OO Req. Analysis Methods

OO Req. Analysis Methods

[sbailin]

About a month ago I posted a request for information about OO
Requirements Analysis methods that people are using.  I'd like to
thank those who responded to me directly, and am posting this note
to summarize those responses.

I received 12 replies, several of them indicating use of a combination
of methods.  The counts of those using particular methods are as
follows:

Booch:                  3
Coad Yourdon            2
Fusion                  2
Jacobsen                3
OMT (Rumbaugh et al):   3
Shlaer-Mellor           1
Unified                 3
Wirfs-Brock             2

In addition, 3 replies indicated they were using some sort of business
rule modeling.  One reply cited Taylor's Business Engineering method,
and another cited the Business Object Notation of Walder and Nerson.

Combinations cited were Coad-Yourdon + Jacobsen + CRC cards,
OMT + Fusion (the latter for modeling aggregation), and
Booch + Jacobsen (use cases) + business rules.

I doubt that the numbers here mean anything about the respective
popularity of the methods cited, but I think the replies give one
an idea of what methods are being used.

On the basis of these replies, and the fact that Booch, Jacobsen, and
Rumbaugh are consolidating their approaches in the Unified method,
I have decided to update my OORA article with summaries of the
following methods:  Coad-Yourdon, Shlaer-Mellor, Wirfs-Brock, Fusion,
and Unified.  Fusion was not included in the Encyclopedia of Software
Engineering version of the article, and Unified will replace OMT.
A couple of the other methods I described in the earlier version will
drop out.  In addition, Shlaer-Mellor has moved some distance in the
intervening years, and I plan to document this.
 
Many thanks again for those who replied, and I hope this summary is
of some interest.
 
Sidney Bailin
Knowledge Evolution, Inc.