(responding to Sri) ... If translation does indeed do almost all your architectural thinking for you, then one would expect translation to do really well in ...
Sri ... Thanks! ... Yes, that's essentially it. ... Yes, you have the option of doing this. Some years ago, for fun, I implemented a multiple lift (elevator)...
... Sorry if I confused the issue :( I think from later posts headway has been made in creating an understanding of Ashley's position, so I won't try and...
... Ashley, My I interpret approach A as: Develop a PSM, test this PSM and then carefully re engineer a PIM out of it? Then continue with this PIM as usual. ...
... as a ... model /must/ ... I find this a strange statement, by definition a PIM (assuming you meant a PIM) is platform independent and my be transformed to...
Rersponding to H. S. Lahman ... The global wall-clock time is totally predictable and not easily realisable. The local wall-clock time is unpredictable but...
Jelle ... I don't think so. Step A1 produces a PIM -- as it does not assume a platform. OK, approach A does base the PIM on the assumption of serialised,...
REsponding to Ashley and Jelle ... Yes and No. Analyze the principles behind the PIM to PSM conversion. Express these principles as a re-usable transformation....
Responding to Oldfield... ... Did the regression test suite execute correctly after the refactoring but before you added the comms code? If so, it would...
Responding to McNeile... ... By 'object state machine' I mean that each object with significant behavior responsibilities has its own finite state machine...
Responding to McNeile... ... The operative word is 'ability'. The PIM must be unambiguously mappable into some computing space execution semantics. The major...
Responding to Nedunuri... ... This was certainly very nearly true in the '70s. The acronyms of USER and CRUD pretty much summarized a world where a key...
Responding to jdgerbrandy... ... That is exactly what I intended: I am saying that a manual system in the customer's domain is just another "platform". The...
Responding to Willink... ... My position is that one can address execution with one of three valid concurrency models: (1) serialized. Only one operation or...
... I just re-read what I wrote and realized it was somewhat ambiguous. So let me try and make it clearer. What I meant was that if you consider the two ...
H.S. ... Here, perhaps, is a core point of difference between us. I don't think that a "computing space" has any execution semantics at all. I think the thing...
... Fair enough, but you still haven't addressed my earlier point about how you verify inherent concurrency in the model. Are you familiar with the Dining ...
Sri ... Yes (at least as it is described in Tony Hoare's book on CSP). ... Do you really mean inherent in the *domain model*, or in the *domain*? For instance,...
Hello, How does products get permission to display the MDA stamp on their publication ? I have several books on MDA. Is there a public spec on MDA (beside the...
There is no formal MDA spec as such, and as AFAIK, you only have to agree to OMG's marketing/trademark usage guidelines. Basically, this means that you can...
Hi, Sorry, AFAIK=? OK, if OMG marketing department is stating a kind of product compliance, on what elements are they accepting the stamp usage ? Is there any...
Message text written by INTERNET:mda-discussion@yahoogroups.com ... Responding to Oldfield... ... Yes; the tests (such as we had) were run regularly. Owing to...
Responding to McNeile... ... You're right, it is a major disconnect. None of common the 3GLs, including all the OOPLs, support either flavor of concurrency;...
Responding to Nedunuri... ... I like this example vis a vis both my discussions with McNeile (asynchronicity vs. concurrency and whether eUML is suitable for...
Responding to Oldfield... ... I have to disagree here strongly for any significant application. Most applications require at least three partitions: the...
(responding to H.S.) ... I think we misunderstood each other to some degree; I had not been considering partitioning into different software components - but...
Responding to Oldfield... ... I'm not talking about things like distributed boundaries (though the USER/CRUD layered models usually have that in mind rather...
H.S. ... True! Thanks for the discussion. Your questions and comments have certainly helped me to clarify my thinking! :-) Best regards Ashley [Non-text...
... Thanks to both of you. It has been a great discussion and great source of thinking material. Cheers, Anthony Richardson [Non-text portions of this message...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
