the longest path you simply follow each path according to the arrows, adding the durations along the way. (Hint: the longest path in this diagram is 16 days – can you find it?) The variance between the longest path and a shorter path is float. The concept of float makes a lot more sense when you are adding the durations and comparing the length of each possible path through the network diagram.
Primavera used the AOA Diagram Method exclusively in its first scheduling program for personal computers, Primavera Project Planner (known today as “Primavera P3”). This was a DOS program running on computers that today would be considered comical. Primavera P3 worked perfectly fine on a machine with a measly 12 Megahertz processor and a 40 MB hard drive. Today’s dishwashers have more powerful processors.
Nevertheless, understanding logic in the AOA Diagram Method is pretty easy. It looks a lot like a road map where the distances are shown between towns. After all, you would not show “Sacramento” on a map with a value of 30 miles. What the heck would that mean? In today’s terminology we would refer to all of the activities in the network above as being Finish to Start; the preceding task must finish before its successor(s) can start.
Primavera dropped the AOA Diagram Method in favor of the Precedence Diagram Method when it released the first Microsoft Windows version of Primavera P3 in the mid-1990s. Also called Activity on Node, PDM introduced other options for describing relationships between activities:
- Start to Start (SS)
- Finish to Finish (FF)
- Start to Finish (SF)
While these relationships make schedules more flexible, it has nearly destroyed the ability of most young schedulers to appreciate the meaning of float. When activities overlap, manually figuring out the longest path is very difficult because you do not simply add all of the durations shown along each path. This can be seen below:
Activity E has a Start to Start relationship with Activity I, with 10 days of “lag”. When calculating the longest path, the duration of Activity E (18 days) is ignored because Activity I starts 10 days after Activity E starts (SS 10). It does not wait for Activity E to finish. The path through Activities E and I would be 8 days longer if this relationship was changed to Finish to Start. These other types of relationships have simply introduced a lot of complexity to schedules.
Programs like Microsoft Project also use PDM. The reality is that Primavera’s switch to PDM pretty much standardized PDM in most industries. Activity on Arrow software is very hard to find these days. Yet Primavera’s most recent programs like Primavera P6 still do a poor job of expressing PDM logic. The above diagram was actually created in P3, which Oracle (which now owns Primavera) discontinued in 2010.
Technology has improved construction scheduling, but it has introduced quite a few bad habits. Most young schedulers build the schedule inside programs like P6 and then struggle to understand why the project end date is “off”. The “old school” method of producing a logic diagram first and then inputting the information into the computer has been skipped by an entire generation of schedulers. An experienced scheduler can visualize the logic in his head even while creating a schedule entirely on the computer. For anyone else this is difficult at best.