Meeting Pearls 1/Meeting Pearls Vol 1 (1994).iso/installed_progs/text/faqs/software-eng.part1

Internet Message Format | 1994-04-17 | 12.9 KB

Subject: FAQ 1: comp.software-eng questions and answers
Newsgroups: comp.software-eng,comp.answers,news.answers
From: [e-mail] (David Alex Lamb)
Date: Sat, 16 Apr 1994 12:05:14 GMT

Last-Modified: 30 Nov 1993
Archive-name: software-eng/part1

This message gives brief answers to questions that have occurred in
comp.software-eng; in many cases they are also topics many readers would like
NOT to see discussed again soon.  Questions are:
   What's a CASE Tool?
   What's a 'function point'?
   What's the 'spiral model'?
   What is a 'specmark'?
   Where can I find a public-domain tool to compute metrics?
   How do I write good C style?
   What is 'Hungarian Notation'?
   Are lines-of-code (LOC) a useful productivity measure?
   Should software professionals be licenced/certified?
   How do I get in touch with the SEI?
   What is the SEI maturity model?
   Where can I get information on API?
   What's a 'bug'?
   Where can I get copies of standards??



------------------------------------------------------------------------
Subject: What's a CASE Tool?
Date: 26 Mar 1993


(see also the archive file "casemsg") (thanks to Scott McGregor
<[e-mail]> for inspiring this question)
CASE stands for Computer Aided Software Engineering;  it can be used to mean
any computer-based tool for software planning, development, and evolution.
Various people regularly call the following 'CASE': Structured Analysis (SA),
Structured Design (SD), Editors, Compilers, Debuggers, Edit-Compile-Debug
environments, Code Generators, Documentation Generators, Configuration
Management, Release Management, Project Management, Scheduling, Tracking,
Requirements Tracing, Change Management (CM), Defect Tracking, Structured
Discourse, Documentation editing, Collaboration tools, Access Control,
Integrated Project Support Environments (IPSEs), Intertool message systems,
Reverse Engineering, Metric Analyzers.


------------------------------------------------------------------------
Subject: What's a 'function point'?
Date: 31 Jul 1993


(see also the archive file "funcpoints")

Function points and feature points are methods of estimating the "amount of
functionality" required for a program, and are thus used to estimate project
completion time.  The basic idea involves counting inputs, outputs, and other
features of a description of functionality.  If interested, for a fee you can
join:
    International Function Point Users Group
    5008-28 Pine Creek Drive
    Blendonview Office Park
    Westerville, Ohio  43081-4899


------------------------------------------------------------------------
Subject: What's the 'spiral model'?
Date: 19 Sep 1991


(see also the archive file "spiral")
(1)  Barry Boehm, "A Spiral Model of Software Development and Enhancement",
     ACM SIGSOFT Software Engineering Notes, August 1986.
(2)  Barry Boehm "A Spiral Model of Software Development and Enhancement" IEEE
     Computer, vol.21, #5, May 1988, pp 61-72.

Basically, the idea is incremental development, using the waterfall model for
each step; it's intended to help manage risks.  Don't define in detail the
entire system at first.  The developers should only define the highest
priority features. Define and implement those.  With this knowledge, they
should then go back to define and implement more features in smaller chunks.


------------------------------------------------------------------------
Subject: What is a 'specmark'?
Date: 19 Sep 1991


(see also the archive file "specmark") The SPECmark is the geometric mean of a
series of benchmarks done by the SPEC group. There are a couple of suites, but
in general SPECmark refers to the results of the first suite.  The suite
includes FORTRAN and C codes, mostly well known codes but slightly hacked
versions.
    SPEC
    c/o NCGA
    2722 Merrilee Drive, Suite 200
    Fairfax, VA 22031
    Phone: [phone]
    FAX:   [phone]


------------------------------------------------------------------------
Subject: Where can I find a public-domain tool to compute metrics?
Date: 17 Jan 1992


(see also the archive file "static") Volume 20 of newsgroup comp.sources.unix
contained a public-domain package called "metrics", which computes McCabe and
Halstead metrics.  There are many comp.sources.unix archives around the net.


------------------------------------------------------------------------
Subject: How do I write good C style?
Date: 19 Sep 1991


This is answered regularly in the comp.lang.c FAQ.  Try "Recommended C style
and Coding Standards", on host archive.cis.ohio-state.edu (128.146.8.52) via
anonymous ftp in directory pub/style-guide.


------------------------------------------------------------------------
Subject: What is 'Hungarian Notation'?
Date: 19 Sep 1991


(see also the archive file "hungarian") A naming convention for C code.  See
Charles Simonyi and Martin Heller, "The Hungarian Revolution", BYTE, Aug. 1991
(vol. 16, no. 8).  There are other naming conventions;  see, e.g.  "A Guide to
Natural Naming", Daniel Keller, ETH, Projekt-Zentrum IDA, CH-8092 Zurich,
Switzerland. Published in SIGPLAN Notices, Vol. 25, No. 5, pages 95-102.


------------------------------------------------------------------------
Subject: Are lines-of-code (LOC) a useful productivity measure?
Date: 26 Jan 1993


(see also the archive file "static") Not unless you are very careful.  Capers
Jones' book has a detailed and insightful discussion of Lines of Code,
including anomalies, and shows how to use it sensibly (eg in a single job
shop, with a single language, and a standard company coding style).  It is
easy to cook up anomalies where LOC gives different numbers for code written
in different styles, but pathological cases should get caught in code
inspections.  References:
-    T. Capers Jones, Programming Productivity, McGraw-Hill, New York, 1986
-    Capers Jones, Applied Software Measurement: Assuring Productivity and
     Quality, McGraw-Hill, Inc., 1991, 494 pages ISBN 0-07-032813-7

The appendices of the latter give rules for counting procedural source code,
as well as rules for counting function points and feature points.  The
following study, cited in Boehm's _S_o_f_t_w_a_r_e _E_n_g_i_n_e_e_r_i_n_g _E_c_o_n_o_m_i_c_s, claims that
anomalies that seriously "fool" the LOC metric show up rarely in real code.
-    R. Nelson _S_o_f_t_w_a_r_e _D_a_t_e _C_o_l_l_e_c_t_i_o_n _a_n_d _A_n_a_l_y_s_i_s _a_t _R_A_D_C, Rome Air
     Development Center, Rome, NY.  1978.


------------------------------------------------------------------------
Subject: Should software professionals be licenced/certified?
Date: 19 Sep 1991


This is a very controversial and political question.  Generally, certification
is something voluntary, while licencing is regulated by governments.
Certification generally means some agency warrants you meet its standards;
licencing generally means that to claim to practice a certain profession
requires a government licence, often administered through a professional
organization.  In theory both are supposed to help judge if someone is capable
of doing certain jobs.

Licencing isn't currently required for computing professionals;  some people
would like to see some jobs require it, as with established branches of
engineering.  Others don't like government intervention, and/or believe many
people who wouldn't get licenced are perfectly competent.

Computing professionals in the USA have had a certification program for years,
administered by the Institute for Certification of Computer Professionals
(708-299-4227), a meta-organization with representatives from ACM, IEEE-CS,
ADAPSO, ICCA, IACE, AIM, DPMA, AISP, COMMON, ASM, CIPS, and AWC.  There are
three certificates aimed at different broad types of practitioner, and many
areas of specialization.  To keep a certificate requires at least 40 hours of
continuing education each year; credit can also be obtained for self-study,
teaching, publication, etc.


------------------------------------------------------------------------
Subject: How do I get in touch with the SEI?
Date: 29 Nov 1993


For general information about the SEI, contact the customer relations
department of the Software Engineering Institute at:
    internet:  [e-mail]
    Phone:  [phone]
A subscriber service is available to U.S. mailing addresses. Subscribers
receive the SEI quarterly newsletter, Bridge; invitations to SEI public
events; and first notification of course offerings and new publications.  To
become a subscriber, contact Customer Relations.

To order an SEI publication, contact NTIS, DTIC, or RAI directly:
    National Technical Information Service (NTIS)
    U.S. Department of Commerce
    Springfield, VA 22161-2103
    Telephone: [phone]

    Defense Technical Information Center (DTIC)
    ATTN: FDRA Cameron Station
    Alexandria, VA 22304-6145
    Telephone: [phone]

    Research Access Inc. (RAI)
    3400 Forbes Avenue
    Suite 302
    Pittsburgh, PA 15213
    Telephone: [phone]
    FAX: [phone]


------------------------------------------------------------------------
Subject: What is the SEI maturity model?
Date: 31 Jan 1992
Originally-From: [e-mail] (Mark Paulk)


(see also the archive file "maturity")

Maturity is not an easy concept to get down to a single paragraph, but
consider this.

Premise:  The quality of a software system is largely governed by the quality
of the process used to develop and maintain the software.  Basics:  The first
step in improving the existing situation is to get management buy-in and
management action to clean up the software management processes (walk the
talk, as TQMers frequently say).  Integration:  The second step is to get
everyone working together as a team.  Measurement:  The third step is to
establish objective ways of understanding status and predict where things are
going in your process.  Continuous improvement:  Understand that this is
building a foundation for continually getting better.


------------------------------------------------------------------------
Subject: Where can I get information on API?
Date: 14 Feb 1992



API stands for Application Programming Interface.  For general information on
API standards, you can look at NIST Spec. Pub.  500-187, "Application
Portability Profile." (contact Barbara Blickenstaff, 301-975-2816). Many of
the open systems APIs are being developed in the IEEE POSIX groups.  An
article in the Dec. 1991 IEEE Spectrum describes these and related API
standards. IEEE standards aren't distributed electronically, but both of the
documents above tell how to obtain copies.


------------------------------------------------------------------------
Subject: What's a 'bug'?
Date: 12 May 1992



You can take your pick:
(1)  Don't use "bug", use "fault" (an incorrect instruction or definition),
     "failure" (an incorrect result), or "mistake" (a human action leading to
     a failure).  Paraphrased from
         IEEE Standard Computer Dictionary
         Standard 610, ISBN 1-55937-079-3
         Institute of Electrical and Electronic Engineers, Inc.
         345 East 47th Street
         New York, NY 10017-2394  USA
         $49.50 (US$) for IEEE members
(2)  Beizer, in a footnote on page 33 of the second edition of _S_o_f_t_w_a_r_e
     _T_e_s_t_i_n_g _T_e_c_h_n_i_q_u_e_s says (paraphrased):
         I'm sticking with "bug" because everyone  knows  what  it  means,
         there  are  several  "standards"  for other terms that are incon-
         sistent with each other, the OED says that the conventional  com-
         puter  meaning  of  "bug" is ancient, and short Anglo-Saxon words
         are preferable to long Norman ones.


------------------------------------------------------------------------
Subject: Where can I get copies of standards??
Date: 26 Jan 1993



ISO, ANSI, and IEEE standards are usually sold to raise some of the funds
that the various national and international standards bodies (who usually
own the copyright) need to keep afloat; thus they are not normally avail-
able electronically.  Also, the organizations are concerned that electron-
ic copies would make it too easy for people to disseminate doctored ver-
sions of the standards.

ISO standards may be purchased from:
In Canada:
    Standards Council of Canada / Conseil canadien des normes
    1200-45 O'Connor,
    Ottawa K1P 6N7
    Phone: [phone]
    Fax:   [phone]
On CD-ROM:
    Omnicom, Inc.
    115 Park St. SE
    Vienna, VA 22180-4607
    1-800-OMNICOM
    Also available through the National Technical Information Service
    (NTIS), 5284 Port Royal Rd., Springfield, VA 22161, (703)
    487-4650.

There were once some CCITT standards on-line at the University of Colorado,
but the arrangement to make them available via the Internet was terminated at
the end of 1991.
--
Software Technology Laboratory      [e-mail] (David Alex Lamb)
Computing and Information Science   phone: [phone]
Queen's University, Kingston, Ontario, Canada K7L 3N6