They must not produce different results.
Is like in the mathematics, when Youhave different method to solve a particular problem. You follow different algorithms, formulas, what else You want but the results must be same.
Here we are in the same situation.
For an end user point of view all this is hiden in the software algorithm, so we saw just the calendar dates that are maching between the two.

May be if You want to start any discussion could be on the best one to use from a "legal" point of view!
As One impose to start and the end of the previus shift, just one second before the end. The other method the start is at the beginning of the next shift.
So in the first You can say I am ready to start, but we arrived at the end of the working shift, so the work can not start. But this is pure philosophy.

PS I noticed before this difference reading different books and attending to different seminar, but I never thought we can say are two different methods as the different look like formal one. A formality that on a pc algorithms, as per Ronald statement, oblige to have two different algorithms.
I do not want enter in the discussion on out-of-sequence, multiple calendar, etc. But I like to consider them as not real added feature, just like constrain to the problem. Like in a mathematical problem, You have Your system of equation (linear, non linear, etc …) and then the boundary condition to the system.

The point of the article is that the two different methods are currently being taught as the only method to calculate a CPM. The two different methods produce differnt work days for the same activity. This is a significant difference.

As I said most clearly, the referenced article still does not address other issues such as how to handle out-of-sequence progress, multiple calendars, lag calendars, interruptible activities, status date considerations, etc. These are not ’added features;’ they drive to the core of the agrument of being able to use the CPM for analysis. Most experianced P3 users understand this.

I will work on developing the rest further. I am not the only person in the world who understands these concepts. This is not a secret science. I would appreciate a little help here. It is time for someone else to step up to the plate.

I have prepared copies of my reviews of this matter as published in the AACEi’s journal, Cost Engineering and placed them on my website at www.RonWinterConsulting.com/published.htm. The article that directly describes the two methods in use is called, How to Befuddle a College Professor (without Really Trying). The other AACEi article is the textbook review that introduced me to this particular issue of CPM calculations.

The above referenced article still does not address other issues such as how to handle out-of-sequence progress, multiple calendars, lag calendars, interruptible activities, status date considerations, etc.; but it is a start. I suppose that when I next speak at the Primavera User’s Meeting, the AACEi local, regional, or international meetings, or at the PMI College of Scheduling international meeting, perhaps I should consider this topic.

The articles appear to not be public any more. Professionals still have to make a living and that often than means not angering your support base. I will work on compiling my own list but even I have to make a living. I cannot do for you in a couple of lines of text that has taken me years to learn.

These sorts of things are discussed in detail at various professional meetings. Perhaps you should consider getting more involved with organizations such as the AACEi, PMI, Primavera User Conventions, etc. A profession requires professional participation, not just an check of the chat boards.

Good luck!

Ron Winter.

Ron (or Paul), could you point me to one of the papers you referred to on Paul’s site discussing the differences in the CPM calculations? I only found book abstracts for each system (and they did not appear to analyze differences between systems).

Bernard Ertl
InterPlan Systems - eTaskMaker Project Planning Software

Ronald,

Totally agree with your point of view. And it seem we have a sales person in the discussion. I am not against new software but this forum is to disucss MICROSOFT PROJECT DISCUSSION and that’s what we discuss. If others would like to discuss spider sw. It’s welcome to open a new thread.

Your faith is founded on the popularity of your software, on the fact that this issue is too easy to do wrong, and you have never noticed any differences? This DOES sound like a lot of faith is involved.

Paul Harris has done extensive research and has published many fine books and papers on the differences in the approach that MSP and P3 take to calculation the CPM. His website is at www.eh.com.au.

I have published a paper about their being two distinctly different methods for CPM calculation in use today in the October 2003 edition of Cost Engineering, the Journal of the Association for the Advancement of Cost Engineering (AACE) International. Their website is located at www.aacei.org. Anyone who is serious about professional scheduling should be a member of AACEi.

Lastly, forget about your bells and whistles. If you are not well grounded in the basics of CPM theory, then resource and cost leveling and all of the extra add-ons might help with sales but this won’t make your product any better. If you do not fully understand the differences in how MSP calculates a CPM schedule versus how P3 does, then you do not know enough about CPM. Good luck.

Ron Winter.

Where does this FAITH in the purity of CPM come from? Has Microsoft ever published the exact procedures that it uses in MSP to calculate the CPM? Has Primavera? The answer is “No,’ they have never published how they compute the CPM using multiple calendars and out-of-sequence progress. Forget the constraints and resources; that is another issue entirely. I am just talking about plain old CPM, but with normal statusing conditions found every day on any project.

Is it because that you were taught only one way to calculate the CPM that you believe that there can be no other way? Would you be surprised to learn that two different methods are being taught today, sometimes in the same university at the same time? So I ask again, since you are arguing not from facts but from faith, why do you believe this fallacy so strongly?

Ron Winter.

Last post I will put in this discussion.
I am with Vladimir and Alex point of view, the theory is the same, but different mathematical engine and different constrains at the problem.
As is a simple linear mathematical problem (system of equations) more and more with the increase number of activities. So the two engine can use different solution methods (first possibility of differences), different precision ... etc ...

One more point, when the CPM bring life the only relationship was possible to use were the finish-start one.
After the method were extended and different kind of relationship were developed or allowed (Metra Potential Method).

With the same set of relationships in the CPM network both packages (and other packages as well) will produce the same CPM schedule if resources are not restricted. It is pure and easy mathematics.
If there are resource constraints project schedules may be different because all packages use different heuristics.

I am not 100% agree with Your statement "The CPM algorithm used by MSP is NOT the same as used by P3. Only a superficial comparison will show otherwise", as also after few line You come to my side.
The theory set up years ago is the same based on the forward and backward steps, but the package has different boundaries option, like as you write about the resource calendar and lag, and the relationship limitation of MSP, etc.

While I am agree with Your statement "MSP computes progressed activities differently than P3"
This is one of the main reason if I am using it is just in the early stages for presentation purpose, after I do all using P3.

I assume that the main question was if they use the same theory and I still convinced yes, but after all the boundary condition that the two packages allow You to set are are different and this is giving the difference in the scheduling.

Agree that with small schedule the difference can be not appreciable but with huge one you start to have "discrepancies". All the two are good and incorrect in the same time.
With is the degree of confidence of Your constrains, durations, etc ...

The CPM theory is the same for both software.
Forward and backward pass are giving Your Early date (first) and late one (second) with floats.
The differences You can have time by time can be due to different calendar set up.
I try a simple example (taking care to have all the boundary consdition set the same) and I had the same results, as should be.
I am agree, more over, it can be a computational error due to the engine, as both use a mathematical engine to solve all the equations. So the difference can come from this as well as You are using different precision in the two package.
But if You do a simple example by hand, with Project, P3, P3e or Artemis and so on, the results must be the same.