home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dove!swe.ncsl.nist.gov!kuhn
- From: kuhn@swe.ncsl.nist.gov (Rick Kuhn)
- Newsgroups: comp.software-eng
- Subject: Re: Putting Engineering into Software Engineering
- Message-ID: <7679@dove.nist.gov>
- Date: 24 Dec 92 14:57:13 GMT
- References: <1992Dec20.234713.10503@latcs1.lat.oz.au>
- Sender: news@dove.nist.gov
- Lines: 50
- X-Newsreader: TIN [version 1.1 PL6]
-
- Jason Baragry (baragry@amdahl1.lat.oz.au) wrote:
-
- ..
-
- :
- : I think the difference is that in other engineering disciplines these
- : modelling / analysis techniques provide a way to analyse the properties of the
- : components which constitute the system. In these engineering fields there is
- : a high level of reuse and to compare components you need some way of
- : representing the properties of these components.
-
- Exactly, that's why reusable software components will require some type of
- formal specifications for their properties. The current approach of
- relying on English descriptions of components makes it hard to say how
- well a component matches a requirement specification.
-
- : Sure, algorithms in SE can be measured by speed but if I am looking
- : at which electronic components I would want to reuse to meet my needs I can
- : use circuit analysis, frequency & phase analysis, transmission line theory,
- : etc to do it. How can I analytically compare to high-level SE components to
- : see which one meets my needs the best? (assumming people may one day actually
-
- I've been interested in this problem for some time. I have a paper in the
- Dec. 92 Computer Journal (BCS) that has a start to answering this
- question. It describes a technique for analyzing the effects of changes
- in formal specifications, which can also be used to determine which of two
- specifications more closely matches a requirement.
-
- : think like this!) I have to rely on written documentation of what the thing
- : does and how the thing does it. Perhaps there isn't any way of modelling high
- : level SE components. Perhaps the only way to compare SE modules without
- : simply relying on documentation (which is very quantifiable) is to animate
- : those components and compare their operating performance.
- :
- : Jason.
- :
- : -------------------------------------------------------------------------------
- : Jason Baragry. | Amdahl Australian Intelligent
- : Dept Comp. Sci. & Comp. Eng., | Tools Program
- : La Trobe University., | baragry@latcs1.lat.oz.au
- : Bundoora. 3083. | Phone: +61 3 479 1477
- : Australia. | Fax: +61 3 470 4915
- : --------------------------------------------------------------------------------
-
- --
-
- Rick Kuhn Telephone: +1 301 975 3337
- Natl Institute of Standards & Technology Fax: +1 301 590 0932
- Technology Bldg. B266 Internet: kuhn@nist.gov
- Gaithersburg, Md. 20899 USA DRKuhn@dockmaster.ncsc.mil
-