February 28, 2013

From rules to principles : COSMIC


Up to 1998 all methods to express functional size of software were rule-based. Performing the measurement rules led to a number of points that we call the size. Whether they were object points, use case points or function points, all these methods have a measurement procedure to award a part of the Fuctional User Requirements that satisfies a number of assessment criteria, with a defined number of points. To determine the number of points you must apply the rules. A group of people who were involved in the conception of ISO/IEC 14143 wanted to use the principles they had described to create a new generation of Functional Size Measurement method with a clear and defined measurement unit. Based on that unit a method could be created, based on principles to identify instances of that unit, rather than on rules. Out of that process, COSMIC was born.
Here is a piece of COSMIC history.

London, December 1998

The group of people that had worked very long to create ISO/IEC 14143 were officially known as ISO/IEC JTC1 SC7 WG12, which stands for Working Group 12 (Functional Size Measurement) of Sub Committee 7 (Software and Systems Engineering) of the Joint ISO/IEC Technical Committee (Information Technology). At that meeting Alain Abran, Charles Symons, Grant Rule, Jolijn Onvlee, Pam Morris, Peter Fagg, Serge Oligny and Tony Rollo decided to start a project to define a new FSM method, based on the ISO/IEC 14143 principles. The project organisation was named COSMIC, the Common Software Measurement International Consortium.
 

Back to Canada, 1997

A research group from the École de Technologie Supérieure of the Université de Québec in Canada had designed a method with four 'control functions' (later known as data movements) to describe the functional size of software. Although the four 'control functions' were a superset of the IFPUG functions, the design was presented to IFPUG as an add-on subset that could be added to the existing set of IFPUG functions. The whole combination was referred to as Full Function Points.
The new design was not accepted by IFPUG and the research group decided to drop the IFPUG part and proceed with the new design only. It was now called FFP version 2.0 and was tested in industry.
 

Email, 1999

In London, Alain Abran and Charles Symons had been elected joint project leaders of COSMIC. FFP version 2.0 was the basis of the new method. During the year a lot of email communication took place around the globe to discuss the principles of the method. The description of the method evolved and the results of the industry tests came back positive.
 

Again Canada, september 1999

September 8-10 the IWSM conference was held in Montréal. The second day 9-9-99 was almost totally filled with discussions about COSMIC-FFP and the implications this new method for Functional Size Measurement would have. To the conference a fourth workshop day was added in the ski resort of Mont Tremblant north of Montréal. On this fourth day  Alain Abran, Charles Symons, Denis Saint-Pierre, Jean-Marc Desharnais, Peter Fagg and Serge Oligny came together to make the principles of the COSMIC method final.  The first official version of the method, ‘COSMIC-FFP v2.0,’ was published in October 1999.
 

Field trials, 2000-2001, brought back to London

During this period, extensive and successful field trials were carried out to put the workings of the new method to the test. The results were presented in april 2001 on the ESCOM conference in London. You can download the paper from the COSMIC website.
 

ISO-proof, January 2003

In January of 2003 version 2.2 of the COSMIC-FFP method was published to make it fully consistent with becoming an international standard, known as ISO/IEC 19761.
 

Drop the FFP, simply COSMIC, September 2007

In September 2007, the name of the method was simplified to the COSMIC method, in version 3.0 of the method. The most important changes were the dropping of FFP from the method name and the change of the name of the measurement unit from Cfsu (COSMIC functional sizing unit) to CFP (COSMIC Function Point). Especially for people with a Roman language as first language Cfsu happened to be a real tongue-twister, so a lot of users started to call it CFP informally already.
 

Acknowledgement

This piece of history could not have been written without the memory and archives of Alain Abran and Charles Symons. I have been working with the COSMIC method since late 2002 and became interested in the piece of COSMIC history that I had not been part of. I think it is worth to be recollected, since this is really a world-wide initiative to create something new.

No comments: