home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.specification
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!ukma!daffodil!news.cs.odu.edu!zeil
- From: zeil@renenutet.cs.odu.edu (Steven J. Zeil)
- Subject: Re: Testing
- In-Reply-To: emcdnnll@cs.tcd.ie's message of Wed, 11 Nov 1992 12: 02:50 GMT
- Message-ID: <ZEIL.92Nov19164524@renenutet.cs.odu.edu>
- Sender: news@cs.odu.edu (News File Owner)
- Nntp-Posting-Host: renenutet.cs.odu.edu
- Organization: Old Dominion University, Norfolk, VA, USA
- References: <1992Nov11.120250.12884@cs.tcd.ie>
- Date: Thu, 19 Nov 1992 21:45:24 GMT
- Lines: 92
-
- In article <1992Nov11.120250.12884@cs.tcd.ie> emcdnnll@cs.tcd.ie (Eoin McDonnell) writes:
-
- Newsgroups: comp.specification
- Path: daffodil!concert!gatech!destroyer!uunet!mcsun!ieunet!tcdcs!emcdnnll
- From: emcdnnll@cs.tcd.ie (Eoin McDonnell)
- Summary: Ref's req'd for use of formal spec's in software testing
- Keywords: Testing,formal specification
- Organization: Computer Science, Trinity College Dublin
- Date: Wed, 11 Nov 1992 12:02:50 GMT
- Lines: 10
-
-
- Hi all,
-
- I'm interested in finding out how formal specifications can be used to
- help design test programmes for software. Any references would be
- appreciated.
-
- Thanks,
-
- Eoin Mc Donnell. emcdnnll@cs.tcd.ie
-
-
-
- Formal specs can come into play in testing in two ways.
-
- First, all testing depends upon the existence of an "oracle" to
- examine test inputs and program outputs and detect the presence of
- failures (incorrect results). This oracle is most often provided via
- human inspection of the input/output data or via textual comparison
- against a file of previously recorded expected results. However, any
- scheme for animating/evaluating specifications can also help to serve
- as an oracle (at the risk of allowing errors in the specification that
- were faithfully copied into the design to go undetected). Such
- automated aids can be of immense use in situations where many
- thousands or tens of thousands of tests are being run.
-
-
-
- Second, the spec can be used as the source of information from which
- test cases are generated or against which the suitability of tests
- sets is measured. Here's a sampler of several very different
- approaches to such specification-based testing:
-
- @article{Bouge/test-gen-logic,
- author = "L. Bouge and N. Choquet and L. Fribourg and M. C. Gaudel",
- journal = "Journal of Systems and Software",
- month = Nov,
- number = "4",
- pages = "343--360",
- title = "Test Sets Generation From Algebraic Specifications Using
- Logic Programming",
- volume = "6",
- year = "1986"
- }
-
- @article{Gannon/daists,
- author = "John Gannon and Paul Mc{M}ullin and Richard Hamlet",
- journal = "ACM Transactions on Programming Languages and Systesm",
- month = Jul,
- number = "3",
- pages = "211--223",
- title = "Data-Abstraction Implementation, Specification, and Testing",
- volume = "3",
- year = "1981"
- }
-
-
- @techreport{Luckham/anna-debugging,
- author = "David C. Luckham and Sriram Sankar and Shuzo Takahashi",
- address = "Stanford CA",
- institution = "Stanford University Computer Systems Laboratory",
- month = Apr,
- number = "CSL--TR--89--379",
- title = "Two Dimensional Pinpointing: An Application of Formal
- Specification to Debugging Packages",
- year = "1989
- "
- }
-
- @article{Wild/refinement-jtvr,
- author = "Christian Wild and Steven Zeil and Ji Chen and Gao Feng",
- journal = "{\rm to appear in} Journal of Software Testing,
- Verification, and Reliability",
- title = "Employing Accumulated Knowledge to Refine Test Cases",
- year = "1992-93"
- note = "also available as Old Dominion University report TR-91-14"
- }
-
- (Well, I couldn't resist putting ONE of my own papers in there.)
-
- Steve Zeil
-