home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!metro!otc!tbird2!chrisf
- From: chrisf@nms.otca.oz.au (Christopher Fraser)
- Newsgroups: comp.software-eng
- Subject: Graphical Object Design Tools (OMT, Booch..)
- Keywords: object oriented design tools
- Message-ID: <chrisf.722257945@tbird2>
- Date: 20 Nov 92 11:12:25 GMT
- Sender: news@otc.otca.oz
- Lines: 44
-
- I've been looking at James Rumhaugh's OMT design methodology over the past few
- days (well, actually, attending a seminar he's been presenting), and was
- wondering if anyone would like to share their experiences with actually
- applying a methodology like OMT or Booch (even if you just used the object
- notations to document your own design methodology). I'm particularly interested
- if you used a graphical design tool.
-
- My general impression of OMT itself was fairly positive. The object model
- basicly a sanitised version of ER techniques, with additional notations like
- inheritance (I though his concept of information paths was especially useful).
- The Dynamic model was based on state transition diagrams, with some nice
- techniques for modularising the schema. The functional model was, well,
- functional decomposition and was presented as something that could feed back
- into the object model (Rumhaugh talked a lot about iterative design ... which
- was comforting).
-
- Methodology aside I though the actual OMT notion itself captures a lot of design
- information in a fairly succinctly, formal form. However, the thought of manually
- maintaining the diagrams didn't particularly appeal to me; hence this
- post. (Much more appealing was the thought of design tools developing into tools
- which could actually help the developer track design constraints, like Rhumbah's
- information paths).
-
- I had a look at the Rational tool, Rose, and wasn't especially impressed.
- It's a nice drawing tool (but then, so is MacPaint!) using Booch's
- constructs. Mostly, it seemed lacking in the ways it could present the actual
- schema information you had painstaking entered; you couldn't filter out
- information on the schema, you couldn't merge them together and queries
- on the relationships an object was involved in only returned a textual list.
- (I probably shouldn't really comment, since I'm not especially familiar
- with building them, but schemas built using Booch's notation seem more
- cluttered and less design information dense than the equivalent OMT
- representation. The Booch notation certainly captures a lot more implementation
- semantics. Opinions?).
-
- I was also told Ratonal were working on tools to reverse-engineer a Booch
- schema for C++ code. They also demonstrated a beta version of a tool for
- generating code stubs (which seemed lacking in the required configurability).
- Does anyone else know of work being done in these areas?
-
- Cheers,
- --
- Christopher Fraser "First time surrealists are often confused by the
- chrisf@nms.otca.oz.au similarities between fish and telephones"
-