National Board of Patents and Registration of Finland (PRH) is an experienced organization to outsource software and systems development. The PRH administers information of Finnish companies, products, patents and trademarks. The national membership of European Union since 1995 has brought also international projects with new stakeholders and requirements along. The organization has experienced great project success, but also some failures in the past.
In the process of looking new ways to develop software for their specific needs, PRH contracted with a new, innovative partner Alfame Systems Oy. The purpose of these projects was to build up two applications, using agile management and development methods (e.g. SCRUM) in 2013 – 2014. Feeling of success in these efforts was so strong that the PRH decided to make a triangle benchmarking survey on these projects, to find out if the feeling was right or wrong. The motivation was to verify that projects’ results come up to the positive impression of cost efficiency.
The triangle benchmarking service conducted by 4SUM Partners showed that indeed both project deliveries were exceptionally efficient and fast, compared to normal productivity of public sector both in Finland and internationally. Functional sizes of the software delivered in these two projects were typical, close to the average FP size in the benchmark dataset, but both the speed of delivery (FP/month) and the unit price (€/FP) were far better than the benchmark. The project management team first believed the high performance was due to agile approach alone, but that was not the whole truth. Several other projects in the benchmarking repositories had applied almost the same set of methods and tools, without reaching the same level of performance.
A closer exploration of the projects and the system documents provided revealed that more than just the agile methods, the key success factor was the product owner’s active leading, and capability to elaborate and represent functional and quality requirements. In this case the developers found it easy to build the expected functionality while in the beginning of iterations, use cases were described on an effective level of accuracy and also the descriptions of business processes and conceptual entities were available for making the big picture.
It is encouraging to notice that it is possible to reach this kind of fine results within our industry. Success is even probable, and should be expected when the client and software vendor have the right capabilities and skills, and they both use those for mutual benefit.
‘Function point counting’, i.e. the process of measuring functional size of software, has claimed to be difficult and time consuming. Several experiments, including the ones organized by FiSMA and introduced below, have proven that the claim is not true, if the measurer knows his method, has fair tools and well-documented information available. Regrettably often system documentation (Product Requirements Document) is fuzzy and incomplete so that finding the functionality there is almost impossible, or at least extremely laborious. In the cases below we had organized the experiment so that the documentation was fairly well done and made it easy to find all the software functions required for counting. This was the way to recognize the true speed of size measurement process, without any disturbance from external sources, like exploring through poor documentation or explaining the details of measurement method to software developers and users.
23 Finnish software experts were invited to FiSMA office in Espoo in November 2015. The purpose of the meeting was to try, how fast they can measure functional size of software, if the circumstances are as good as they can. They were given the information of five systems or pieces of software in terms of numbers of different kind of functional components. For comparison, a similar type of experiment was arranged also 18 months earlier for a larger group of members of the FiSMA Scope Manager Forum. In that case they had measured software size using a professional tool, Experience® Service. The old “world record” from that case was measuring five applications with average speed of 2.7 minutes per software. The average measurement error of 60 measurers was only 1 % compared to the correct result. There they also completed effort estimation in a few extra minutes, still with astonishing success rate. This experiment has documented and published in a paper of MetriKon conference.
In the new ‘world record trial’ our experts concentrated only in software size measurement, using the freeware FiSMA function point calculator instead of professional Experience® Service. They measured the same applications in half a minute per application. Though the tool is simplified, the measurement errors didn’t exceed six percentages. Counting function points is really fast, if you have the information available, system architecture and amount of functional components of the system. In a well-managed software project all this information should be available with ease.
Validating the Design Theory for Managing Project Scope during Software Sourcing and Delivery
Most software project estimation and measurement (PEM) processes have been designed for providers. Customers need to leverage PEM to better direct software sourcing. A design theory for software project-scoping has been developed that supports the PEM processes of both customers and providers. This paper probes the validity of the theory from providers’ and customers’ viewpoints through three longitudinal case studies. A software provider used a preliminary version of the theory in dozens of software development projects annually, systematically (1) achieving higher customer satisfaction through better estimation accuracy and (2) improving productivity. Two government organizations used the theory to speed up their software sourcing and reduce their sourcing costs to a half without giving up any functional or quality requirements. Similar results have not been obtained before. Future research is needed to generalize the findings.
4SUM Partners is hosting a 5 day NorthernSCOPE training event in October 2016. The event will start on Monday 24th and will last till Friday, the 28th. In this training the participants will get all necessary skills a Certified Scope Manager needs, and a certification exam will be held on Friday.
The International Software Benchmarking Standards Group has now released R13 of its New Developments and Enhancements (D&E) repository, containing the project data of 6,760 completed software realization projects. A broad range of project types from many industries and many business areas are available for you to use for estimating, awareness of trends, comparison of platforms and languages or benchmarking. Compared to Release 12 (2012) a total number of 754 projects have been added, 660 of these which were completed in the year 2012 or later. The data in the projects have come from over thirty countries, with 53% of the projects being completed in the last decade.
This repository will be added to Experience® Service later. Keep on reading our news!
(Pekka Forselius and Erkki Savioja, FiSMA, Finland)
Functional size of software is one of the most important cost drivers of software development projects. However, function points are very rarely used in the industry, although they are the only internationally standardized way to measure size of software. Sometimes decision makers have heard about function points, but their knowledge of them is old, or they have been misguided to believe that the methods are difficult to use, extremely time consuming, and that the counting results may vary significantly between two measurement experts.
Finnish Software Measurement Association (FiSMA) arranged a speed test for software size estimation and project effort estimation in summer 2013. The results are impressive and very promising.
The world number one tool for estimating and measuring IT software projects is waiting for YOU in the cloud. This is an acquisition that you will never regret. Knowledge data base with more than 5000 data points, 600+ java projects, 600+ government projects, etc.
We are happy to announce that Experience® Service is now published. This web-based software service is for estimating and measuring IT software projects using function points as a base of measurement. The service application is a Project Estimation and Measurement System (PEMS). With it, the users are capable to measure the functional size of a software application, estimate the effort and cost of the project and collect the project data into the Experience® database.
Order it here. Note, that credit card payments are not yet possible. We are working on it.
Controlling software development – A breakthrough for business.
An article in Software Quality & Process Improvement by mr. Terry Wright (2000).
Businesses must be able to change the way they operate quickly – and often they need to change their software to do so. With this in mind, over the last four years the Victorian Government has developed an invaluable new method for acquiring custom-built software, reports Terry Wright (pdf)