NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/software/5010

Text File | 1992-12-16 | 23.7 KB | 515 lines

Newsgroups: comp.software-eng
Path: sparky!uunet!haven.umd.edu!decuac!pa.dec.com!space!sp_pm
From: [e-mail] (Paul Moore)
Message-ID: <[e-mail]>
Subject: HOOD - Summary of responses
Date: Wed, 16 Dec 92 15:41:16 +0100
X-Received: by usenet.pa.dec.com; id AA04721; Wed, 16 Dec 92 08:22:55 -0800
X-Received: by inet-gw-1.pa.dec.com; id AA18033; Wed, 16 Dec 92 08:22:51 -0800
X-Received: from ub4b.buug.be by mcsun.EU.net with SMTP
    id AA17751 (5.65b/CWI-2.198); Wed, 16 Dec 1992 17:21:06 +0100
X-Received: from alcbel.UUCP by ub4b.buug.be (5.65c/ub4b_02)
    id AA14868; Wed, 16 Dec 1992 17:21:14 +0100
X-Received: from se.alcbel.be by alcbel.be (4.1/SMI-4.1)
    id AA24368; Wed, 16 Dec 92 15:34:20 +0100
X-Received: by se.alcbel.be (5.57/SE_04)
    id AA06633; Wed, 16 Dec 92 15:42:27 +0100
X-Received: from btmpjw.space by space.alcbel.be (4.1/SMI-4.0)
    id AA25259; Wed, 16 Dec 92 15:41:16 +0100
X-To: comp.software-eng.usenet
Lines: 493

Hello,

Some time ago, placed a note requesting information and comments about HOOD. This note summarises the responses received. My thanks to the following people for offering their comments and advice:
    [e-mail] (Eoin Woods)
    [e-mail] (Pete Cornwell ?)
    [e-mail] (Abdrew Ormsby)
    [e-mail] (?)
    [e-mail] (Laurent MANGANE)
    [e-mail] (Francisco Javier Campos)
    [e-mail] (Gerard Gobillard)

In summary, it appears that HOOD is a method that does not align itself very well with traditional structured analysis techniques (Yourdon, Ward/Mellor, Hatley, etc), due to its being an object-oriented design technique. It also doesn't appear to integrate very well with the currently in-vogue object-
oriented analysis techniques. HOOD appears to have been forced on the European Space Agency by a French aerospace company (Matra, I believe, but open to correction). However, it appears to be well supported by a number of CASE tool suppliers, even though the documentation describing the methodology in detail is more or less non-existent.

Anyway, my apologies for taking so long to reply. By the way, I didn't have any "Please let me have what you have received" type of requests, but I feel that the responses will be of sufficient interest anyway.

If you have any comments/observations, please email me, as I don't have a News feed.

Paul Moore

---
email: [e-mail]

Here follows my original posting, followed by the responses I received back.

===============================================================================


Hello,

I have just started working for a company working in the aerospace field. I
have some questions about the use of HOOD, and I would be grateful for some advice.

1. I have had some experience with the Ward/Mellor and Hatley/Pirbhai
   structured analysis methodologies. There are two excellent books available
   that explain how to use these methodologies [1] and [2], each with
   lots of examples showing how the method is used. It seems that the only
   documents available explaining how to use HOOD are the HOOD User Manual,
   and the HOOD Reference Manual. These two manuals do not explain how to use
   the methodology, and certainly do not explain the methodology to new users.
   Are there any better books available that teach the use of HOOD? It should
   *NOT* be necessary to attend a training course to learn how to use HOOD.

2. It seems that HOOD is used at the structured design level of the
   analysis/design/implementation software project cycle, as defined by
   Constantine/Page-Jones and Yourdon/Ward/Mellor. This means that its use
   should be preceeded by a structured analysis phase. Is this correct? How
   does it relate to modern structured analysis techniques, such as the two
   mentioned in point 1 above, and which structured analysis methodology is
   recommended for use with the HOOD methodology?

3. It seems that HOOD is oriented towards Ada as the implementation language.
   Does this mean that HOOD is still mandated by ESA when the language used
   is not Ada, for example C and C++. If so, can HOOD still be used if the
   implementation language is not Ada. For example, HOOD appears to take
   advantage of Ada's built-in inter-process-communication facilities. What
   if these facilities are not present in the implementation language used?

I will summarise responses I get to the net, if sufficient interest warrants.

Please respond by email, as I don't have a connection to the News.

Regards, Paul

[1] Structured Development for Real-Time Systems, Vols 1-3,
    Paul T. Ward and Stephen J. Mellor, 1985.
[2] Strategies for Real-Time System Specification,
    Derek J. Hatley and Imtiaz A. Pirbhai, 1987.

---
Paul Moore,                      email: [e-mail]
Alcatel Bell Telephone,
Space Dept. TS 664,              phone: [phone]
Berkenrodelei 33,
2660 Hoboken,
Belgium

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

From [e-mail] Tue Oct 13 19:08:57 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA03938; Tue, 13 Oct 92 19:08:55 +0100
Received: from uk03.bull.co.uk by alcbel.be (4.1/SMI-4.1)
    id AA18597; Tue, 13 Oct 92 19:04:46 +0100
Received: from bertha.uk03.bull.co.uk by ub4b.buug.be (5.65c/ub4b_02)
    id AA09826; Tue, 13 Oct 1992 16:39:48 +0100
Received: from brno.uk03.bull.co.uk by bertha.uk03.bull.co.uk with SMTP
    (15.11.1.4/15.6) id AA29325; Tue, 13 Oct 92 12:39:43 bst
Received: by brno.uk03.bull.co.uk; Tue, 13 Oct 92 11:40:52 GMT
Date: Tue, 13 Oct 92 11:40:52 GMT
From: [e-mail] (Eoin Woods)
Message-Id: <[e-mail]>
To: [e-mail]
Subject: Re: Repost: Please reread: Has anyone successfully used HOOD?
Newsgroups: comp.object
References: <[e-mail]>
Status: RO

There is a collection of papers published by UNICOM in the
UK which gives interesting background information. I think the
training course is the best option (I learned HOOD as part of my MSc
course).

Yes, HOOD is primarily a design method, not an analysis one. I have
used W&M with it and it seems to work OK! I don't think there is an
"official" analysis method for it.

It is *very* Ada biased. It has been used with other
languages, but presumably they built a library to simulate the Ada
facilities they needed.

I did hear recently that ESA are using OOD (based on Booch) as well as
HOOD (particularly for C++ project). However this is just hearsay, not
fact.

Eoin.
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| Eoin Woods                         | Tel : [phone] (Fax 4570)     |
| Application Services Product Unit  | BVN : (UK) 773-4467                  |
| Bull Information Systems           |                                      |
| Maxted Road                        | [e-mail]              |
| Hemel Hempstead, Herts HP2 7DZ, UK | ...!relay1.uu.net!uknet!brno!ewoods  |

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

From [e-mail] Wed Oct 14 17:08:41 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA04123; Wed, 14 Oct 92 17:08:39 +0100
Received: from minster.york.ac.uk by alcbel.be (4.1/SMI-4.1)
    id AA28579; Wed, 14 Oct 92 17:04:27 +0100
Received: from minster.york.ac.uk by ub4b.buug.be (5.65c/ub4b_02)
    id AA24380; Wed, 14 Oct 1992 15:02:14 +0100
From: [e-mail]
Date: Wed, 14 Oct 92 14:44:48
Message-Id: <[e-mail]>
To: [e-mail]
Subject: HOOD
Status: RO

I work at York University on Hard real-time HOOD which is essentially a
superset of the original notation, so I've had a bit of experience in
some of the areas you mentioned in your news posting...

to summarise....

1) I've only come across one other book that deals with HOOD in any
depth and thats coolings book on real-time system design which gives
a very general introduction to the method.  When I originally started looking
at HOOD all that was avaliable however was the reference and user
manuals, and yes I agree these are generally pretty lousy.

2) I believe HOOD takes the view of "here is the requirement document,
lets start the design". Using the methods you cite would be difficult as
HOOD is such a weird compromise between top-down and OO design methodologies.

3) Although the method is strongly related to Ada, there is movement
within the HUG (HOOD USer Group) to gradually move away from Ada and
attempt to make HOOD more generic, and hence more appealing to non-Ada
using designers.  At the moment however the various mapping rules only
cover HOOD to Ada'83 , sorry I should say "guidelines" :-) version 3.1
has guidelines rather than rules!  Basically all the guidelines state
for other languages is that as long as it has similar modularity /
encapsulation features to Ada you should be OK (!).

NB: I believe that version 4.0 of HOOD will have inheritance, but when this
is due out is anyones guess.  Also I dont know if this is an attempt to
keep in step with the proposed changes to Ada 9X or to appeal to the
OOP using community at large.

...anyway hope my limited experience in this area has been of some use.

Pete.

([e-mail])

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

From [e-mail] Thu Oct 15 21:47:40 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA04401; Thu, 15 Oct 92 21:47:39 +0100
Received: from aberystwyth.ac.uk by alcbel.be (4.1/SMI-4.1)
    id AA09518; Thu, 15 Oct 92 21:43:24 +0100
Received: from elem4.vub.ac.be by ub4b.buug.be (5.65c/ub4b_02)
    id AA15206; Thu, 15 Oct 1992 12:08:46 +0100
X400-Received: by mta elem4.vub.ac.be in /PRMD=iihe/ADMD=rtt/C=be/; Relayed;
               Thu, 15 Oct 1992 12:06:25 +0100
X400-Received: by /PRMD=uk.ac/ADMD= /C=gb/; Relayed;
               Thu, 15 Oct 1992 12:06:34 +0100
X400-Received: by /PRMD=UK.AC/ADMD= /C=GB/; Relayed;
               Thu, 15 Oct 1992 12:06:54 +0100
Date: Thu, 15 Oct 1992 12:06:54 +0100
X400-Originator: [e-mail]
X400-Recipients: [e-mail]
X400-Mts-Identifier: [/PRMD=uk.ac/ADMD= /C=gb/;<[e-mail].]
X400-Content-Type: P2-1984 (2)
Content-Identifier: Repost: Pleas...
From: [e-mail]
Message-Id: <[e-mail]>
To: " (Paul Moore)" <[e-mail]>
Subject: Repost: Please reread: Has anyone successfully used HOOD?
Reply-To: " (Andrew Ormsby)" <[e-mail]>
Status: R

I know of only one book which purports to explain how to use HOOD.
This is "Object-Oriented Design" edited by Peter Robinson. I'm not
terribly impressed by it; it is just a collection of papers given at a
UNICOM seminar in London, is not well produced and is very, very
expensive. If you are still interested, the book is published by
Chapman & Hall, a UK publisher.

I'm not sure what the official position is on what you should use as
the precusor to HOOD. The best people to ask might be those currently
involved in its use on ESA and related projects. My view would be that
some sort of object modelling analysis method would be far more
appropriate start. You might find some useful hints in the book by
Coad and Yourdon, "Object-Oriented Analysis", (Yourdon
Press/Prentice-Hall), though their notation is (of course) different
to the HOOD notation. A better analysis method is described in
"Object-oriented modellinand design" by Rumbaugh and others. Also
published by Prentice-Hall.

You might also talk to the various companies that are selling CASE
tools for supporting use of HOOD and see what they think you should be
doing. The company I know about is IPSYS; there are a number of others.

From my reading on HOOD, the method seems to be highly Ada dependent.
Although the HOOD manual asserts that the notation is not just another
Ada design notation, there seems to be no guidance on how one might
hope to map HOOD concepts onto (for example) C++. Perhaps things have
improved since I looked.

It certainly isn't clear to me how people manage to use HOOD in
practice. I'd be very interested to hear from you if you find out
more.

Thanks,

Andrew Ormsby, [e-mail]
University of Wales, Aberystwyth.

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

From [e-mail] Fri Oct 16 01:34:05 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA04448; Fri, 16 Oct 92 01:34:03 +0100
Received: from cs.ualberta.ca by alcbel.be (4.1/SMI-4.1)
    id AA10155; Fri, 16 Oct 92 01:29:48 +0100
Received: from scapa.cs.ualberta.ca by ub4b.buug.be (5.65c/ub4b_02)
    id AA03690; Thu, 15 Oct 1992 18:00:55 +0100
Received: from scapa.cs.ualberta.ca by scapa.cs.ualberta.ca with UUCP id <42133-1>; Thu, 15 Oct 1992 10:59:55 -0600
Received: from agt (rnd-svrn) by agt.uucp (4.1/SMI-4.1)
    id AA22544; Thu, 15 Oct 92 10:24:13 MDT
Received: from [128.9.32.22] by agt (4.1/SMI-4.1)
    id AA17159; Thu, 15 Oct 92 10:24:09 MDT
Date:     Thu, 15 Oct 1992 10:24:08 -0600
Message-Id: <9210151624.AA17159@agt>
To: [e-mail]
From: [e-mail]
Subject: HOOD
Status: R

Hi!

Have you looked at Shlear/Mellor's latest book (1991)
'Modeling the World in States'?

This may give you some good ideas.

+ August Guillaume      Internet:  [e-mail]
+ Senior Researcher
+ AGT Research & Development
+ 500 Capitol Square
+ 10065 - Jasper Ave            Voice: [phone]
+ Edmonton, Alberta CAN         Fax:   [phone]
+ T5J-3B1

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

From [e-mail] Mon Oct 19 01:34:20 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA05371; Mon, 19 Oct 92 01:34:18 +0100
Received: from SOLEIL.matra-espace.fr by alcbel.be (4.1/SMI-4.1)
    id AA11323; Mon, 19 Oct 92 01:29:55 +0100
Received: from corton.inria.fr by ub4b.buug.be (5.65c/ub4b_02)
    id AA25932; Fri, 16 Oct 1992 18:02:25 +0100
Received: from MATRASP.UUCP by corton.inria.fr (5.65c8d/92.02.29)
    via Fnet-EUnet id AA06714; Fri, 16 Oct 1992 16:08:34 +0100 (MET)
Received: from SOLEIL.matra-espace.fr by MATRASP.matra-espace.fr, Fri, 16 Oct 92 15:49:12 GMT
Received: by SOLEIL.matra-espace.fr, Fri, 16 Oct 92 15:37:58 +0100
Date: Fri, 16 Oct 92 15:37:58 +0100
From: [e-mail] (Laurent 7557)
Message-Id: <[e-mail]>
To: [e-mail]
Subject: Somes answers to your questions on HOOD
Status: RO


Dear Paul:

Your first sentence is partly the answer to the question of your subject. HOOD
has already been successly used in this particuliar field and in Europe because
it is a European method. But as it exists only since 1987, I think no system
operational today has been designed with HOOD, I believe the first ones will be
Ariane 5 or EFA.

1. You are totally right. For the same reason as mentioned above, there is no
other OFFICIAL documentation on HOOD than HOOD User Manual 3.0 (December 89)
and HOOD Reference Manual 3.1.1 (August 92 to be published soon by Prentice
Hall). As the method has been improved since December 89, everybody is
expecting an updated version of the HOOD User Manual. This update is already
available internally in my company thru a three-day course which can be
attended by everybody on request. It should be necessary to learn HOOD (and
then to experiment it on short examples) *BEFORE* being forced to use it on a
real project and not *AFTER*: you should mention it to the aerospace company
which you are working for.

2. HOOD is also suitable for unusual software project cycle (Boehm's spiral for
instance) and not only for the waterfall model mentioned by
Constantin/Page-Jones and Yourdon/Ward/Mellor (but not defined by them because
it existed before the creation of any method) which is an old model only used
with old methods. Anyway, be aware that HOOD is not a structured design method
but a HIERARCHICAL OBJECT-ORIENTED design method which implies a totally
different approach difficult to learn (especially alone) for people who have an
important experience in classical (not so modern) approachs such as structured
or functional analysis. You just have to remember that a design method must be
used *AFTER* an analysis phase and *BEFORE* implementation: do not laugh, we
met people who had used HOOD after having implemented in ADA the requirements
and obviously their implementation was unusable and it was impossible to
translate it in HOOD.

HOOD like most design methods can be used with any analysis method (SA,
IDEF...) with no other links with it than a dictionnary which can be reused to
name HOOD objects' operations (it is the same case with all other
object-oriented design methods because no widely used analysis method allows
today a complete and easy traceability with an OO design). No analysis method
is recommended because none fully suits HOOD's approach of design.

3. HOOD is mandated by ESA for any implementation language and has still been used
with ADA, Pascal, Assembler and C. Naturally, it is easier to use ADA because
the method includes rules to derive ADA code from the ODS but some companies
have established such rules with C code (it works properly in most cases).

As for C++, ESA cannot mandated HOOD with it as it never mandated the use of
C++ for coding phase. By the way, if you know a genuine DESIGN method (I mean
not graphical programming as OOSD) which can be used with C++ LARGE projects
(if this coding language can be used on large projects..??), please let me know
(I am still searching such a method for years).

HOOD allows to design Real-Time systems and if your implementation language
and/or environment does not provide R/T facilities, you just have to create it
by yourself: it has already been done for projects using pure assembler on a
sole processor with no R/T monitor. The method only provide a frame to express
these aspects in the OBCS by using any formalism of your choice like in SA/RT
(SDL, Esterel, Petri nets, Finite-states automata... etc) depending on what is
supported by your HOOD tool or by you but the implementation still depends (and
will always) on the target environment.

I hope this will help you in using this new method (only 5 years old but still
adult) and you can ask further questions if needed to some of the persons who
have created it and which are all listed in the ACKNOWLEDGEMENTS part of the
HOOD Reference Manual. You should consider it as you considered Kernighan &
Ritchie's for C programming or Wirth and Jensen's for Pascal i.e. unusable to
master those languages but essential to be able to use them intensively.

| Laurent MANGANE                 |
| Matra Marconi Space France             |
| 31, rue des Cosmonautes             |
| 31077 Toulouse Cedex                 |
| France                     |
| Tel: [phone]             |
| Fax: [phone]             |
| e-mail (UUNET): [e-mail] |
|          or  [e-mail]    |
|        (X400) : C=FR ; PD=MATES400 ; S=MANGANE |

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

Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA06237; Wed, 21 Oct 92 14:42:36 +0100
Received: from gmv.es by alcbel.be (4.1/SMI-4.1)
    id AA06109; Wed, 21 Oct 92 14:38:04 +0100
Received: from goya.uu.es by ub4b.buug.be (5.65c/ub4b_02)
    id AA28779; Wed, 21 Oct 1992 10:54:58 +0100
Received: from eonwe.UUCP by goya.uu.es (EUnet) (5.65c/8.12); Wed, 21 Oct 1992 10:54:41 +0100
Received: from conejo.gmv.es by gmv.es (4.1/GMV-1.6)
    id AA21881; Wed, 21 Oct 92 10:44:30 +0100
Date: Wed, 21 Oct 92 10:44:30 +0100
From: [e-mail] (Francisco Javier Campos)
Message-Id: <[e-mail]>
To: [e-mail]
Subject: Re:  Can you give me some HOOD information, please?
Status: RO


Hello; sorry for the delay in answering you.

1. Unfortunately, there are no books available on the
method. I know that there will be one text book
available very soon; the author is P. Robinson
(formerly in charge of HOOD at ESTEC), but I have no
exact dates for this. The only way to get in contact
with the method is to attend the training courses. A
good way to be in contact with the HOOD community and
to be able to talk to the "gurus" is being meber of the
HOOD Users Group. If you are interested, I can give you
the contact point.

2. HOOD is indicated for the Architectural and Detailed
Design phases (high0level and low-level design), where
a pseudo code is produced that may be easily
transformed into Ada. There is no link between
rqeuirement analysis and the Design in the HOOD method;
that is, only the expertise of the enginner is
applicable to get the HOOD objects from the
requirements. With this approach, it is not
important whether you use a structured analysis or
object oriented analysis method. There is a lot of
literature about the convenience or not of object
oriented analysis methods for object oriented design
methods, and there is no consensus at all.

3. Certainly HOOD is, at the time being, oriented towards
Ada code. No guidelines have been defined yet for
production of software in other languages. The basic
idea is that if you use other languages you will have
to resemble the Ada facilities by yourself. I have no
experience on this topic, but some tools (e.g. AdaNICE,
by INTECS) generate C code from a HOOD design.


I hope this will answer your questions.



Javier Campos                            Ph. [phone]
GMV, S.A.                                Fax [phone]
C/ Isaac Newton, s/n                     Telex  48487 GMEV E
TRES CANTOS                              [e-mail]
28760 MADRID                             [e-mail]
SPAIN                                    uunet!mcsun!gmv.es!jcampos

-------------------------------------------------------------------
From [e-mail] Fri Oct 23 17:08:50 1992
Return-Path: <[e-mail]>
Received: from alcbel.be ([138.203.64.80]) by space.alcbel.be (4.1/SMI-4.0)
    id AA06685; Fri, 23 Oct 92 17:08:48 +0100
Received: from rcvie.co.at by alcbel.be (4.1/SMI-4.1)
    id AA06920; Fri, 23 Oct 92 17:04:10 +0100
Received: from hp4at.eunet.co.at by ub4b.buug.be (5.65c/ub4b_02)
    id AA20208; Fri, 23 Oct 1992 12:16:55 +0100
Received: from rcvie.rcvie.co.at (rcvie-c.rcvie.co.at) by hp4at.eunet.co.at with SMTP id AA00976
  (5.65c8/IDA-1.4.4 for <[e-mail]>); Fri, 23 Oct 1992 12:16:11 +0100
Received: from rcsw26.rcvie.co.at by rcvie.rcvie.co.at (4.1/RCVIE-Main-1)
    id AA01167; Fri, 23 Oct 92 12:14:38 +0100
Date: Fri, 23 Oct 92 12:14:38 +0100
From: [e-mail]
Message-Id: <[e-mail]>
To: [e-mail]
Subject: HOOD
Status: RO


Dear Paul,

Forgive me if my information is fuzzy. If you are really
interested, I can spend some time making it more accurate.

There is a book in French called " conception orientee object,
application a HOOD" which describes the method. The guy (Michel LAI) also
developed a C++ customization of HOOD. I had the book in my hands,
it seems very quickly written.

There is a book in English by a guy called ROBINSON called spomething like
"software engineering with HOOD". I saw it in a bookstore in Paris.
It is expensive (~1ooo ff), and seems technically sound.

I also heard that Maurice HEITZ is preparing a book (he is the inventor
of HOOD). I don't know if it is in F or E. The guy is very nice and I think he
answers questions by phone.

Don't hesitate to call me if you need more help from me.

Gerard Gobillard
mail: [e-mail]
phone: [phone]