Re: [boost] [Boost.Msm] On the necessity of state-boundary-crossingtransitions
Hi Andreas, I am not sure what is the point of such a challenge considering that I already explained that entry and exit pseudo states are not supported yet but will probably be in the future. Their absence is not by design as you seem to think but simply a decision of when to bring a version 1.0 and what should be inside. Anyway, as you offered me this challenge, I offer a solution which you will find in the vault ( http://www.boostpro.com/vault/index.php?direction=0&order=&directory=Msm&). I think this solution is a good example of what you can do with Msm and will hopefully make it clear how sub machines can make life easier. I want to thank you for the chance you give me to improve Msm (a few bugs were corrected on the way) and for your reaction to my opinion, which made it clear I need to more clearly draw a line between what I personally think and what users are used to and expect to find in Msm. Again, although I think no good of entry and exit pseudo states and consider them more than doubtful practice (please have a look at the doc for an explanation), a future version of Msm will offer them in some form with corresponding warnings. Regards, Christophe
Hi Christophe
I am not sure what is the point of such a challenge considering that I already explained that entry and exit pseudo states are not supported yet but will probably be in the future.
To me, it wasn't all that clear whether you are committed to provide such states, but I guess I've misinterpreted things.
Their absence is not by design as you seem to think but simply a decision of when to bring a version 1.0 and what should be inside.
Excellent, that's good to hear!
Anyway, as you offered me this challenge, I offer a solution which you will find in the vault ( http://www.boostpro.com/vault/index.php?direction=0&order=&directory=Msm&). I think this solution is a good example of what you can do with Msm and will hopefully make it clear how sub machines can make life easier.
I guess we agree that, strictly speaking, the challenge remains unsolved (your FSM has a different observable behavior). However, I do agree that using orthogonal regions for the alarms is superior to Harel's original. Anyway, I'll shut up now, as entry & exit points will definitely address my current concerns :-). Kudos, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.
participants (2)
-
Andreas Huber -
Christophe Henry