School Work

METKIT: A Metrics Educational Toolkit

Description
METKIT: A Metrics Educational Toolkit
Categories
Published
of 3
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  MARTIN BUSH NICHOLAS ASHLEY Brameur ltd. zyxwvutsrqpo Mekit was one of the SPRIT I1 information-processing system and software proiects. The project, which involved seven companies and one university from five countries, lasted from 1 989 to 1992. It was funded with 4.6 million KUs (about 5.7 million) and involved 51 man-years of effort. ~ 52 Metkit responded to the need for more material with which to teach measurement techniques as they apply to software engineering. The materials the project team produced have greatly aided companies and universities throughout Europe; more than 5,000 industry users have studied a cross-section of Metkit modules. More important, team members have noted a change in attitude toward metrics, as software engineers begin to regard measurement as an integral part of software engineering. zyx or metrics to be adopted F n a broad scale, there must be some way to dissemi- nate knowledge about the sub- ject, both the techques and the concepts underlying it. Yet there is surprisingly ittle avail- able to educate individuals. Metkit was organized to ad- dress lus issue, both for indus- zyxwvu ry and universities. ORGANIZATION The Metkit project had three phases. In the first, we did background studies into what teaching material was thought to be required, what was available but not being used, and what was already being used. In parallel, we de- vised a conceptual fi-amework to underlie the subject of soft- ware engineering as a whole.’ The second phase involved he production of separate proto- type teachmg packages for ac- ademic and industrial audi- ences, and extensive trials using th~s material wihn the eight participating organiza- tions plus two other compa- nies and 1 I other universities. In the third phase, we pro- duced the ha1 versions of the teaching packages. Although not funded as part of the project, we have conducted a series of follow-up evaluations of companies using the teaching Inaterial to iden- tify trends in attitude and use. Information gathering. The background studies in the first phase involved extensive sur- veys of industry and academia, involving the information- technology departments of 400 European organizations and 445 universities worldwide. The results were illuininat- ing. We found that more teaching was tahg place than we had thought, but that topics were normally presented in a rather fragmented way. Only seven percent of the academic respondents felt that sufficient teaching aids were already available to them.* zyxwv   more predicable result was that available techniques weren’t being used. The in- dustry survey revealed that fewer than five percent of the information-technology de- pamnents use measurement to improve their software-devel- opinent process. But many noted that the main reason was they simply didn’t understand how to use measurement ef- fectively. Rdts. The project pro- duced concrete results in the form of teaching modules and a prototype computer-aided instruction system. The proj- ect also partially funded a text based on the teaching package written by Norman Fenton of City University, London, in collaboration with the Metlut team.’ A less tangible but equally important result is the change in attitude we have observed toward software measurement as these materials are being used. This is gratifying be- cause we deliberately inserted the term “software-engineer- ing measurement” in all the Metkt materials in an attempt to encourage software engi- neers to thmk less in terms of the restrictive concept of soft- ware metrics and more in terms of integrated measure- ~ NOVEMBER 1993  Conducting zyxwvut   defect onolyris zyxwvuts Figure zyxwvutsrqpo   zyxwvutsrqponmlk mtmt ojthe Metkrt inrliisn-ial eaibmg paikage. Enib rectangle repl-esezts zyxwvuts   module zyxwvutsr ment, in the broadest sense, wih oftware engineering. TEACHING PACKAGES There are two distinct teachmg packages of modules: one for industrial use and one for academic use. Foundational concepts. The material is based on several un- derlying principles. In keeping with our broad outlook on measurement, the packages cover measurement tech- niques applicable not only to software products but also to development activities and to the resources used in software production ncluding the technology likely to be used and even the developers themselves. The packages also embody the notion ofgoal-driven mea- surement. Tlus notion is based on Vic Basili’s goal-question- metric paradigm,4 which em- phasizes responding to “What do we want to know?” rather than “What can we measure?” Content. Both packages con- tain an overview of software- engineering measurement as well as topic-specific modules. The rainer can select modules to fit a range of courses, de- pending on the time available and the students’ particular in- terests. Indwtrialpackage. Figure  1  shows the content of the in- dustrial package, which is con- cerned primarilywith showing how to use measurement as a management tool and deals with specific areas like cost prediction, defect analysis, process improvement, and quaky specification and mea- surement. In many cases, we have tailored modules to in- dustrial users other than main- stream software engmeers, such as senior and middle man- agers and quality engineers. Academic package. The academic package is somewhat smaller than the industrial package, concentrating more on the theory underlying the subject. It is intended to be slotted into undergraduate and postgraduate courses in coin- puting and related subjects. There is more emphasis on consistency and cross-refer- encing than in the industrial modules. Acceptance. The packages h ave b e en well-received. So far, we have sold them to 45 universities both within and outside Europe and to 28 companies ot count- ing the orga- nizations that were given the material as a zyxw   conduct a survey to confirm zy S. Companies that use the modules have found that they save teachers 80 to 100 percent of the preparation time nor- mally required for courses. For example, preparation for a short course on cost esti- INDUSTRY TRAINERS REPORT THAT THE PACKAGES SAVE THEM AS MUCH AS 10 DAYS OF PREPARATION. mation may normally take two weeks; with the Metlut mate- rials, the time is typically one day or shorter. In all, we know that reward for participating in the trials. We also know that the kits are being used. We have sur- veyed industrial recipients and found that all are using the in- dustrial package to train their staff. We also assume that most academic recipients are using the material, but we plan to ~~ ~ ~~ ~~ ~ more than 5 000 indus- ny-based users have studied a cross-section of Metlut mod- ules. Furthermore, at least 16 companies have also used Metkit to support process- imp rove men initiatives Brameur Ltd. has used the in- dustrial package as a blueprint for setting up and running IEEE SOFTWARE zyxwvutsrqponml    measurement programs for four companies zyxwvutsr n he UK. Another indication of ac- ceptance is that Fenton’s text, which is closely tied to the modules, has had considerable commercial success and has greatly influenced progress in the field. zyxwvutsrq CAI SYSTEM The prototype zyxwvutsrqp AI sys- tem is designed to support software engineers as they work by providing knowl- edge on demand about the concepts of software-engi- neering measurement and their interrelationship^.^ The system integrates hypertext concepts with those of computer-aided instruction to provide different levels of access to the conceptual knowledge it contains. The system is implemented in Smalltalk 80 and runs on Apple Macintosh I1 and Sun Sparcstation workstations under Unix and Open Win- dows. he Metlat project ended T y delivering completed packages ofteahg materials; there were no loose ends. Nev- ertheless, updating and ex- OFFERED BY THE SCHOOL OF ENGINEERING AND APPLIED SCIENCE AT SOUTHERN METHODIST UNIVERSITY TheprogramisofferedviaVIDEOTAPEnation- wide as well as on the SMU campus. The program is centered about the problems facing the working professional in the field. Emphasis is zyxwvutsrqpo n both the fundamental principles of software and system design and the practi- cal problems of commercialization. Much of the subject material of the core is based on the curriculum recommended by the Software En- gineering Institute. The elective courses enable the student to focus on particular interests. ADMISSION REQUIREMENTS: BS/BA in one of the sciences or engineering disciplines with a 3 014 0 GPA and one year experience in software development or maintenance. tending the industrial modules is an ongoing process. We have not yet felt the need to revise and update the academic modules, but it is only a matter of time before we also bep work on these. User feedback has been ex- tremely positive, both from teachers, who remark on the materials’ comprehensiveness, and students, who appreciate the relevance and immediate practical use of the subject matter. We have also received a number of constructive sugges- tions on how to enhance the material; in particular, users seem to want more case studies We believe Metlut will be significant in the growing trend to accept and use mea- surement techniques. General industry awareness has signifi- cantly improved; measure- ment is now being mentioned frequently at conferences and in journals dealing with main- stream software-engineering issues. Software-engineering measurement is now dy e- coming regarded as an insepa- rable part of software engi- neering and software-quality management. Indeed, the distinction between these areas has become quite blurred. The emphasis in the future z  ll increasingly be to set up measurement programs in some form within organiza- tions. It is our hope that Metkit can help the information- technology industry achieve zy his objective. zy EFEREN ES 1. M. Bush and N. Fenton “Software Measurement.” AConceptuaJ Frame- 2. M. Russell “Internanonal Survey of Software Measurement Fducation and zyx 3 N. Fenton S@mre Menim R~om~u.4pproach hapman Hall IWl. 4. V Basili and H. Rombach “The TAME Project: Towards Improvement- Oriented Software Environments ” EEE Eum. Sojwure Eng., June 1988 5 M. Mullerburg and D. Meyerhoff “The METKIT CAI System ”?. Ifor- worV7 System Sofiwuw, Dec. 1990 pp. 223-231. Training ”?. zyxwvut ystenzs Sojwure Dec. 1990, pp. 233-241. pp. 758-773. m tion Software Technohgy No. 2 1993, pp. 120 126. Martin Bush is a senior lecturer in computing at South Bank University and was project leader of Metkit’s South Bank team. Bush received a first-class honors degree in electrical and electronic engineering om London University King’s College. He is a chartered engineer and a member of the IEE and British Computer Society. Nicholas Ashley S a consultant with Brameur Ltd. a consultancy specializing in software-quality ssurance where his i~in oles are to help information-technology depamients set up and run pmcess-improvement program and use measurement as a management tool. University Cmllege of North Wales. Ashley received a PhD in pure mathematics from the Address questions about this summary or requests for Metkit documents to Bush at South Bank University Center for System and Software ling. 103 Borough Rd. London SEI 0 Q England; fax 44) 1 028 1?84. NOVEMBER 1993
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x