Lahman's census of OOAD tools using Action Semantic Languages::
RJLRef: ~/01f522/smu/smuActionSemanticsCensus.lahman

Note: I don't know why Lahman ignores 'round-trippers' below.
To me, it means back-translation or reverse-engineering of code to models,
after models have been converted to code which might be modified
directly. Within what I suspect are serious coding constraints, 
code changes are reflected back to model changes to keep them both 
in sync. Another term for this capability is back-annotation.

[I have never seen back-annotation in practice, working or not.
A famous paper in CACM by Bohmm & Jacopini showed that [pointer-less?] 
code modeled as a flowchart could be reverse-engineered into a structured 
flowchart (and code) by adding a stack of bits. This stack eliminated 
the  binary decisions that branched to the notorius GO-TO instruction 
labels  by adding nested control states as needed. These augmented 
the flowchart's structured branch-points, to produce a structured set 
of nestable state models and code. This might lead to a constructive 
proof that back-anotation works. (With pointers all bets are off.)]
 
Forwarde  message:
> From owner-shlaer-mellor-users@phoenix.projtech.com  Mon Sep 10 17:22:37 2001
> From: "H. S. Lahman" <h.s.lahman@worldnet.att.net>
> To: shlaer-mellor-users@projtech.com
> Subject: Re: (SMU) Asbestos Necessary
> 
> "H. S. Lahman" <h.s.lahman@worldnet.att.net> writes to shlaer-mellor-users:
> --------------------------------------------------------------------
> 
> Responding to Case...
> 
> > (1) It should be well understood that the UML (either 1.4 or 2.0 [as it
> > stands]) is *not* executable. The SM experience (of old?) is, however,
> > completely executable.
> 
> Actually, I think v2.0 will be.  Last I heard, that is when the action
> language meta model is supposed to be released.  That is the last piece
> necessary to make UML executable.  [Though it leaves a lot of vendor
> options open.  B-)]
> 
> > (2) How many of the tooling vendors out there that (truly) support model
> > based compilation have their technology rooted in SM and key method
> > contributors? My guess is most of them. Discounting those (with no ill
> > intent) of "insert your <action sematics> code here" pathology.
> 
> If one <correctly> ignores the round-trippers, then all the tools that I
> know of are:
> 
> ARTiSAN -- S-M (abstract AL)
> Bridgepoint -- S-M (abstract AL)
> Intelligent-OOA -- S-M (abstract AL)
> ObjectBench -- nondenominational UML (C++)
> Rhasody -- ROPES (C++)
> Rose/Pathfinder -- S-M (Abstract AL)
> Rose/RT -- ROOM (C++)  [nee ObjecTime]
> System Architect -- S-M (Abstract AL)
> 
> Note the interesting correlation between S-M and abstract AL vs. non-S-M
> and C++.  Amazingly, I never noticed that until I built this list!  [The
> S-M/abstract tie is somewhat biased since Pathfinder provides
> compilation for three of them.  B-)]
> 
> *************
> There is nothing wrong with me that could
> not be cured by a capful of Drano.
> 
> H. S. Lahman
> hsl@pathfindersol.com
> Pathfinder Solutions  -- We Make UML Work
> http://www.pathfindersol.com
> (888)-OOA-PATH
>