Linking of Summary Tasks

From: http://www.stakeholdermap.com/ms-project/link-summary-tasks-ms-project.html

The PMI's scheduling best practice standard does not recommend the use of links on summary activities. Here is what they say:

 

It generally is not recommended to use links on summary activities because the logic can be difficult to follow and the practice may not be supported by all scheduling tools. Use of links on summary activities may produce logic errors and create circular logic within the schedule model.

Practice Standard for Scheduling – Second Edition, 2011, Project Management Institute, Pennsylvania, pg 38.

In bold my friends, take the bull by the horns!

RE: It generally is not recommended to use links on summary activities because the logic can be difficult to follow.

• Well implemented Pacing Hammock functionality should be able to display all links and therefore logic shall be as easy or as difficult to follow as in any other schedule without any Summary Tasks (Pacing Hammock) links.

Re: The practice may not be supported by all scheduling tools.

• If we only allow the practice supported by all tools most probably we will end with nothing.
• Incredible it is naive to believe all software will behave equally or very similar if you limit the options.
• This shall never be an excuse to ban functionality that works well under the hands of a good scheduler, let the end user decide what functionality he needs.
• Guess optimization algorithms will be banned because most tools cannot handle them.
• Guess Multiple WBS will also be banned
• Guess variable partial quantity and workload will also be banned
• Guess a very long list of unique functionalities of each software will be banned.

RE: Logic can be difficult to follow

• Any complex project is a complex spaghetti no Gantt chart can make it visible, impossible to follow thousands of link lines across thousands of activities.
• You need a links table to see the logic as well as you need the activities table. Software lacking a links table shall be declared as not apt for good practice.

RE:Summary task links can produce errors in logic often artificially extending the plan

• The examples provided represent different logic, the error in logic by the planner, not the software, is because the poor intellectual capacity of the planner or lack of knowledge. Limiting the options to make it easier is not going to make the scheduler any better nor real life situations easier to model.

RE: Linking summary tasks can lead to bench time

• The example provided is poor, from a scheduler that created a link between summary task when he should have not. Who said all summary task got to be linked?

To me the PMI does more harm than good by looking for poor excuses, perhaps to favor their favorite software that cannot model the Pacing Hammock. This archaic vision promotes that software developers do not look for the Pacing Hammock, promotes not slowing but stopping the search for more and better functionality.

The PMI should be promoting education of the scheduler, not the pampering of those who shall not be doing scheduling.

I am in no way surprised Primavera/ORACLE promotes users of Primavera Risk to avoid linking summary tasks, I suspect it is because P6 cannot handle them. 

RE: I still feel, that the Mr. Plotnik’s definition of the Pacing Hammock also to do with some algorithm of calculation of it’s duration, that is not just summary task, but maybe I am wrong.

• He is talking about his preference for roll up calculation. I do not play much with it but believe most software give the user some control on how roll up is done.

RE: Anyway, the task is very simple now – just convince Vladimir to implement the feature.

• I have tried many, many times without success.

RE: But so far, based on his response in the parallel thread, I can see that even he treats summary task and WBS Phase as the same thing.

• No they are not the same thing otherwise I would not be asking for the Pacing Hammock. Vladimir has no other option than to convert MSP Summary Task into something Spider has, cannot be a Pacing Hammock as Spider does not have such thing. Can be a Spider Hammock that can create even more confusion, within his options he opted to eliminate the activity and group the child activities under a same WBS Phase.
• I do not believe P6 have the Pacing Hammock, maybe Primavera Risk has it with a different engine originated from other developer. Primavera Risk in conjunction with P6 is kind of an hybrid car.

RE: For me personally linking to and from summary tasks is just another flexibility, which you may have in the tool and which you may or may not use. So, I can see advantage of having it, however I can see, that it will conflict with multiple WBS structures feature, which you also praise.

• For me it is not just that I want to use it it is also about the logic lost when importing from MSP.
• I cannot cover my eyes and pretend MSP universe does not matter. There are many MSP user that can make good schedules and trowing down the drain a substantial amount of their schedules logic does not makes sense to me.
• In no way a Pacing Hammock will have conflict with multiple WBS, it is just an activity. What is impossible is to have multiple outline views as implemented in MSP as each outline will have different summary tasks with different links invisible to the user, different logic even when not clearly visible makes no sense at all.

RE: Representing a schedule from different point of view can be achieved by using grouping and sorting features of MS Project, though I agree, that multiple WBS is more flexible.

• At a single mouse click Spider can create such views from unlimited number of codes while other software have severe limitations on the number of codes available. Believe MSP is limited to 10 and calls them "outline codes". Other software combine the amount of codes with the available code values to establish the limit.
• In Spider just click "Create by Codes ...".
• 

Best Regards,

Rafael

Evgeny,

Could not be better, his 3 pumps example is typical, is similar to the 3 finishes of Building A in my example. The installation of the 3 pumps is independent, the delivery might  also but the 3 activities are under a common Pacing Hammock. Links can come in and out directly from the individual activities as well in and out the Pacing Hammock.

The details he speaks about coding I do not understand, I care from the standpoint of the end user not the programming details which can be way complicated even for the expert programmers. Looks to me that the programming of the Pacing Hammock in the absence of the awkward outlining method is difficult for those who abandoned for good this method. That the outlining method is awkward does not invalidate the concept in any way.

In reality all are outline methods and Work Breakdown can model all, actually Spider WBS functionality can model Activity Data Items using unlimited number of user defined activity codes/fields, if you add Pacing Hammock it will be able to model a superior Ouline method as it will have all the functionality of the WBS method in addition to the Outline. You shall be able to move activities at will without changing internal logic of child activities, and in the same way an unorganized job can be difficult to understand you can do it, but with the links table logic will always be visible.

Regards,

Rafael

Evgeny,

Converting an active MSP Summary Task ( activity) into a WBS Phase does not makes sense to me. MSP Summary tasks are activities that should be converted into Pacing Hammocks (can be called a Spider Summary Task). But traditional Hammocks cannot be linked while Summary Tasks can be linked, more functionality is needed, something else is needed. A MSP activity shall be converted into some Spider activity not a WBS Phase.

Although not shown MSP Summary Task are linked to child activities. In MSP when you move child activities these links are changed, this I find insane, invisible links defined by some position of activities.

Pacing Hammocks are similar but more functional than traditional Hammocks and MSP Summary Tasks. Pacing Hammocks shall display all links while MSP Summary Tasks misses displaying some links, is incomplete functionality working on an organizing system that is awkward.

If you transfer the following MSP job into Spider too much of the logic is lost.

To get back all logic from the 4 Summary Task converted into WBS you  must first create 4 hammocks that will be defined by 8 new milestones, 2 per hammock. Each of these new milestones shall be linked to every child activity, 2 new links per every child activity, this adds up to 24 new links. In summary 4 new hammocks, 8 new milestones, 24 new links. And this is for a ridiculously simple job.

To me, this is not good enough, better not providing such incomplete conversion.

Best Regards,

Rafael

RE: In Rafael's Subject, The question is, is the Summary Task will control the Dates of the Activities or the Activities will control the Dates of the Summary?

Mr. Plotnick's Pacing Hammock (nothing new just a new name) same as MSP Summary Task and same as SureTrak Topic Activity is considered bad practice by some and I am asking why if it can even be modeled with any software after much trouble using 2 milestones and a hammock?

It is a procedure that can be programmed by all developers in the same way as MSP, Open Plan and SureTrak developers did.

I am asking for a single sample job using free Open Plan where the model is wrong or does not works as expected. As up to now I have only found the model do not show some predecessors and that if you move around summary tasks or its member activities the logic might be changed. The outline model is prone to create user errors in logic when moving activities, these are not software errors but user errors due to the inherent behavior of the outline method.

I happen to agree with Mr Plotnick's that the issue is on how it is modeled by actual software. As far as I remember he was one of the old time Gurus that once questioned why they cannot follow the logic in their schedules.

I believe if the logic is there, and it works, then it should be disclosed by the software developers. The problem is on the software not in the logic. As long as circular references are rejected it works, it works in SureTrak, it works in Microsoft Project and works in Open Project. The problem is they do not show full predecessor logic, it is there just they have not figured it out. Mr. Plotnick suggest his coding system is needed to get the Pacing Hammock to display full logic, I am not sure. I have my doubts and believe it is on the will of the developers to observe and react to what hundreds of thousands do with MSP. MSP developers are also included as they still miss to display full logic.

I am not sure the links displayed by P6 work or are merely decorative, similar to Spider links with Hammocks. This I believe is also misleading. If hammock links do not work shall not be displayed.

Best Regards,

Rafael

Evgeny,

About Spider WBS Phase:

• You can have WBS Phase within another WBS Phase.
• 
• Phases cannot be linked and cannot have cost nor resources assigned.
• WBS Phase is not an activity.
• A Summary Task, not available in Spider, is a Task/Activity, not a WBS Phase.
• WBS is a hierarchical system for organizing activities.
• On any given dictionary / structure an activity can only be under a single WBS Phase.
• You need multiple WBS dictionaries / structures in order to organize an activity under different WBS Phase.

About WBS and Outlining Methods:

• Some time ago I asked Vladimir to include Outlining method in Spider and learned latter that it can yield to unexpected results. Results that will respect the rules of your software but that might change the originally intended logic.
• Primavera nor Spider support the "Outlaw" Method, it is good they do not, but Spider makes it many times as good as it allows for multiple WBS structures.

Summary Tasks are not the traditional hammock:

• Child/Member activities of a Summary Task can belong to another Summary Task within the same WBS Structure, within the same or different WBS Phase.
• Summary task can be linked to one another but there are certain limitations on the type of links allowed. You ca try for yourself using MSP or Open Project.
• Summary Task might be resource and cost loaded but will not be considered in resource leveling. A limitation that makes much sense if you are to make use of efficient resource leveling algorithms.

Spider Hammocks:

• Are defined by two activities, one defines the start and the other the finish.
• Hammocks might be resource and cost loaded but will not be considered in resource leveling. A limitation that makes much sense.
• In Spider hammocks can be defined by activities belonging to different and independent paths.
• In Spider the activity that define the start of the hammok can occur befor the activity that defines the end of the hammock allowing for the "vanishing hammock".

Summary Task shall provide for better compatibility with MSP just import an MSP schedule into Spider and you will see summary tasks are converted to hammocks with most logic lost.

Summary Task implemented as what Plotnick calls a Pacing Hammock shall provide for the linking between other Pacing Hammocks as well as to/from other activities. If well implemented it shall provide a view for the missing logic not shown in the Summary Taks as implemented by MSP or Open Project.

Best Regards,

Rafael

Mike,

To display pictures that do not vanish I use an external file hosting, I use Photobucket which is free and now without space limitations and know other PP members are using different provider without any problem. I even do have a Photobucket app on my iPhone.

I enter text using rich-text, for posting the links I disable rich-text and paste the HTML code, to get the view I activate back rich-text.

I usually write XXXXXXX in rich text format the place where I want the picture, then when I disable rich text I can see where to put the picture link in HTML code.

Gary,

I am interested in your figures, please try to post them in a way they remain visible.

RE:  it's very dangerous because all too often one or more of the summarized tasks actually needs a different link to the rest of them.

• I do not see what is the problem other than how limited the outlining method is, it is 100% dependent on the placing of your activities, in any case outlining method as implemented shall be banned from CPM practice, not the Pacing Hammock.
• I do not see any problem with concrete activities on different buildings be under a Concrete Summary Bar.
• I do not see any problem with the concrete of the different buildings de independent of each other.
• I do not see any problem with concrete activities depending on something else than concrete. Say footings depend on excavation activities, activities that might be under separate Summary Bar like "Earth Work".
• I do not see any problem with a requirement that can be contractual for a contractor to limit work on different structures to be occurring at the same time. But remember within concrete activities some might not be linked. The main issue with the outline approach is that it misses to display logic, it misses to display some predecessors, even in MSP, not to my surprise. If you look at the Open Project figure in one of my prior posting, sorry PP do not keep fixed references, you will see no logic is shown for activities 15, 16 and 17. But by magical logic they get moved! The problem is with the outline method not with the linking that can make so much sense.

Best Regards,

Rafael

Evgeny,

RE: For me this functionality is not that important. I use it reary and mainly in the initial versions of the schedule for rough estimations of durations.

• For me as well as for Microsoft that for decades has provided such functionality as well as for Mr. Plotnick that in his paper mention the desirability of the pacing hammock ii is.

RE: I have the following questions though: how can you combine feature of linking to and from summary tasks with Spider’s feature of multiple WBS?

• Summary tasks as well as any task do not belong to WBS, that in the event Spider ever implement summary tasks as Microsoft Project and other software do.

RE: P.S. by the way, the main thing I would like to see in Spider in relation to summary tasks is to be able to place them at the location of my choice (Spider always places summary tasks below the none-summary tasks of the same level see here)

• Spider never places summary task at any place, Spider simply do not have summary tasks, it cannot place what it does not have. What you see are WBS Phases. WBS phases and Sub-phases that are located via their member activities, it is their member activities that have links. If there were pacing hammock these by definition, as per Mr. Plotnick, could have predecessors and successors within certain rules. They would be activities but not Phases. Like any activity could be assigned multiple WBS, will be placed in the respective WBS depending on the dictionary (structure) you are viewing. The logic shall be visible in the Links Table. In schedules with hundreds of activities the link lines are useless. Software that do not display a links table I consider primitive, CPM is about logic, no logic is bar charting.
• WBS members as well as pacing hammock child do not need to be linked one to each other, any member activity can finish last while in the traditional hammock this is not allowed. BTW Spider hammocks are somewhat different and more functional than in other software, the pacing hammock have even more functionality.
• Because Spider Hammocks are not required to belong to the same path with the start activity prior to the finish activity this added functionality allows for some strange hammocks, no longer the Twisted Hammock is available, my favorite, but still some strange occurrences can happen.
• 
• I still like the idea for Spider Hammocks to allow for lag. Good for modeling cash flow off-set, surely could be used in combination with the pacing hammock.

Best Regards,

Rafael

Mike,

Can you show it. I do not understand you. A figure can spell a thousand words. I have heard such argument but never seen a visible example of the claim. I do not see any difference between the following two schedules, both are equivalent and both show same critical path with the addition of a few dummy milestones on the long method, more prone to error than the first.

I have always found lazy those who make claims without taking the time to prove it. Maybe Asta is so cheap it cannot display criticality but even free Open Project can.

I agree with Mr. Plotnick that good coding will allow for the Pacing Hammock, with smart use of programming it can be done. Prove it with Open Project, it is free and everyone of us can learn from you for free.

Best Regards,

Rafael

P.S. Meanwhile I will go for another touch and go in our neighboring islands with my son "driving" blindfolded, except on landings, he is learning to fly by instruments. The video is from early today.

http://www.youtube.com/watch?v=pOI2uEq-EPs&feature=youtu.be