| Author | Message | ||
     Carlos Eduardo Vazquez |
Dear Fellows and CPC, Is it correct to assert: a) The very same software can have different function point counts depending on how the boundaries are placed. b) There is not a unique, correct, boundary positioning detached from a particular user view; this subjective user view is the foremost reference to identifying all the function types. c) IFPUG defines that boundary positioning is rather a decision, a choice among different options, than the way the software solution is packed. Best regards, Carlos Eduardo Vazquez, CFPS carlos.vazquez@fattoCS.com.br Fone: (27) 8123-9100 begin_of_the_skype_highlighting (27) 8123-9100 end_of_the_skype_highlighting / (21) 9750-1259 FATTO Consultoria e Sistemas http://www.fattoCS.com.br/ Medição, Estimativa e Requisitos de Software Vitória: (27) 3084-7304 Fax : (27) 3223-8472 Brasília: (61) 3717-1974 São Paulo: (11) 3717-2974 | ||
| Abinash Sahoo |
Carlos, You need to make one point clear that when you are saying 'software' are you talking of an integrated software package that might include several 'applications' talking to each other. Or you meant just an application only. Just to refer the examples of CPM, an HR application talking to a Finance application and a couple of such applications might be packaged into a Organization level software system. My understanding on your questions : a) So if you are talking of such a software system, then the FP counts would be dependent on which application you are counting, whether it is the HR or finance application. But I am not sure if it would be correct to say that the sum of the FP of individual applications would be equal to the FP of the overall software. Just because fundamentally counting FP for an individual application would have to be done by defining the boundaries at the application level and if you have to count the FP for the whole Software , then the boundary gets redefined to encompass all applications in which case the inter-application transactions would not get counted if they do not cross the boundary. b) I do not think there would be anything called as a 'correct application boundary' or 'incorrect'. Boundary would have to be defined from the user's point of view, what data storage requirements the user has , what transactions the user would want to maintain data. For the same application different users would view differntly and hence there would be differnt boundary for them. CPM mentions one such example about HR application having the HR user and security user. Both have differnt needs and hence FP counted for each user separately depending on what data they use and what transactions they have access to. c) I am not sure what exactly you mean by this statement. May be the above two explanations help in answering this as well. Else you will have to wait for some experts on this forum to answer.  Hope this helps. Abinash. | ||
| Steve Neuendorf |
Hello Carlos, a) Very different. Since FP is largely based on boundary definitions and events, the granularity of a portfolio of applications and projects defined by boundary decisions is a significant determinant of the number of FP. Your FP do not equal my FP because of boundary considerations. Hopefully your FP counting process equals my FP counting process, but FP is operates as an interval number system. We would need to consider the variable "magnitude" or "total size" to treat FP as a ratio type value. And, of course, that impacts the dependent variables also; productivity for example. Here is an interesting blog post (3 parts) that discusses this: http://www.qsm.com/blog/2010/size-productivity-paradox-part-i/index.html b) True, it is entirely up to the counting organization to decide the counting boundaries. There is a dependence locally; that is how you make any boundary decision will influence other boundary decisions, but outside of a particular portfolio, there is no relationship or dependence. I have published a sources of variation matrix on this board a few times, and you may find that useful. Once you get data aggregated above company departments, size (magnitude) differences become a significant source of variation. c) True. It has to be a well considered decision too. I tell everyone that how you make the boundary decision determines whether FP counting is easy or hard and useful or useless. Most ill considered boundary decisions look like an application development or purchase history and have little relationship to how the applications are used and viewed by the users. These groups usually have difficultly with and realize limited value from their measurement efforts. The most important consideration, IMO, is the cost (or managerial) accounting system. With the right chart of accounts and cost data collection process, the boundary decision is easy and the results are useful and very valuable. If you really have not made a conscious effort to acquire the expertise that understands this, the suggestion "don't try this at home" should get some serious consideration first. Hope this helps, Steve | ||
| Carlos Eduardo Vazquez |
Dear Steve, It is always nice to hear from you, and I do admire and respect the way you articulate the ideas. In my opinion and for a long time, you are one of the more engaged FP professionals trying hard to drive people toward the right direction when it comes to FP counting. I, myself, do understand those three items I proposed as correct propositions. However, I'm facing some difficulties with a client of mine who understands the boundary is comparable to “gravity law”… He states, positioning of application boundaries does not come from a *decision* among a series of *possible choices*, but rather is engraved in stone by IFPUG CPM regardless of any business purpose… of any measurement and analysis goals. I do thank your reply acknowledging this understanding of mine, although it’s not an official IFPUG message, at least gives me a lot of reassurance I have the proper understanding of the counting rules regarding those three items. Carlos Eduardo Vazquez, CFPS carlos.vazquez@fattoCS.com.br Fone: (27) 8123-9100 begin_of_the_skype_highlighting (27) 8123-9100 end_of_the_skype_highlighting / (21) 9750-1259 FATTO Consultoria e Sistemas http://www.fattoCS.com.br/ Medição, Estimativa e Requisitos de Software Vitória: (27) 3084-7304 Fax : (27) 3223-8472 Brasília: (61) 3717-1974 São Paulo: (11) 3717-2974 | ||
| Anand Bharadwaj |
IMO: As per CPM 4.2.1, Boundary is defined as 'border between the software to be measured and user' Hence, in our count if we come accross an application that need not be measured, it could be considered as user and all interactions with it hitherto. The term 'software' unambigously defines whether it is 1 or many applicaions (existing or to be built) I think it is important to 'ignore the user' and analyse only the software to identify the boundary, as the 'user' gets defined by default.. Analogically, 'articles in the Ring is finite, but outside world is infinite' and which ever is not inside (i.e. outside) becomes (or has potential to become) an user.. awaiting review from Steve.. | ||
| Steve Neuendorf |
Hi all, First I want to thank Carlos for his kind remarks. I am glad I can be helpful. Looking at your client position, I guess I would want to know more. I am not so sure about the boundary rules in 4.3. Before, they were quite clear. In 4.3, I have to wonder if they may have just fallen off the table. The definition, example and discussion seem consistent with the prior version, but the rest of the discussion is just not there. I did a very extensive boundary evaluation with a client about 10 years ago and came to the conclusion that the boundary rules were quite broad. While they did define a region among the possibilities of potential boundaries, inside that region, any boundary definition would survive analysis by the rules (a tautology). And, you know something, just like in management, there is absolutely nothing wrong with having two (or more) sets of books, as in cost accounting and financial accounting; there would not be anything wrong with having two sets of boundaries; one that satisfies your clients' needs and one for your own needs. If you look closely at the 4.2 boundary rules and discussion in general, it seems a little ambiguous. The users define the boundary and the boundary defines the users. Since the consequence of any particular boundary decision becomes evident only after the boundary decision is made, the decision to change a boundary decision is made on a different set of facts than was used to make the original decision. My number one FP teaching aid is a pencil with a 3 inch lead and a 4 inch eraser. Seems to me the universe of users and functionality would be a finite set, and any boundary decisions would be a partitioning of that set. But that is probably one of those discussions best served with a few cold ones. Take care, Steve | ||
| Carlos Eduardo Vazquez |
Has somebody else got a position regarding this issue. When evaluating R3, one and only one function point comes from applying IFPUG rules, or the count varies depending on how the boundaries are placed according to the purpose of the count? Doesn´t the purpose of the count influence how the boundaries are placed among applications? It´s been raised a doubt regarding 4.3.1 changing boundaries rules. Does CPC have something to say? Is there no decision to be made since IFPUG dictates in an objective maner one and olny one way to place the application boundary? Thanks all. | ||
| Vajee Uddin |
I have similar query on determining the type and Application boundary The business need for a project is to supply the necessary data to the business, a more up to date information used for business analysis and decisions. Here Application 'A' is to supply the necessary data to Application 'B'. My counting scope is limited to supply of data from Application 'A'. 1. Can we consider the type of count for implementation of the FR's as a Development Project Founction Point Count (as a new application) or is it an Enhancement project function point count? 2. Can my boundary be between Application 'A' and Application 'B'? Or Within Application 'A', the boundary is between existing functionality and the new functionality? 3. For this count, as part of a project to meet the objective of data transfer from Application 'A' to Application 'B', if functional requirements to be implemented on Application 'A' needs to access data maintained through other elementary processes of Application 'A', do we consider the logical data on Application 'A' as EIF (as the new functionality does not maintain any of the logical files) and the output file to be sent to Application B as 'ILF'? The logical data is maintained by existing FR's / elementary process and were considered as ILF during the initial measurement of functional size | ||
| N.Venkateswaran Username: Nvenk Post Number: 270 Registered: 07-2009 |
Guys, this is where the industry is moving towards. (Means reusing existing applications and minimize the developement efforts by integrating different applications - SOA?). Answering to Vajee's question, you can consider point 2, if Application A can provide the data to application B for its new functionality. In this case an EIF and EQ are involved for new count. Application A count remains unchanged. In case Application A needs a change in fetching the data for Application B, Application A counts add as well. cheers. |
