| Author | Message | ||
     Rae A. Hamilton |
Hello IFPUG Members, It has been awhile since I was last here, some of you may remember me as Rae Ann Burns. I am with a new company who does not use FPs as a software metric. Currently have an issue with a data warehouse project that was just implemented; number of outstanding issues that need to be resolved and implemented as a result of this project. The issue is this, determining priorities of O&M versus discretionary enhancement work to the application. Estimated that there is over 1000 hours of O&M work to be completed to make the data warehouse environment stable. And another 4000 hours estimated of discretionary project work. IS supports 2 different sets of users that will use the data warehouse and each set of users has their own priorities of what needs to be done first; O&M or discretionary. I believe there is value in FPA. How can FPA be used to assist in what gets worked first; O&M or discretionary or a combination? How to satisfy each user group; customer satisfaction, needs being met? How to use FPA to assist IS in determining support levels for O&M and discretionary? I don't have my IT Measurement book here at work; I know that this is a valuable resource and would probably answer some of my questions. Since I don't have it here at work I am here within the bulletin board to seek advice from the experts. Thank you, Rae | ||
| Steve Neuendorf |
Hello Rae Ann, There are a lot of things you are not saying. What is your shop capacity? Is 5000 hours six months work or several years. What is your ratio of estimate to actual? Will 5000 hours take 3000 hours or 15000 hours based on experience? What are your project sizes? Are there a few very large projects, or lots of small ones. Have you analyzed the relationship between projects? The biggest opportunity is the catalog of functionality offered by using FPA. With that, you can accurately size each project and identify any relationships between projects. Typically, any project is less than 20% value added work. So say you had two 40 hour projects that could reasonably be combined, into one 50 hour project. It would not take too many of these to completely pay for a full set of baselines and project counts. Say you found that several projects that enhance the same report and another project that replaces it. The risk is not that it is not cost effective, but that it is embarrassing that this type of approach has not been used before now given its wide use and well publicized benefits for over 25 years. And of course you can fully engage the customer. You guarantee the capacity, and they identify the size and priority. Good Luck, Steve | ||
| Rae A. Hamilton |
Thanks Steve. The questions you ask in the first paragraph are unknown to me; have been with the company for less then 6 months. What I do know is that projects are worked as releases; so requirements/projects are estimated and scheduled for a release. Release schedules for the remainder of this year are 10/01, 11/06 and 12/16 (implementation dates) with 5 weeks (estimated) given to testing for each release. I haven't been here long enough to get a grasp of the productivity rate for projects. Don't know what the ratio of estimate to actual is for project work; don't even know if this is being measured or captured....I have a lot of unknowns yet. I think this company would be a good candidate for FPA, especially since they are working towards implementing CMMI and process improvement with projects. I guess I was looking for existing benchmarking data that I could use to show "proof of concept" and to assist management in moving towards managing what they can measure.....isn't that a Bill Hufschmidt button.... Rae | ||
| Charley Tichenor |
I'm not sure that function points can be used to help determine which project development efforts are most important. The function point count gives the size of the effort or the application, and the size variable can be used in various formulas and ratios to forecast cost or schedule, or do benchmark comparisons of productivity, quality, or other quantitative analyses. But I see importance as a value judgement, not something that is a function of software size. You might try ISBSG.org for good function point benchmarking data. I have a number of their publications and we have used them here for several important management decisions. Charley | ||
| Marie O'Neill B. Comm. MICS CFPA CFPS CCP |
Hi Rae, Charley, I know what Charley is saying about FP not determining importance, but I've certainly used FP to present options to client groups, so that they can see the relative size of 'bundles'of work for releases and then its amazing how quickly they decide what is 'urgent' 'important' 'we can wait' (especially if its their own budget that gets hit. It also passes the decision to the bill payer. I don't know if that would help Rae? | ||
| Pranay Srivastava |
Dear All, Project size is a pointer of the project cost. And project cost can be used to determine which project yields best return for the organization and thus which should be selected first. Project size determines the effort required to support an application. It you have fixed effort available then theoretically you can determine the support level that you could provide. But none of my previous organizations have done that. Hope that helps, Pranay | ||
| Peter Hudson |
I am not sure that my posting would be what Rae is looking for. But typically I would work with several FP based indicators for discretionary work and maintenance work. For example: how many full time equivalent I (or you) use today to maintain the production function-points. How many FTE I would need and for what duration for a project. And once the project is delivered, how would it impact my first number (i.e. number of full time equivalent per year to maintain). If there is an important influence then it makes sense to give priority to the project. Of course this works mainly in case of a project "improving" existing functionalities. If a project is meant for delivering new functionaltiies then some other means must be used. In all cases keep in mind that there can always be exceptions. One can always have an extremely high cost project (because of high technology investment) but very low functionalities. This can still have a lot of business value. Thus, it can be interesting to measure the business value delivered per Function Point. The challenge here is to define objectively "business value delivered". I would start with a simple definition: like amount of additional annual revenue the business hopes to achieve by this change (the estimate can thereafter be replaced by actuals if they are available). I wish you the best of luck, and do remember to share with us your experience if you try out some of the suggestions. | ||
| ALOK KHARE |
Is it a normal practice to measure sub-processes (SDLC phases) productivity? In our organisation, we execute complete SDLC / only construction / only Testing projects. We face difficulties in consolidating productivity data from different projects, so we are planning to capture productivity (in hours per FP) for the sub-processes i.e SRS, Design, Contruction, Testing, etc separately. I would like to know whether it is a normal practice and are we going in right direction? | ||
| Pranay Srivastava |
Dear Alok, You are going in the right direction. You have 2 options: - Capture productivity for each phase - Capture full SDLC productivity and use effort distribution to break-up and get the effort for each phase. The problem here is how do you merge data for partial SDLC projects into this Regards, pranay | ||
| ALOK KHARE |
Thanks Pranay; Currently, we follow 2nd option i.e. using effort distribution. But merging data and accuracy is the major issue. Also setting productivity improvement target is also difficult as different phases are excuted by different expert groups. So, we are moving to option 1. Could you please let me know organizations, who are maintaining phase wise productivity data or can we get some white paper on this? Regards, Alok | ||
| Pranay Srivastava |
Dear Alok, We have used option 1. What help do you need? Your organization capability would be x FP/hr for complete project. The effort distribution would be - Analysis..................a% - Design....................b% - Coding & Unit testing.....c% - Integration Testing.......d% - System testing............e% - SCM.......................f% - Project Management........g% & so on Hope that helps, Pranay | ||
| Ron Bora |
Can anybody tell me from where I can get the CPM. I m a member of the IFPUG Bulletin Board. Is it enough to get the CPM manual!!! | ||
| Rashmi Sinha |
Hi Ron, Please visit www.ifpug.org homepage and you'll find "MEMBERS ONLY" drop down box. Please select the CPM downloads option. You would need to keep your username and password handy. Hope this helps. Thanks & Regards. Rashmi Sinha | ||
| Rashmi Sinha |
Oops...apologies....you said you are a member of IFPUG Bulletin Board...You need to have a valid IFPUG membership. IFPUG Bulletin Board membership alone is not enough. Regards Rashmi Sinha | ||
| Ron Bora |
Thnx Rashmi, Can you tell me how can I get that! Is there any payment required for tht or its FREE! | ||
| Rashmi Sinha |
Hi Ron, It requires a fee payment of $250. Would suggest you visit http://www.ifpug.org/membership/ for more information. You also have an Online Application Form available there. Kind Regards, Rashmi Sinha | ||
| Ron Bora |
Rashmi, Thanks for the link! Regards- Ron | ||
| Mridul Das |
I am a Professor of an Engg College. While trying to register my college as a member of IFPUG(which is 'free' as it falls under University/College category) it didn't get registered. I finished typing the Online form and didn't enter anything in the Payment section as the registration is free. Lastly when I clicked the submit button it was not responded. I want an IFPUG member or IFPUG organization's official to reply what's wrong? | ||
| Mridul Das |
I am a Professor of an Engg College in eastern part of India. While trying to register my college as a member of IFPUG(which is 'free' as it falls under University/College category) it didn't get registered. I finished typing the Online form and didn't enter anything in the Payment section as the registration is free. Lastly when I clicked the submit button the following error comes- Missing value for Credit Card Type Missing value for Card Name Missing value for Card Number Missing value for Expiration Date I want an IFPUG member or IFPUG organization's official to reply what's wrong? | ||
| Linda K. Hughes |
We are not sure how the website is supposed to read universities, but they are indeed free for membership. Contact the IFPUG office specifically: Christopher L. Decker Association Coordinator CMA Association Management 191 Clarksville Road Princeton Junction, New Jersey 08550 Phone: (609) 799-6000 ext. 33 Fax: (609) 799-7032 cdecker@cmasolutions.com Linda K. Hughes | ||
| Renuka Marwah |
Hi Pranay I read your post where you mentioned "Your organization capability would be x FP/hr for complete project. The effort distribution would be - Analysis..................a% - Design....................b% - Coding & Unit testing.....c% - Integration Testing.......d% - System testing............e% - SCM.......................f% - Project Management........g% & so on " I have a query. If suppose my application size comes to be 2000 FP and say my orgn productivity is 18 FP/MM, then, will the total effort of 112 MM include all the stages of the SDLC, and activities like SCM and Project Management also? | ||
| prakash kumar sk |
Hi Renuka, The effort distribution depends on your organization metrics methodology. If they have used all the abvoe stages to consider productivity, then you can use the productivity of X/PM and divide the resulting effort Y to all the above mentioned activities. Normally the effort considered for productivity calaculation includes all the above mentioned stages | ||
| Luigi Buglione |
Renuka, the answer to your question is: yes and - as Prakash said - it should be modeled on your own way to decompose processes in your Quality Management System. But if you're looking for some benchmarking data, for instance in the ISBSG repository r11(www.isbsg.org) there are 7 main phases (Planning, Specification, Design, Build/Programming, Testing, Implementation/Installation, Project Monitoring & Control) and you - choosing the filters of your interest - can determine a reference % point on such distribution. For a proper analysis you've to take into account is that the current 'nominal' productivity formula (FP/effort) relates a 'product' functional size (FP) with the overall 'project' effort (coming from FUR and NFR, including also the organizational tasks). So, let's suppose that during the project lifetime, in order to improve the quality of your final software you'd spend some more time for quality assurance or project management tasks. The effect would be to formally obtain a lower 'nominal' productivity level, because the upper part of the formula (FP) is not directly affected while the lower part of the formula (effort) will grow, producing a new 'productivity paradox'. A full discussion with some proposal for overcoming this situation are in a white paper (going to be updated by the end of this year) on this issue is available @ http://www.geocities.com/lbu_measure/fpa/fsm-prod-120e.pdf Hoping it can be of some help. Best regards, Luigi ____________________________________________ Luigi Buglione, Ph.D., CSMS Level 3 SEMQ: www.geocities.com/lbu_measure/ - www.semq.eu MIS 3/e: www.francoangeli.it/Ricerca/Scheda_Libro.asp?ID=8054 ____________________________________________ Luigi Buglione, Ph.D., CSMS Level 3 SEMQ: http://www.geocities.com/lbu_measure MIS 3/e: www.francoangeli.it/Ricerca/Scheda_Libro.asp?ID=8054 |
