soft-eng@MITRE.ARPA (Alok Nigam) (01/19/89)
Software Engineering Digest Wednesday, 18 Jan 1989
Volume 6 : Issue 2
Today's Topics:
Shareware
Looking for an SCCS front-end.
Question for COCOMO users
software cost estimation utilities
Clarification of an earlier posting
Realtime Software Development Courses
CASE tools question...
Workshop on Software Metrics
Tools for building a system
Re: software cost estimation utilities
A CASE for the Macintosh
Re: Question for COCOMO users
Software engineering for expert systems
------------------------------------------------------------
Date: Thu, 12 Jan 89 11:11:02 CST
From: Don Branson <C08922DB%WUVMD.BITNET@uicvm.uic.edu>
Subject: Shareware
I am interested in finding out some things about users of shareware, and
would like to take a survey. Please mail your answers to directly to me,
and I will summarize to the list. The more people respond, the more meaningful
the result will be. The primary reason for doing this is to find out what
percentage of users of shareware pay for what they use.
If you have any hesitation about answering, let me say that I will not use
these results for prosecution purposes. Here are the questions:
1) How and when did you become aware of shareware?
2) How many shareware packages do you use?
3) How many of these have you registered?
4) How many of these do you intend to register at a later date?
5) What do you see as your reasons for not registering?
Thanks in advance for your answers to this survey.
------------------------------
Date: 11 Jan 89 17:18:19 GMT
From: mcvax!cernvax!hslrswi!francis@uunet.uu.net (Francis Demierre)
Organization: Hasler AG, CH-3000 Berne 14, Switzerland
Subject: Looking for an SCCS front-end.
I am looking for a front end to the SCCS for maintenance of multi-level
projects. (one 'Standard code release level', several 'Project
specific release levels').
If you have any such PD front end written in C or shell script, I would
be very grateful if you could 'uuencode' it and mail it to me.
(or tell me about a site where I could get it from with uucp).
------------------------------
Date: 11 Jan 89 18:25:56 GMT
From: mcvax!ukc!cs.tcd.ie!csvax1!omahony@uunet.uu.net (Donal O'Mahony - OMAHONY@cs.tcd.ie)
Organization: Computer Science Department, Trinity College Dublin
Subject: Question for COCOMO users
I am preparing a short course in Software Cost Estimation in COCOMO,
and have a question that a seasoned user of the method may be able to
answer - I can't find it in Boehm's "Software Engineering Economics"!
::
In the intermediate model- assuming organic mode development
EFFORTnominal = 2.4(KDSI)**1.05
Effort Multipler = (product of 15 cost drivers)
EFFORTactual = (Effort Multipler).EFFORTnominal
: So far, so good - the difficulty come in the scheduling equation
which of the following is TRUE?
A) TDEV = 2.5(EFFORTnominal)**0.38
or
B) TDEV = 2.5(EFFORTactual)**0.38
: If the correct formula is (B), then if the SCED driver is
very low (1.23), one should get a 75% reduction in schedule. Does
this mean 75% or (A) or (B).
I'd appreciate any help with this.
------------------------------
Date: 12 Jan 89 21:47:07 GMT
From: manta!plantz@nosc.mil (Glen W. Plantz)
Organization: Naval Ocean Systems Center, San Diego
Subject: software cost estimation utilities
Does anyone know the names of any companies that produce software
cost estimation utilities that will run on an at-clone? I know that
one company has a product that implements the COCOMO model, but I
don't remember that name. Please send responses to plantz@nosc.mil
------------------------------
Date: 12 Jan 89 15:18:08 GMT
From: steinmetz!farside!mitchell@uunet.uu.net
Subject: Clarification of an earlier posting
Some clarification on the type of transition I am interested in...
>I am interested in any references which deal substantially with the
>event of transitioning software from a development to a production
......
The clarification (hopefully..)
Specifically, the type of transition I am interested in is one where the
software transfer is a one time deal. That is, group A develops the
software for group B. When development is complete, the software is
transitioned completely to group B. Groups A and B may be down the hall
from each other or they may be separated by a great distance. The
important thing is that they belong to the same basic organization
(company, university, etc.).
------------------------------
Date: 9 Jan 89 22:48:05 GMT
From: cadnetix.COM!robin@uunet.uu.net
Organization: Cadnetix Corp., Boulder, CO
Subject: Realtime Software Development Courses
I am posting this for a friend with no network access:
I am interested in any short courses (also good books/articles)
on Realtime Software Development. I am in the Research Triangle
area, but would have no problem flying anywhere in the U.S. for
a short course. Any information would be appreciated.
Please send replies to {uunet, uucp, nbires, boulder}!cadnetix!robin
and I will forward hardcopies.
------------------------------
Date: 9 Jan 89 21:39:46 GMT
From: nic.MR.NET!eta!woods.unix.eta.com!bstrand@csd4.milw.wisc.edu
Organization: ETA Systems, Inc., St Paul, MN
Subject: CASE tools question...
I am posting this request for a colleague...please respond to me, or
post.
- ------------------------------------------------------------------
I am evaluating CASE tools for the development of real-time
systems. So far, I have looked at the following two packages:
Software through Pictures (tm) - Interactive Development Environments (IDE)
Excelerator/RTS (tm) - Index Technology
Does anyone have experience with either of these packages or
any other CASE tools? What were the problems and benefits?
------------------------------
Date: 11 Jan 89 17:00:04 GMT
From: blake!ogccse!littlei!omepd!psu-cs!warren@beaver.cs.washington.edu (Warren Harrison)
Organization: Dept. of Computer Science, Portland State University; Portland OR
Subject: Workshop on Software Metrics
I'm really sorry to be using the net this way, but I have received a large
number of e-mail requests for information about the Workshop on Software
Metrics scheduled for March 1 and 2. Unfortunately, many requests did not
include a USMail address, and I have had trouble getting through to some
of these people. In particular, Cynthia@nyborg needs to resend me her USMail
address and Brian@huron needs to send me his phone number since my mail keeps
getting returned.
If you sent me an e-mail request and you're not sure if you included a USMail
address, please resend to ensure you will be on the mailing list.
- -------------------------------------------------------------------------------
Warren Harrison CSNET: warren@pdx.edu
Department of Computer Science UUCP: {ucbvax,decvax}!tektronix!psu-cs!warren
Portland State University Internet: warren%pdx.edu@relay.cs.net
Portland, OR 97207-0751
------------------------------
Date: 14 Jan 89 19:17:46 GMT
From: mcvax!enea!sommar@uunet.uu.net (Erland Sommarskog)
Organization: ENEA DATA AB, Sweden
Subject: Tools for building a system
When you build your system, or part of it, you have a good help
of a tool like MMS or some other "make" clone. However, MMS,
requires that you know all your dependencies. Since they may
change during development maintaining the MMS files becomes
an important task that is boring and error prone. At the place
I'm currently located we have a tool to automatically generate MMS
files. Unfortunately, it doesn't work very well, since the
analysis is made on the list file, thus the tool must compile
to find out the dependencies. This is a fatal mistake, since
some of the files required for the compilation is maybe out of
date or non-existent, causing the compilation to fail.
What I have been thinking of is to first get dependency information
which concerns compilation from the source file and then iterating
back, until I don't depend on anything, compile in the correct
order, gather all external references(*) from the list file (or the
object file) until all is compiled and I can construct the link
dependencies for my executable(s). The external references must
have been put in a database of some sort, so I can check that the
corresponding modules are up to date.
(*) External reference: A variable/procedure/function that is not
defined in the referencing module or any of the modules it imports
from. E.g. a routine in another langauge.
I have two questions, one of just general curiousity and one to
possibly save some work.
1) Have such tools been built? And for what langauges and OS:s?
I know of something similar for C/Unix but that is a fairly
trivial case, since C only have an include facility. All Ada
system must have something of the like.
2) Is there anything we could use? We have the combination VMS,
VAX-Pascal, Cobol, CDD, RDB and CMS. Any ideas to avoid work
that already has been done are appreciated.
------------------------------
Date: 13 Jan 89 17:24:41 GMT
From: hp-pcd!hpcvlx!boses@hplabs.hp.com (Sharada Bose)
Organization: Hewlett-Packard Co., Corvallis, OR, USA
Subject: Re: software cost estimation utilities
Iconix Software Engineering, Inc. markets CoCoPro that implements
CoCoMo for the Macintosh.
All I know is that the company is located in the Los Angeles
area. If you want more info (address, phone number) let me know. I can
get it for you.
------------------------------
Date: 16 Jan 89 22:28:40 GMT
From: pitstop!sundc!netxcom!dmcintee@sun.com (Dave McIntee)
Organization: NetExpress Communications, Inc., Vienna, VA
Subject: A CASE for the Macintosh
Numerous products have come on the market in recent years
in the category of Computer-Aided Software Engineering (CASE). These
are systems for building systems.
CASE tools aid systems developers in the creation and validation
of various types of diagrams. Techniques include DeMarco/Gane and Sarson
dataflow diagrams, state-transition charts, structure charts, Chen or
Martin entity-relationship diagrams, and others. Some even allow you to
define your own symbols and modeling rules to set up your own diagramming
techniques.
These tools represent, for a Macintosh user, a paradox: They are
all graphics applications - with mouse, icons, windows, menu bars - and
they all run on IBM PCs. (This is a slight exaggeration, as there are a
handful running on VAX/VMS, UNIX and IBM MVS, but the vast majority run
under MS-DOS.) What adds to the puzzlement is that these are CPU- and
memory-intensive applications; some won't boot in 640KB of memory, while
others "will run on a 286 machine, but we URGE you to consider a 386-
based platform". So why isn't the Macintosh at the fore, or moving to be,
in this market?
If you have information or comments, please mail to me; I would
be delighted to be proven incorrect. I'll summarize any responses.
------------------------------
Date: 13 Jan 89 21:01:42 GMT
From: ncar!mailrus!sharkey!oxtrap!metavax!marc@gatech.edu (Marc Paige)
Organization: Meta Systems, Ltd. -- Ann Arbor, MI
Subject: Re: Question for COCOMO users
>A) TDEV = 2.5(EFFORTnominal)**0.38
>or
>B) TDEV = 2.5(EFFORTactual)**0.38
>
I believe tha answer is A. This is from looking back on notes from the
same book (similar course I hope :-) ). The organic model in a few pages
back (pg 85 fig 6-5) shows the formula for MM as EFFORTnominal.
------------------------------
Date: 12 Jan 89 13:25:58 GMT
From: mcvax!unido!ztivax!jack@uunet.uu.net (Jack Schiff)
Organization: Siemens AG in Munich, W-Germany
Subject: Software engineering for expert systems
Hello everybody! I'm currently collecting references on methods and/or
tools which have been developed for the systematic design, testing and
maintenance of expert systems; also lifecycle models for expert systems
etc. (I'm crossposting this to comp.cog-eng because I'm especially
interested in requirements analysis for expert systems.)
Any pointers that you can give me are greatly appreciated; if there are
sufficient responses, I'll summarize and post them to the net. Please
mail to: jack@ztivax.uucp (from UUCP) or jack@ztivax.siemens.com (from
ARPANet), or Jack Schiff c/o Siemens AG, ZFE F2 INF32, Otto-Hahn-Ring 6,
D-8 Munich 83, West Germany (if you have any papers that you could send
me.) Thanks a lot for your help! -- Jack
------------------------------
End of Software Engineering Digest
**********************************rsalz@bbn.com (Rich Salz) (01/19/89)
Why are the Software Engineering Digests appearing here? I'm a new reader of this group, please enlighten me. It seems that all us Usenet folks are just spending time and money to double the volume of the group... /rich $alz -- Please send comp.sources.unix-related mail to rsalz@uunet.uu.net.