[comp.software-eng] Software Engineering Digest v5n45

soft-eng@MITRE.ARPA (Alok Nigam) (11/30/88)

       Software Engineering Digest          Tuesday, 29 Nov 1988

                          Volume 5 : Issue 45

                            Today's Topics:

                               Quotes (2)
                             Re: ethics (2)
                            SE Digest V4 #8
                     Off-Shore Software Development
         cont. discussion re: uniformity legislation (SUMMARY)
          Are methods and CASE-tools REALLY used in projects?
                     Software Engineering in Prolog
                   PROLOG as a Specification Language
                      Workshop on Software Metrics
                Requirements for SEE Information Systems
- ------------------------------------------------------------

Date: 17 Nov 88 23:10:25 GMT
From: netsys!lamc!well!shf@ames.arpa  (Stuart H. Ferguson)
Organization: The Blue Planet
Subject: Quotes

Sorry if this is not appropriate for this group, but I'm looking for a
reference to a couple of quotes that (I think) I heard sometime in this
newsgroup.  They related to tools and went something like:

        If the only tool you have is an axe, every problem looks like
        a tree.

        If I had six hours to chop down a tree, I'd spend four hours
        sharpening my axe.

Does anyone know the exact quotes and who said them?  Thanx in advance.

- ------------------------------

Date: 22 Nov 88 14:53:47 GMT
From: mailrus!uflorida!haven!decuac!hadron!jsdy@ohio-state.arpa  (Joseph S. D. Yao)
Organization: Hadron, Inc., Fairfax, VA
Subject: Re: Quotes

>       If the only tool you have is an axe, every problem looks like
>       a tree.

This is a re-phrase of an old proverb out of "folk wisdom".  I've
heard it more often phrased with a hammer and nail, though.

>       If I had six hours to chop down a tree, I'd spend four hours
>       sharpening my axe.

I hadn't heard this one before, but I'm glad I have now!

- ------------------------------

Date: 17 Nov 88 13:20:18 GMT
From: vsi!friedl@uunet.uu.net  (Stephen J. Friedl)
Organization: V-Systems, Inc. -- Santa Ana, CA
Subject: Re: ethics

< For those of you faced with this problem I offer the method.
< I call it `Estimating by Threes'.

Alternate mechanism:

* take your best guess as to how long the project will take,
  double it, then go up to the next unit of time :-).

- ------------------------------

Date: 22 Nov 88 16:29:54 GMT
From: mcvax!cernvax!cgch!cgchd6!whbr@uunet.uu.net  (Hellmuth Broda)
Organization: c/o CIBA-GEIGY Ltd, Sci.Comp.Centre, CH-4002 Basel, Switzerland
Subject: Re: ethics

>* take your best guess as to how long the project will take,
>  double it, then go up to the next unit of time :-).

>From /usr/lib/cookies:

        Westheimer's Rule
                To estimate the time to do a task: estimate
                the time it should take, multiply by 2, and
                change  the unit  of measure  to  the  next
                highest unit. Thus we allocate 2 days for a
                one hour task.

This wording sounds so scientific that you can put it into your design
documentation as a motto on page 3 ;-

- ------------------------------

Date: 22 Nov 88 10:20:00 PDT
From: "FRYER RICHARD E" <jiawg@nwc.arpa>
Subject: SE Digest V4 #8

You published a message from ddrg@skl-crc.arpa (Duncan Glendinning)
regarding cost modeling for maintenance.  I've been unable to locate
a current address for Mr. Glendinning.  Can you supply one, or was
there any dialog or follow-up references?

Worried about maintenance costs!

- ------------------------------

Date: 23 Nov 88 12:34:36 GMT
From: att!cbnews!trb@bloom-beacon.mit.edu  (Tom Balent)
Organization: AT&T Bell Laboratories, Columbus
Subject: Off-Shore Software Development

I read this in att.compete...

>    Monday, November 21, 1988
>
>    SINGAPORE -- The Singapore government and US telecommunications
>    giant, AT&T will jointly set up an information communication
>    institute. ... To be known as the Information Communication
>    Institute of Singapore (ICIS), it will train Singaporeans in
>    telecommunications software at the postgraduate level. ... The
>    initial period of cooperation between Singapore and AT&T will be
>    from 1989 to 1995 when the ICIS will be full operational. ... IT
>    Asia, p. 1, Nov.
>

Just last week at the 2nd Annual Columbus Operations Systems Symposium
I was told by Steve Chappell (Director, Software Technology Laboratory
AT&T Bell Laboratories, Liberty Corner) that AT&T is "testing" the
waters of doing off-shore software development. They have decided to use
Indians (living in India) to provide software maintenance for the
project in question. The idea being (at least this is what I got) if
all goes well they will farm out development to India also. Why?
Because studies have shown that code developed in India costs 30% of
what it costs here in the USofA.

He also mentioned that "everybody is doing it", and rattled off a long
list of corporations who use off-shore software developers.

Then along comes this news item. Sounds like AT&T is seeding its next
generation of software developers. What do you people think?  Are you
the least bit concerned or worried? This is not, as mentioned, only an
AT&T problem.

I like a comment I heard a while ago (on t.v. I think) concerning the
jobs of the future in the US, I forgot who said it, but it went like
this:
        "Soon the only job that will be available will be as a
         MacDonald's worker or a lawyer."

- ------------------------------

Date: 23 Nov 88 18:05:40 GMT
From: dulcimer.cis.ohio-state.edu!cso@ohio-state.arpa  (Conleth OConnell)
Organization: The Ohio State University Dept of Computer and Information Science
Subject: cont. discussion re: uniformity legislation

I received the following articles and thought that they gave both
sides good arguments (for and against).  Also, thanks to all who have
offered opinions, I have forwarded all of the comments.

Con O'Connell
- - ------------- Messages appear below ------------------

To: cso@dulcimer.cis.ohio-state.edu
Subject: Re: Software Uniformity Legislation
Newsgroups: comp.software-eng
In-Reply-To: <27462@tut.cis.ohio-state.edu>
Organization: The Ohio State University Dept of Computer and Information Science
Cc:
Bcc:

>From gnu@toad.com Tue Nov 22 02:43:46 1988
Date: Mon, 21 Nov 88 22:00:05 PST
From: gnu@toad.com (John Gilmore)
To: cso@tut.cis.ohio-state.edu (Conleth S. O'Connell)
Cc: rms@toad.com, tower@prep.ai.mit.edu
Subject: Re: uniformity legislation for software
In-Reply-To: RISKS DIGEST 7.79

> A government committee is now considering if uniformity legislation
> for software is necessary, warranted, or desirable.  For example,
> should software suppliers be required to warranty their products?
> should suppliers be required to inform users of known bugs?  should
> bug-fixes be distributed at cost? who should be responsible for
> viruses in object code? etc.

I think that no further legislation should be passed regarding
commercial computer software.  I am amazed at the things that these
legislators sit around and dream up.  Why don't they require warrantees
and bug fixes for slide rules, assign responsibility for damages
resulting from misuse or accident, etc?  If a bridge blows down, should
the engineer's slide rule manufacturer who didn't have a completely
>From gnu@toad.com Tue Nov 22 02:43:46 1988
Date: Mon, 21 Nov 88 22:00:05 PST
From: gnu@toad.com (John Gilmore)
To: cso@tut.cis.ohio-state.edu (Conleth S. O'Connell)
Cc: rms@toad.com, tower@prep.ai.mit.edu
Subject: Re: uniformity legislation for software
In-Reply-To: RISKS DIGEST 7.79

> A government committee is now considering if uniformity legislation
> for software is necessary, warranted, or desirable.  For example,
> should software suppliers be required to warranty their products?
> should suppliers be required to inform users of known bugs?  should
> bug-fixes be distributed at cost? who should be responsible for
> viruses in object code? etc.

I think that no further legislation should be passed regarding
commercial computer software.  I am amazed at the things that these
legislators sit around and dream up.  Why don't they require warrantees
and bug fixes for slide rules, assign responsibility for damages
resulting from misuse or accident, etc?  If a bridge blows down, should
the engineer's slide rule manufacturer who didn't have a completely
transparent cursor be responsible?  (Note: a "program proving" company
in England, Praxis, has been pushing the government there to make
software authors criminally and financially responsible for harm caused
when users run their programs.  This is a great way to make people stop
programming.  If they pass a law like that, *I* will never export a
program to England!)

The computer industry is where it is today, because it moved faster
than the government.  Laws could not be created to regulate and stifle
it, as quickly as they would be obsoleted by new technologies.
And computer innovators were free to innovate because they did not
have a government albatross hung around their necks (or rather, their
albatross was the same size as other businesses').  But computers and
software are made in huge volumes today, and this requires more careful
design and planning -- basically, more time (measured in years) and money.

Now the government sees that computers are a big deal, and they want to
get in on the act.  Japan is winning market share, so they pass laws
about it.  (These laws cripple the U.S. computer manufacturers, and
multiply the Japanese profits.)  Someone reads someone else's
electronic mail, so they pass laws about it.  (The laws have riders
that overturn the fundamental freedom of radio reception which has
existed since radio was discovered; make it easier to get wiretaps and
to plant "bugs" on people; and allow the government to force computer
service bureaus to act against their customers' best interest by
providing secret copies of the customers' files -- and they call this a
"Privacy" act!)  Are you getting my drift here?  The government does
a horrible job when it tries to regulate computers; even when its motives
are reasonable, the result has wide ranging side-effects that make the
total effect worse than leaving things alone.  (I tend to think that this
is true across *all* government regulation -- say, drug laws as another
prime example -- but I know computers and I know best how they screw it up
here.)

"Uniform" rules for software just means "somebody besides the seller
gets to set the rules for their software".  And you can bet that these
rules will not be any good for the small, innovative software creators.
They'll be designed by the lobbyists -- either from the huge computer
companies, or from large user groups like the government or the auto
industry.  These people, it will be claimed, know better than the
seller and their customer.  The customer will be prohibited from buying
a package with different terms, because the seller will be prohibited
from selling such a package.  This is a great way to stifle innovation.

Examples:  if someone sells public domain software, do they have to
warrantee it?  If they sell copyrighted software that they didn't write
and don't maintain, ditto?  If you let those two groups off the hook,
then it becomes a competitive disadvantage to write your own software,
since you have to follow more rules if you do.  If you don't let them
off the hook, then you basically put them out of business, and stop the
distribution of free software.

Of course, you could only make the law cover companies that both sell
and provide maintenance for their products.  Then it would be a
competitive disadvantage to maintain your products yourself -- it would
be better to spin off a subsidiary (at a cost in legal fees, accounting
overhead, and communication between the support & engineering staff)
so that you could offer maintenance without being stuck with the
burden of the law.  The law could make it illegal to maintain software
that you didn't write or don't sell, but that would just put a bunch
more companies out of business and relagate a bunch more software
to the dustbin because it can't be maintained any more.

Of course, the people drafting the laws would not intend any of the above.
They probably don't even know that free software exists (or, the commercial
companies' lobbyists that do know about it are opposed to it).  But
that wouldn't stop them from destroying free software in the process of
trying to put 5-MPH bumpers on Lotus-1-2-3.

Of course, the political process has its own effects.  If a bunch of
software companies originated this effort, they probably have something
in there that legitimizes "shink-wrap" licenses.  Then the big user groups
come in and say "we'll let you have that if we get better warrantees".
This goes on.  Unfortunately, the people who want no regulations at all,
don't have anything to bargain with, so they are left out of the process.

(Free software and software licensing are just examples that I'm
involved in.  If the proposed law takes account of this stuff, it will
just railroad some other part of the industry -- say, network users,
biocomputers, commercial hypertext systems, or publishing networks --
that may not even exist yet.)

If the Congress really wanted the U.S. computer industry to be
competitive, it would stop passing laws shackling the industry.  Go
back to increasing milk prices artificially and paying off farmers to
fallow their land.  We don't need or want government regulation of the
computer industry.

My background:  I've been working in the computer industry since I was
15; I'm now 33.  I have written applications for customers.  I have
written manuals.  I have worked for a time-sharing service bureau.  I
have built system software.  I have worked on computer languages.  I've
worked on IBM and Burroughs mainframes, DEC and DG minicomputers,
assorted micros, and assorted supermicros and workstations.  I helped
to found a startup (Sun Microsystems) and made my fortune thereby.  I
have helped to design hardware and have debugged lots of it.  I helped
to design and debug the Motorola 68020 CPU chip.  I've been a computer
consultant.  I've written free software.  I am currently a partner in a
4-person company selling software that runs on Macintoshes.  I have
never attended college.  I am not a professional actor.

> If you have an opinion on software uniformity legislation, please
> express it publicly, and I will forward your thoughts to one of the
> committee members.  If you feel moved to "second" an opinion already
> expressed, please send me e-mail.

I'm not sure what you mean by "express it publicly".  Feel free to pass
on my opinion to anyone, including mailing lists, newsgroups, or especially
Congresspeople.

- - --------  Now from the other side...----------

>From mergvax!tfli!mikej@philabs.philips.com Wed Nov 23 07:29:25 1988
To: cso@tut.cis.ohio-state.edu
Subject: Call for comments on uniformity legislation for software
Date: 21 Nov 88 15:55:14 EST (Mon)
From: tfli!mikej@mergvax (Michael R.  Johnston)
>From rms@wheaties.ai.mit.edu Tue Nov 22 21:00:50 1988
Date: Tue, 22 Nov 88 21:00:46 EST
From: rms@wheaties.ai.mit.edu (Richard Stallman)
To: cso@tut.cis.ohio-state.edu
In-Reply-To: <8811220600.AA06519@hop.toad.com> (gnu's message of Mon, 21 Nov 88 22:00:05 PST)
Subject: uniformity legislation for software

This is Richard Stallman, founder of the GNU project to give everyone
free software to do all that they want software to do.

I agree with the general sentiment expressed by John Gilmore; however,
the danger that suppliers of free software could be held liable for
bugs in it is not hypothetical.  Right now, we can in principle be
held liable.  Although our public-service nature might shield us by
causing judges or jurors to sympathize with us, this is not
guaranteed, and we cannot do anything to protect ourselves because we
cannot sign contracts with each user.  (Doing so would destroy the
free nature of the software and defeat the purpose of our activities.)

Therefore, I would be happy to see a law passed that would state explicitly
that people who provide free software are not liable for the costs
caused by bugs in the software.

I am one of the principle developers of free software today, and I am also
very interested in political and legal issues that relate to free software
and to society's need for the freedom to share information.
If you find any interest in hearing a non-commercial point of view
on such issues, I could probably fill a few ears.

- - -------

>From mergvax!tfli!mikej@philabs.philips.com Wed Nov 23 07:29:25 1988
To: cso@tut.cis.ohio-state.edu
Subject: Call for comments on uniformity legislation for software
Date: 21 Nov 88 15:55:14 EST (Mon)
From: tfli!mikej@mergvax (Michael R.  Johnston)
>From mergvax!tfli!mikej@philabs.philips.com Wed Nov 23 07:29:25 1988
To: cso@tut.cis.ohio-state.edu
Subject: Call for comments on uniformity legislation for software
Date: 21 Nov 88 15:55:14 EST (Mon)
From: tfli!mikej@mergvax (Michael R.  Johnston)

I am absolutely in favor of this legislation so long as it does not
"go too far". For too long there has been a discrepency in the way
software vendors handle waranties and refunds on their products. Any
legislation that provides for a minimum amount of protection for both
the vendor and the purchaser is in my opinion good for the industry.

Many times I have been apprehensive about purchasing a product because
I had no recourse but to keep it once I open the "shrink wrap". If you
purchase a car do you have to keep it when you find out that it is a 4
speed rather than the five speed you ordered?

Please send any additional information you have on this legislation to
me. Thank you.

Mike Johnston
- - -------

- - ------- End of mail messages --------

- ------------------------------

Date: 18 Nov 88 13:59:25 GMT
From: mcvax!enea!dkuug!imada!thhj@uunet.uu.net  (Hjelm Thomas)
Organization: Dept. of Mathematics and Computer Science, Odense University, Denmark
Subject: Are methods and CASE-tools REALLY used in projects?

In my current work at IFAD (Institute of Applied Computer
Science) I am preparing a danish report comparing software
design methods (which implies the use of CASE). I evaluate
the methods on small examples so I dont know how well they
would do on real projects. However the problems I had on
these small examples could indicate that the methods would be
hard to use on real projects.

Even though CASE tools are spreading at the moment, I believe
the use of methods is very limited, even when tools have
been bought. The amount of discussion of methods/CASE tools
on this newsgroup the time I have been reading it, confirms
this.

If somebody reading this newsgroup have experience using
methods and tools (like Structured Analysis, Jackson
methodologies, Vienna Development Method, PAISley or any other
method/tool), why not share it with all the readers?

I believe the following questions would be of general
interest:

- - - which method?
- - - which tools?
- - - why and how was the method/tool introduced?
- - - what kind of project?
- - - what was the size of the project (sourcelines, persons)?
- - - what is your experience using the method/tool?
     did it make the job easier? did increase the quality of
     your work? did you enjoy using it? will you continue
     using it?

- ------------------------------

Date: 21 Nov 88 20:18:21 GMT
From: ugacs!csun1!weyrich@gatech.edu  (Orville Weyrich)
Organization: University of Georgia, Computer Science Department
Subject: Software Engineering in Prolog

In connection with a large project implemented in Prolog, I am preparing a
survey of software engineering techniques which are applicable to Prolog
programs.

I would greatly appreciate any information, references, or programs dealing
with:

  1) Examining Prolog programs with the intent of finding glitches,
     anomalies, and errors;

  2) Generating test cases to exercise all relevant aspects of a Prolog
     program being tested;

  3) Providing test coverage reports of a Prolog program being tested.

Please help! If you don't, I may be forced to implement the stuff myself :-).

Any suggestions regarding how the above might be done would also be appreciated.

Please respond by mail and I will summarize to the net.  Thanks in advance.

- - --
Orville R. Weyrich, Jr.          | UUCP    : ...gatech!ugacs!csun1!weyrich
Department of Computer Science   |
University of Georgia            |
Athens, GA 30602 USA             | MA BELL : (404) 542-1082

- ------------------------------

Date: 26 Nov 88 17:16:22 GMT
From: oliveb!intelca!mipos3!omepd!psu-cs!warren@ames.arpa  (Warren Harrison)
Organization: Dept. of Computer Science, Portland State University; Portland OR
Subject: PROLOG as a Specification Language

I'm currently looking at using a variant of PROLOG as a formal specification
language.  Could anyone out there supply some pointers to papers which deal
with the subject of PROLOG as a formal speciifcation language?  Many thanks.

- ------------------------------

Date: 23 Nov 88 18:40:27 GMT
From: voder!pyramid!oliveb!intelca!mipos3!omepd!psu-cs!warren@bloom-beacon.mit.edu  (Warren Harrison)
Organization: Dept. of Computer Science, Portland State University; Portland OR
Subject: Workshop on Software Metrics

    The First Annual Oregon Workshop on Software Metrics


The First Annual Oregon Workshop on Software Metrics will be
held in Portland Oregon on March 1 and 2, 1989.  The
workshop will bring together both practioners and
researchers from throughout the country to discuss the
application and use of software metrics.  In order to ensure
close interaction among participants, attendance will be
limited to 75.

The first day of the workshop will consist of a tutorial on
software metrics, followed by presentations from
practitioners currently using metrics.  The second day of
the workshop will include presentations by metrics
researchers and a panel session addressing how well current
metrics serve the needs of the software development
community.

The First Annual Oregon Workshop on Software Metrics is
being jointly sponsored by the Oregon Center for Advanced
Technology Education, Portland State University and Oregon
State University.  For more information contact:




                      Warren Harrison
               Department of Computer Science
                 Portland State University
                         PO Box 751
                  Portland, OR 97207-0751

                       (503) 464-3108

- ------------------------------

Date: 25 Nov 88 21:44:12 GMT
From: mcvax!inria!crin!charoy@uunet.uu.net  (Francois CHAROY)
Organization: C.R.I.N
Subject: Requirements for SEE Information Systems

A lot of paper emphasize that current data models and database systems
doesn't fit with software engineering environments requirements (CAD,
CASE and so on...) but I found few reports on such requirements

Yet, I have only the following Report:
        RAC : CAIS Requirements.
        EURAC : Requirements For Tools Support Interface (PCTE+)

 So, I'm looking for reports, articles, and even personnal opinions
dealing with these requirements.

Please e-mail me the answers, I'll summarize to the NET

- ------------------------------

End of Software Engineering Digest
***************************************