home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.object:5091 comp.software-eng:5493
- Newsgroups: comp.object,comp.software-eng
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!decwrl!sdd.hp.com!ncr-sd!SanDiego.NCR.COM!aces!khalsa
- From: khalsa@SanDiego.NCR.com (G.K. Khalsa)
- Subject: Re: How About Using Bootch for Analysis?
- Message-ID: <1993Jan27.014937.20191@donner.SanDiego.NCR.COM>
- Sender: news@donner.SanDiego.NCR.COM (News Owner)
- Reply-To: g.k.khalsa@SanDiego.NCR.com (G.K. Khalsa)
- Organization: NCR Engineering and Manufacturing, San Diego, CA
- References: <mmaccorm.68@fox.nstn.ns.ca>
- Date: Wed, 27 Jan 1993 01:49:37 GMT
- Lines: 61
-
- In article <mmaccorm.68@fox.nstn.ns.ca> mmaccorm@fox.nstn.ns.ca
- (Mike Mac Cormack) writes:
-
- >My company is currently trying to decide which object oriented methodology
- >to use for our next project (approx. about 10,000 lines of code) and we had
- >the following questions:
- >
- >1) Can you use Bootch for both analysis and design? The first 3 steps of
- >his method are almost identical to Coad/Yourdon's OOA.
-
- Booch has a rich set of modeling concepts and a notation to support them.
- My personal opinion is that the handling of aggregation (Coad/Yourdon's
- whole-part) is the weakest part. But the real issue is _what_ you are
- modeling. This distinction divides analysis from design. Although there
- is an inevitable intertwining of specification and implementation, we
- must do our best to model the 'what' in analysis and add the 'how' in
- design. Some notation features from Booch may be inappropriate for
- analysis modeling. Items like scope and visibility seem to be
- software architecture issues - best left for design.
-
- The one other area where Booch's notation may be weak is in the
- ability to develop high level views of the model. For large systems,
- it is not practical to represent a specification as a flat, huge
- set of classes/objects and their interactions. Coad/Yourdon and
- Embley, Kurtz and Woodfield are two texts that describe ways to
- present a model in manageable chunks.
-
- >
- >2) Are there any case tools available (either now or in the next year) which
- >will allow you to reverse engineer Bootch diagrams from Ada code?
-
- I think one by Mark VI (or some roman numeral) has this capability.
-
- >
- >3) How about using Coad/Yourdon for analysis and Bootch for design?
-
- Sure. However, you give up some of the supposed benefits of OOA in
- terms of the seamless transition to design. A mapping between the two
- notations is possible, but the terminology will change too. In the
- long run, it could be a benefit to gain a second perspective by
- employing two modeling texts (since no one text is very comprehensive),
- but be prepared for some project confusion.
- >
- >Any comments/suggestions would be greatly appreciated!
- >
- >Mike Mac Cormack
- >Micronav International Inc
- >P.O.Box 1523 / 104 Marine Drive
- >Sydport Industrial Park,
- >Sydney, N.S. Canada B1P 6R7
- >Tel: 902-564-8833 Fax: 902-564-8764
-
- ...............................................................
- | | On Contract To: |
- | GK Khalsa | NCR Engineering and Manufacturing |
- |....................| 16550 W. Bernardo Dr. |
- | (619) 485-2460 | San Diego, CA 92127 |
- !....................!........................................!
- !.................g.k.khalsa@sandiego.ncr.com.................!
-
-
-