Re: Shlaer-Mellor OOA/RD with Ada

[helliott@losat.redstone.army.mil (Doc Elliott)]

In article <3iat22$pl6@butch.lmsc.lockheed.com>, 
brockwel@zuni.litc.lockheed.com says...
>
>Are there any subsribers who have experience using
>Shlaer-Mellor OOA/RD to develop real-time Ada
>applications?  
>
>Specifically, has anyone automated the architecture
>design / mapping rule definition process?
>
>Please email me if you have any information or advice to
>offer a newcomer to Shlaer-Mellor for Ada applications.
>
>brockwel@lightning.hunts.lmsc.lockheed.com
>
>(205) 722-4080 (phone)
>.
>

We did a study comparing all of the OOD methods available at the time 
(~92?) and looked at tool support, mapping into Ada, availability of 
training etc etc.  I can probably get you a report released through our 
security weenies, but the bottom line was Shlaer - Mellor sucked for real 
time systems.  It has very little capability for defining system 
boundaries - the subtitle of their book was "Modelling the world" and 
that's kinda what you end up doing, rather than modelling your problem 
and solution space.  SM is a data flow oriented model, it depends on what 
appears to be enhanced data flow diagrams for almost all of its database. 
 This is not very good for interrupt driven systems which must be 
deterministic.  Also, the tools available to support it required use of 
some additional methods (such as Harel statecharts) to allow any 
simulation of the system model.

Let me know directly if you want me to get the report released to you - 
it may take a while, and the info is probably somehwat out of date, 
especially the tool support part.
-- 
Doc Elliott
KE4KUZ
Internet: helliott@losat.redstone.army.mil
packet: ke4kuz@k4ry.#cenal.al.usa.noam
The opinions expressed herein are mine, and do not
reflect those of my employer or anyone else unless
specifically stated as such.