OOram: upcoming book, review 
Author Message
 OOram: upcoming book, review

I recently reviewed a preprint manuscript for a new OOSE methodology book
due out later this summer:

        Working With Objects:
        The OOram Software Engineering Method

        Trygve Reenskaug, with P. Wold and O.A. Lehne

        Manning Publications Co., 1995

This article is a brief summary of the full review at:

        http://www.*-*-*.com/ ~alyon/BBC/Reenskaug/Sutter_Review_OOram.html

(Note: The full review was originally intended for the publisher and
authors only, so the body reads more like an email than a normal published
review.)

OOram approaches OO from the point of view that the same object may play
different roles throughout its lifetime, and that the same roles may be
played at different times by different objects; further, a given method in a
role might be callable only from specific other roles.  OOram's main goal is
to formalise this concept and express it as a methodology and a notation,
with a view specifically to creating manufacturable software components such
as domain-specific application frameworks.  An appendix includes a first cut
at a language specification supporting these techniques, but the language
itself is incomplete and is at this point only suitable for generating
skeletons (rather than working programs) in some other complete OO language.

As a methodology, OOram embodies many good ideas.  In my opinion it makes a
great stride forward to realising its goal of manufacturable software
components, just as we already have standard manufacturable hardware
components in the matured engineering disciplines.  While I'm not convinced
OOram alone will be enough to take us all the way there (and the authors
note they don't see it as a silver bullet either -- perhaps an overused
phrase), it does have substantial advantages over existing OO methodologies
-- although at the cost of giving up (unnecessarily, in some cases) some of
the notational power of existing methodologies like Booch's.

Generic programming (e.g., the Standard Template Library described in
[Stepanov 95], now part of the C++ draft standard), is a different step
toward generic software parts, though at a lower level than OOram.  The
authors compare OOram with existing software engineering methodologies
(Booch, Jacobson, Rumbaugh, Wirfs-Brock) and design patterns efforts ([Gamma
95]); I recommended that they also mention how OOram would be applied to, or
work with, the ideas of genericity to get us to the stated goal.  In
particular, the authors specifically mention that OOram doesn't express
details like parameterised classes (pg. 36), which Booch's system does and
which is essential for [Stepanov 95]'s style of generic programming.  STL
relies completely on the template as the fundamental building block of its
approach.

The book is well-written and accessible; from the very first chapter the
text relates OOram to existing methodologies and even pattern studies,
mentioning books as current as Design Patterns (Gamma et al).  This helps
the well-read reader to relate this new information to existing knowledge.

I expect that this is a book we'll be hearing more about when it becomes
generally available in the late summer.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Herb Sutter                 2228 Urwin, Ste 102         voice (416) 618-0184
Connected Object Solutions  Oakville ON Canada L6L 2T2    fax (905) 847-6019



Tue, 02 Dec 1997 03:00:00 GMT  
 
 [ 1 post ] 

 Relevant Pages 

1. OOram: upcoming book, review

2. Book review review

3. Request for book review : OOP books

4. proofreading books, was Re: Book REview

5. Request for book review : OOP books

6. Authors wanted for upcoming book

7. Upcoming Book on Object REXX for OS/2

8. Info on upcoming Ruby books?

9. Authors wanted for upcoming book

10. Accidental second sending of first book review

11. Need Book Review

12. Clarion Book Reviews online

 

 
Powered by phpBB® Forum Software