Comments on: What can UML do for you? http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/ A company obsessed with one single goal: stopping people from writing so much code Thu, 09 Sep 2010 10:59:01 +0000 http://wordpress.org/?v=2.5.1 By: rafael.chaves http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/#comment-557 rafael.chaves Fri, 07 Nov 2008 16:33:52 +0000 http://abstratt.com/blog/?p=123#comment-557 The specification itself does not refer to it as 'action semantics'. It is basically the chapter on Actions. The Superstructure has it since 2.0 (in 2.1.1, it is in chapter 11). In 1.5, it was the entire part 5. The specification itself does not refer to it as ‘action semantics’.

It is basically the chapter on Actions. The Superstructure has it since 2.0 (in 2.1.1, it is in chapter 11). In 1.5, it was the entire part 5.

]]> By: bnz http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/#comment-556 bnz Fri, 07 Nov 2008 11:06:53 +0000 http://abstratt.com/blog/?p=123#comment-556 Huh. Interesting. Never heard of "action semantics" in the context of UML. Could you please provide me with a pointer in what OMG standard i have to look this up? A quick search through the superstructure and infrastructure standard did help much. I would be very glad to read a little more about that. Huh. Interesting. Never heard of “action semantics” in the context of UML. Could you please provide me with a pointer in what OMG standard i have to look this up? A quick search through the superstructure and infrastructure standard did help much. I would be very glad to read a little more about that.

]]> By: Ed Merks http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/#comment-554 Ed Merks Mon, 03 Nov 2008 16:50:47 +0000 http://abstratt.com/blog/?p=123#comment-554 Rafael, EMF has EOperations but no mechanism for declaring their behavior nor for invoking them reflectively, i.e., EOperations just declare the signature. I've been prototyping designs for https://bugs.eclipse.org/bugs/show_bug.cgi?id=216701 and thinking about similar mechanisms to describe the behavior of an operation along with the ability to reflectively invoke an EOperation. We do already support specifying the text of the Java method body, but I'd like something that would work for dynamic models. So yes, I'm imagining indeed using unparsed text, but to me that's "just" an XMI serialization story. The runtime story would be that expression languages are pluggable. I think you're exploring an area that's becoming increasingly interesting and relevant. It just seems that building a runtime model for UML is a pretty challenging task in and of itself and that this is a prerequisite for the rest of the story. While adding something to an Ecore base could produce useful and used results every quickly... Rafael, EMF has EOperations but no mechanism for declaring their behavior nor for invoking them reflectively, i.e., EOperations just declare the signature. I’ve been prototyping designs for https://bugs.eclipse.org/bugs/show_bug.cgi?id=216701 and thinking about similar mechanisms to describe the behavior of an operation along with the ability to reflectively invoke an EOperation. We do already support specifying the text of the Java method body, but I’d like something that would work for dynamic models. So yes, I’m imagining indeed using unparsed text, but to me that’s “just” an XMI serialization story. The runtime story would be that expression languages are pluggable.

I think you’re exploring an area that’s becoming increasingly interesting and relevant. It just seems that building a runtime model for UML is a pretty challenging task in and of itself and that this is a prerequisite for the rest of the story. While adding something to an Ecore base could produce useful and used results every quickly…

]]> By: rafael.chaves http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/#comment-553 rafael.chaves Mon, 03 Nov 2008 15:51:32 +0000 http://abstratt.com/blog/?p=123#comment-553 Ed, correct me if I'm wrong, but EMF's runtime capabilities are limited to structural features (data), no support for behavior. That sort of thing cannot be done at the tooling level, the metamodel itself has to support it. That is bound to lead to a heavy metamodel, which is the opposite of EMOF/Ecore's reason-d'être. Unless you go with unparsed text, but that is more or less the same as ignoring the issue altogether and pushing it to the application. Ed, correct me if I’m wrong, but EMF’s runtime capabilities are limited to structural features (data), no support for behavior. That sort of thing cannot be done at the tooling level, the metamodel itself has to support it. That is bound to lead to a heavy metamodel, which is the opposite of EMOF/Ecore’s reason-d’être. Unless you go with unparsed text, but that is more or less the same as ignoring the issue altogether and pushing it to the application.

]]> By: Ed Merks http://abstratt.com/blog/2008/11/02/what-can-uml-do-for-you/#comment-552 Ed Merks Mon, 03 Nov 2008 14:36:12 +0000 http://abstratt.com/blog/?p=123#comment-552 If only we had this for Ecore instead. Given that we already have an efficient Ecore-based runtime, it seems more immediately applicable... Goodness knows Emfatic needs some love and attention... If only we had this for Ecore instead. Given that we already have an efficient Ecore-based runtime, it seems more immediately applicable… Goodness knows Emfatic needs some love and attention…

]]>