Talk:Shlaer–Mellor method

Discussion 2006
This is promotional and commercial material. —The preceding unsigned comment was added by 65.83.52.198 (talk • contribs).


 * How exactly is this page promotional or commercial? What is it promoting, Shlaer-Mellor?  GeorgeBills 06:02, 15 June 2006 (UTC)

Wrong Steve Mellor
It looks as if the link to Stephen Mellor goes to the American actor of that name, not the English methodologist. --Jhebley 16:44, 18 January 2007 (UTC)

Fractured sentence
Taken as written, the following sentence says that 'describing [a certain thing] becomes a design'. This is clearly not what the writer intended to say. 'Describing' in functional analysis isn't analogous to 'a design' in semantic analysis.

"That is to say, describing the control of a passenger train as load passengers, close doors, start train, stop train, open doors, unload passengers becomes a design focused on the behavior of doors, brakes, and engines, and how those "domains" (doors, brakes, etc.) are related and interact.

I think that the writer's intent was to convey the difference in how a system is decomposed in *functional* and *semantic* decomposition, respectively.

A first try to reword it:

"Functional decomposition means that the system under study is broken down into a series of functions--that is, actions or operations performed by the system. In the case of a system to control a passenger train, one might describe it as 'load passengers, close doors, start train, stop train, open doors, unload passengers.'  In functional decomposition, each element starts with an action verb; the common noun--the class of thing being acted upon--is secondary. "In contrast, semantic decomposition first breaks down a system into a set of classes of things (ie., a set of common nouns). The passenger train analysis becomes 'train, passenger, doors'. The class described by the common noun "door" has two behaviours: "open" and "close". Thus, functional analysis's "open doors" and "close doors" (appearing at different places in the specification!) become the (closely grouped) facts that "doors [can] open" and "doors [can] close". Since semantic analysis at first gives only the possible behaviors of various kinds of people, places, and things, of course one has not specified a system that meets functional requirements until one has done a second "process modeling" step. "The advantage claimed is that semantic analysis collects all the currently relevant knowledge about "doors" in one place, whereas in functional analysis (and thus, in the resulting designs and implementations), it is spread out all over. If it becomes necessary to improve the knowledge about doors in the analysis, either because (a) a bug was found in the first system, or (b) because the analysis is later being re-used in a different system or (c) the original system is being functionally extended, then the change can be done more quickly, or more reliably, or with less time and resource.

Comments? Mark.camp (talk) 21:10, 1 August 2009 (UTC)

Design dependence
I believe the author's intention with the sentence "The analysis domain is considered to be dependent upon the design domain." was in fact the reverse of what is written, as in "The design domain is considered to be dependent upon the analysis domain."

Duncan George —Preceding unsigned comment added by 99.247.135.102 (talk) 15:32, 6 October 2009 (UTC)

Shlaer–Mellor, OOSA and OOA
The (new) reference to OOSA had been removed, see here with the argument. Shlaer-Mellor has never been commonly known as Object-Oriented Structured Analysis. Now this could be true in the States, but apparently not in Germany.

I came up with the idea to mention OOSA here, because this name is listed image of the UML history (see images). Both images are drawn by the associates of the German Dienstleistungen für Innovative Informatik GmbH, nowadays Oose Innovative Informatik GmbH.

The first image is constructed for the German Wikipedia and published since 2004. And the second image is an update from 2008 published in the English Wikipedia since 2008 (after I introduced it, see here).

Now I guess there are a couple of things we can do here:
 * 1) Double check the most common other names/abbreviations of Shlaer–Mellor method in the since the 1980s in the States and World wide.
 * 2) Correct the image (it also  still needs to be translated)
 * 3) And... possibly update the image adding a line to Executable UML if acceptable
 * 4) Possibly add the image to this article.

-- Mdd (talk) 23:45, 3 December 2012 (UTC)


 * The removal wasn't due to the abbreviation, OOSA, but the mistaken expansion as, Object-Oriented Structured Analysis. I had thought of just correcting it to be, Object-Oriented Systems Analysis, which was in the title of the first book, but Shlaer and Mellor had started calling it just Object-Oriented Analysis before the second book. The meta-model for the Shlaer-Mellor method was named, OOA of OOA. Lwriemen (talk) 03:37, 5 December 2012 (UTC)


 * I am a bit confused here. This method was introduced with the 1988 book:
 * Sally Shlaer, Stephen Mellor (1988) Object-Oriented Systems Analysis: Modeling the World in Data, Yourdon Press. ISBN 0-13-629023-X
 * This book is the most cited work of Shlaer and Mellor, according to Google-Beta 866 times (see here). And you don't want to use the term, because Shlaer and Mellor didn't use it anymore in their second book?


 * We have to take in account here, that Wikipedia is not representing the ideas and opinions of single (original) authors. Wikipedia represents the way these ideas have come to life in (reliable) sources. So the argument that Shlaer-Mellor didn't use the term any more in their second book doesn't count much. We have to rely on independent reliable sources.


 * Now Google Beta states the term "Object-Oriented Systems Analysis" is mentioned 1.980 sources. Of these 1.980 already 866 relate to that first book of Shlaer-Mellor's. So it seems only logical to add the term here. -- Mdd (talk) 15:17, 5 December 2012 (UTC)


 * I added the back in the lead, added some more details in the "Criticisms"section. Now the article is more in line with the sources, the image and also the external link to smartdraw.com. -- Mdd (talk) 22:06, 5 December 2012 (UTC)


 * I have updated the image and added xUML. -- Mdd (talk) 01:15, 7 December 2012 (UTC)
 * Your timeline is confusing. xUML predates UML 1.5.Lwriemen (talk) 14:48, 28 April 2013 (UTC)
 * Two papers available here Show migration of Shlaer-Mellor to UML actually started in 1997, and Leon Starr released his Executable UML case study in 2001. I would still use 2002 as the hallmark date. Lwriemen (talk) 16:17, 1 May 2013 (UTC)