| Author | Message | ||
     Peter Hudson |
Hi, We would like to know the experience of different users of estimation tools (vendors please keep out of the discussion in order to retain objectivity). Requirement is that we provide size of project in FP units plus a whole bunch of other necessary parameters (e.g. technology, type of development, ...) and in return the tool gives us the expected costs (mandays) and the optimum duration for the project. The tool should be able to use our own historical data to calculate the outputs. In the absence of our data (as it takes time to build a repository) the tool should have it's own set of industry data. This means that we need to have the possibility to tell the tool which data is to be used for estimating the costs and the duration. Anybody already using some tools and with what level of satisfaction? Your experience can help us in our decision process. Many thanks in anticipation. | ||
| Lonnie Franks |
Peter, I started using SLIM from www.qsm.com in 1989 (when it was running on Lotus 1-2-3). It gives very good estimates and an ability to track progress. I've also used Knowledge Plan from www.spr.com. Both SLIM and Knowledge Plan are relatively expensive, but well worth the money for 1000 function point projects. For small projects I use Estimate from www.construx.com. Lonnie | ||
| Peter Hudson |
Hi Lonnie Thanks for your feedback. Are these tools able to use your own historical data and provide future estimates based on them? Also: I did not understand why you use construx for small projects? If you already have SPR or Slim despite the high price tag, then why do you need to use a separate project for smaller estimations. Did I miss something there? Many thanks. (Message edited by peter on December 07, 2009) | ||
| Ashok Kumar |
Peter, I have SLIM Estimate, COCONO-II and RASS Estimate. I have also compared Estimates from SLIM with COCOMO and FP/Productivity based model. You can see comparison graphs on http://www.ifpug.org/discus/messages/1778/7935.html#POST8808 SLIM can use historical productivity rates. However, if you have just historical costs and no size productivity rates, RASS Estimate's COCUSI model is best as it uses only historical costs and does not require size productivity rates. If you want to know more about SLIM or RASS Estimate, please drop me a personal message. Ashok Creator of RASS Estimate www.rasstools.com | ||
| Peter Hudson |
Thanks Ashok. In my case I have both size and cost (and lots of other information like duration, technology, ...) So getting historical data is not a problem. I just want to be sure that the tool has inbuilt algorithms to churn the data and to determine a productivity function in order to propose me the best possible estimates for future size based estimation. | ||
| Ashok Kumar |
Peter, Most estimation tools I am aware of use single productivity figure and give effect of size and environmental factors. This works well for a limited range of size say 10 fold but will certainly fail if size varies by 100 fold. If you have historical data of size and cost, you do not need to consider environmental factors as these had already impacted the cost figures. I have come with an approach where you can create cost curves and grade the curve with size on five points scale. Now while estimating you have to select which cost curve you want to apply and what is the expected size of the work. The approach has been neetly packaged into CoCuSi (Cost Curve Simulation) model. I think this is what you are looking for. With your approach, you can do estimation in the quickest possible time and minimum effort. Let me know if this http://rasstools.com/CoCuSimodel.aspx reference helps you. Ashok Creator of RASS Estimate www.rasstools.com | ||
| venkataraghava |
Obviously like Peter, I was looking for tools that does FPA estimation and use Historic productivity data to derive effort. Here is my experience in last six months: - Not able to find such combination in one tool among vendors. I end up working on simple FPA estimator tool (J2EE based) and created workflows to apply historic productivity factor (based on size) to derive effort - Productivity does vary based on size of the project. This are my findings upon studying numerous size projects and varies based on technology applied, it will never be linear. Thanks Ashok for your detailed analysis and the model - Now I am thinking to build estimation model into my tool. Once the historic data get captured, estimation model can be taken out Regards to all, | ||
| Linda K. Hughes |
We have been successful with estimation using Function Points along with other key input parameters via the Cost Xpert tool. The tool is based upon a historical data base of numerous technologies. You can down load a trial copy from their website. Linda K. Hughes | ||
| Peter Hudson Username: Peter Post Number: 130 Registered: 02-2004 |
Thanks Vernkataraghava and Linda for your additional input. I will most certainly give a look to Cost Xpert. I had thought that (at least in the past) it worked with SLOC as input because it is built on CoCoMo model. My preference is to input FP and not SLOC. But I will certainly give Cost Xpert a re-look. Venkataraghava: I know that productivity depends on several factors, one of them being the size of the projects. And I did not expect the formula to be linear. Mathemetical algorithms can be built into tools to analyse existing data and then based on a new data pinpoint with sufficient precision in what range it would fall. At least that is what I expect from a professional estimation tool that claims to use historical data. Thanks for your help and suggestions. |
