The following text is a work is in preparation. For any comments and suggestions please send an e-mail to Carla Freericks (Blanck Purper)
Pursuing after a certification is one of the major motivations of software development organizations to apply a process standard. This fact has emerged to what Curtis /Curtis, 1998/ refers to as a "checklist mentality", when process standards are used as a checklist instead of a guidebook to enhance project results. Organizations with such mentality often apply standards mechanistically and rigidly without continuously integrating, improving and tailoring respective process standards to achieve project objectives.
Rather than hampering progress and consisting in a mere checklist, process standards establish a common framework for process participants to carry out a process. Actually, process standards should be seen as a way for:
However, the advent of the Internet has added a new criterion to process standards. Not only should they be innovative to serve as a mean for continual improvement, but also, should their last release be systematically and immediately transferred to users in order to accompany technology.
- modernizing the software development,
- establishing quality standard level,
- improving efficiency and,
- warning users about the known pitfalls.
As the expansion of the Internet/Intranet has broadened the market of web-based systems (WIS) /Isakowitz, 1998/ the number of them have increased remarkably in the last years. Nowadays, WIS systems have assumed such a prominent role in system development that it changed the development priority from What to When /Aoyama, 1998/. Reduced time-to-market has become the competitive cornerstone of this growing marketplace. This priority requires shorten system lifecycles and flexible process models in order to facilitate the adoption of emerging proved solutions, which, in turn, requires agile and dynamic process standards.
Nevertheless, agility and dynamism of process standards are restrained by:
- the medium used for publishing them. Most commonly by means of static documents, usually paper-based, which are difficult to handle and to learn. Few of them are optionally available as:
- postscript (or PDF, RTF, Word documents),
- help-based documents and
- web-based documents. Usually, web-based documents are automatically converted from word-processor format into HTML documents for direct display via Web-browsers, and the existing hyper-links are limited to references within the same document.
- the writing form for describing the standard. Mähönen /Mähönen, 2000/ argues that part of the claims to the hardness in assimilating standards are due to the difficulty in understanding the text rather than in implementing the real technological aspects.
- distribution channels. Still there are many process standards which cannot be accessed on-line, and its paper-based handbook is only available by the official selling-points. However, the selling-points are commonly not on-line, and might only be accessed personally or by phone. Another incovenience is that they are usually not informed which is the lattest version. This problem is accentuated by the conversion of standards between standard development organizations (SDO). For example, the original ISO 9001-3 is found in Germany as DIN EN ISO 9001 Teil 3, and was delivered some months later. Normally, you are not allowed to order the standard from a different selling-point than the one assigned to the region you live. This means, in a normal situation, you cannot get the standard directly from the source.
- monolithic versions.
- decision making process .
- late releases .
GDPA is an environment to support flexible process standards. Some benefits of this environment are:
- Promotes the dissemination of the standard by enabling a expeditious on-line and off-line access to its target public, available for most of the World-Wide-Web (WWW) browsers,
- Supports process modelers to navigate throughout the standard according to their own strategy of learning by supplying a consistent, extensive interwoven hypertext documents, and
- Facilitates the incorporation of the standard increments into the existing process by providing a flexible structure for process elements (elements such as activities, artifacts, agents, etc.)
Figure Intro 1: GDPA as a process web-center for standards
There are so many process standards, methodologies and initiatives on process technologies (
PSS05, IEEE1220, ISO9001-3, ISO12207, IEEE1074, AQAP150, GD250 (V-Model), BOOTSTRAP, CMM, SPICE, Trillium, IDEAL, QIP) that it is quite difficult to decide which one should be used. There are numerous pertained published studies which provide some guidelines to understand them. For example the Quagmire Frameworks /Sheard, 1998/ presents an interwoven graph providing an overview of almost all software processes. Moore /Moore, 1998/ describes a "User's Road Map" to software engineering standards. Emmerich et al. /Emmerich, 1998/ organise the process standardisation initiatives in three main streams: process technologies, assessment methods and improvement. El Eman /El Eman, 1999/ concentrates on software process assessment and improvement.
However, the above mentioned publications do not cover in extension the facts on process technologies. One can obtain information from different sources such as:
Web pages from the tools: In the process web-centre more than 300 homepages are identified. Although it is still not an exhaustive index, it contains various links to programming languages, workflow, graphical notations, process engines, etc.
Standards on software processes: Whether standards are de facto or imposed, Standard Developing Organisations, also known as SDO (ISO, ANSI, IEC, etc.), consortia (W3C, OMG, IETF, etc) and individual organisations (ESA, NATO, HP, etc.) are launching standards. The key driver of Consortia is "productising" the specification of the standard rather than "producing" it. Consortia principle relies on the concept that the market decides about the dominance or obsolescence of incompatible standards and decides which technology becomes dominant. Despite of the fact that it is quite difficult to determine when a standard candidate becomes a "real" standard, standards endeavours on process technologies are flourishing and the number of them has increased notably in the last few years.
Publications on process technologies: Since the first ISPW (International Software Process Workshop) held in 1984, more than 24 sessions together with ICSP (International Conference on Software Process), EWSPT (European Workshop on Process Technologies), IPTW (International Process Technology Workshop) and the german session organized by GI 5.1.1 have been taken place. Together with the research on process technologies published in other conferences such as ICSE, FSE, ESEC, CAISE, EUROMICRO, we find out more than 3000 publications. Meanwhile circa 1000 publications are already recorded in the process web-centre.
Research projects: Research projects on process technologies conducted by partnerships between academia and industry have demonstrated to be a very effective way for innovating concepts, methodologies and tools which are being applied on the market. Most of these research projects are partially sponsored by national governmental organisms. Some of them are supported by multi-national governmental organisations. For example, the European Commission has launched in the last 5 years more than 80 projects concerning process technologies.
Set-up experiences: Mailinglists, invited speaks at conferences, roundtables and panels are outstanding sources for learning about practices in setting-up process technologies. In the process web-centre, more than 350 e-mails sent to the V-Model [GD250] mailinglist are identified and classified in the experience library meta-model.
Altogether, the above cited sources of information do not consist in a comprehensive list of all possibilities, but they provide a hindsight on most of the trends in process technologies. From these sources, all the data for the information system in the process web-centre is captured and stored in a repository. Most of the objects of the repository is organised into meta-models.