Issue 5724: use of multiple stereotype (uml-scheduling-ftf) Source: International Business Machines (Dr. Bruce Powel Douglass, bruce.douglass(at)us.ibm.com) Nature: Uncategorized Issue Severity: Summary: I find the use of multiple stereotypes just to get the tags required to do a simple global RMA analysis, plus all the indirection necessary. ugly, ugly, ugly. A thought occurred to me - why not highly-specific analysis subprofiles, where we would define a set of stereotypes like <<RMAG_Task>> which would be <<CRConcurrnent>>, <<SASchedulable>> and <<SAAction>> all bundled together? (RMA Global is where I got RMAG). Thus we could stereotype elements such as active objects and resources very simply. A small set of internally consistent tags with the right properties that map specifically to the kind of analysis to be perform would make the profile much easier to use. Resolution: Closed, no change Revised Text: Actions taken: October 24, 2002: received issue June 30, 2003: closed issue Discussion: End of Annotations:===== X-Sender: bpd@flash.ilogix.com X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Thu, 24 Oct 2002 14:30:40 -0500 To: uml-scheduling-ftf@omg.org From: Bruce Douglass Subject: Thinking more on this... I find the use of multiple stereotypes just to get the tags required to do a simple global RMA analysis, plus all the indirection necessary. ugly, ugly, ugly. A thought occurred to me - why not highly-specific analysis subprofiles, where we would define a set of stereotypes like <> which would be <>, <> and <> all bundled together? (RMA Global is where I got RMAG). Thus we could stereotype elements such as active objects and resources very simply. A small set of internally consistent tags with the right properties that map specifically to the kind of analysis to be perform would make the profile much easier to use. Opinions? From: "Moore, Alan" To: "'Bruce Douglass'" , uml-scheduling-ftf@omg.org Subject: RE: Thinking more on this... Date: Fri, 25 Oct 2002 11:21:32 +0100 X-Mailer: Internet Mail Service (5.5.2653.19) > -----Original Message----- > From: Bruce Douglass [mailto:bpd@ilogix.com] > Sent: 24 October 2002 20:31 > To: uml-scheduling-ftf@omg.org > Subject: Thinking more on this... > > > I find the use of multiple stereotypes just to get the tags > required to do > a simple global RMA analysis, plus all the indirection what do you mean by indirection here? > necessary. ugly, > ugly, ugly. A thought occurred to me - why not > highly-specific analysis > subprofiles, where we would define a set of stereotypes like > <> > which would be <>, <> and > <> all > bundled together? (RMA Global is where I got RMAG). Thus we could > stereotype elements such as active objects and resources very > simply. A > small set of internally consistent tags with the right > properties that map > specifically to the kind of analysis to be perform would make > the profile > much easier to use. > The most obvious use of this would be to support the use of specific products - I know that Tri-pacific use a subset of the SA profile in their tool integration. > Opinions? The problem would be to get people to agree on what such a subset might me - you raised a point in a previous e-mail to me about the lack of "priority" on schedulable resource - we could of course add it but depending on what you want to use a schedulable resource for, we might need to add all of the tags of SAresponse to SAschedulable. I don't see the problem with allocating multiple stereotypes to the same element - as long as the diagram views don't become too cluttered, which is a tool issue I think. > > Date: Mon, 28 Oct 2002 09:22:03 -0600 From: "Watson, Bennett C" Subject: RE: Thinking more on this... To: "'Moore, Alan'" , "'Bruce Douglass'" , uml-scheduling-ftf@omg.org X-Mailer: Internet Mail Service (5.5.2653.19) Based on my experience with the Tri-Pacific tool, my opinion is that the analysis tool only uses a specialized subset of the tagged values. That's not a proper subset ... there are tags added and certainly many tags that are not used. The approach we used was to stay with the standards as much as possible and add new tags to suit our unique needs. There are a lot of issues with how the UML tools handle stereotypes, constraints/notes, and display real-estate. It may be interesting to talk about "practical" use of the profile for schedulability analysis. I am also working with "another analysis product" on the behalf of LM Aero, and that may be another source of discussion. Mark, what do you think? Ben -----Original Message----- From: Moore, Alan [mailto:AlanM@artisansw.com] Sent: Friday, October 25, 2002 5:22 AM To: 'Bruce Douglass'; uml-scheduling-ftf@omg.org Subject: RE: Thinking more on this... > -----Original Message----- > From: Bruce Douglass [mailto:bpd@ilogix.com] > Sent: 24 October 2002 20:31 > To: uml-scheduling-ftf@omg.org > Subject: Thinking more on this... > > > I find the use of multiple stereotypes just to get the tags > required to do > a simple global RMA analysis, plus all the indirection what do you mean by indirection here? > necessary. ugly, > ugly, ugly. A thought occurred to me - why not > highly-specific analysis > subprofiles, where we would define a set of stereotypes like > <> > which would be <>, <> and > <> all > bundled together? (RMA Global is where I got RMAG). Thus we could > stereotype elements such as active objects and resources very > simply. A > small set of internally consistent tags with the right > properties that map > specifically to the kind of analysis to be perform would make > the profile > much easier to use. > The most obvious use of this would be to support the use of specific products - I know that Tri-pacific use a subset of the SA profile in their tool integration. > Opinions? The problem would be to get people to agree on what such a subset might me - you raised a point in a previous e-mail to me about the lack of "priority" on schedulable resource - we could of course add it but depending on what you want to use a schedulable resource for, we might need to add all of the tags of SAresponse to SAschedulable. I don't see the problem with allocating multiple stereotypes to the same element - as long as the diagram views don't become too cluttered, which is a tool issue I think. > > X-Sender: bpd@flash.ilogix.com X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Tue, 29 Oct 2002 15:54:58 -0600 To: , "'Watson, Bennett C'" , "'Moore, Alan'" , From: Bruce Douglass Subject: RE: Thinking more on this... EXACTLY At 09:02 AM 10/28/2002 -0800, Mark S Gerhardt wrote: Ben is on the same track in thought as I am. A practical subset would go a long way, as well as need things like statistical expression for period and computation as well ( for simulation.) The two tools that Ben referenced use a subset plus statistical specifications as extensions. ________________ Mark S. Gerhardt Principal Applications Architect TimeSys Corporation 1901 S. Bascom Avenue, Suite 220 Campbell, CA 95008 (650)208-3994 mark.gerhardt@timesys.com -----Original Message----- From: Watson, Bennett C [mailto:bennett.c.watson@lmco.com] Sent: Monday, October 28, 2002 7:22 AM To: 'Moore, Alan'; 'Bruce Douglass'; uml-scheduling-ftf@omg.org Subject: RE: Thinking more on this... Based on my experience with the Tri-Pacific tool, my opinion is that the analysis tool only uses a specialized subset of the tagged values. That's not a proper subset ... there are tags added and certainly many tags that are not used. The approach we used was to stay with the standards as much as possible and add new tags to suit our unique needs. There are a lot of issues with how the UML tools handle stereotypes, constraints/notes, and display real-estate. It may be interesting to talk about "practical" use of the profile for schedulability analysis. I am also working with "another analysis product" on the behalf of LM Aero, and that may be another source of discussion. Mark, what do you think? Ben -----Original Message----- From: Moore, Alan [mailto:AlanM@artisansw.com] Sent: Friday, October 25, 2002 5:22 AM To: 'Bruce Douglass'; uml-scheduling-ftf@omg.org Subject: RE: Thinking more on this... > -----Original Message----- > From: Bruce Douglass [mailto:bpd@ilogix.com] > Sent: 24 October 2002 20:31 > To: uml-scheduling-ftf@omg.org > Subject: Thinking more on this... > > > I find the use of multiple stereotypes just to get the tags > required to do > a simple global RMA analysis, plus all the indirection what do you mean by indirection here? > necessary. ugly, > ugly, ugly. A thought occurred to me - why not > highly-specific analysis > subprofiles, where we would define a set of stereotypes like > <> > which would be <>, <> and > <> all > bundled together? (RMA Global is where I got RMAG). Thus we could > stereotype elements such as active objects and resources very > simply. A > small set of internally consistent tags with the right > properties that map > specifically to the kind of analysis to be perform would make > the profile > much easier to use. > The most obvious use of this would be to support the use of specific products - I know that Tri-pacific use a subset of the SA profile in their tool integration. > Opinions? The problem would be to get people to agree on what such a subset might me - you raised a point in a previous e-mail to me about the lack of "priority" on schedulable resource - we could of course add it but depending on what you want to use a schedulable resource for, we might need to add all of the tags of SAresponse to SAschedulable. I don't see the problem with allocating multiple stereotypes to the same element - as long as the diagram views don't become too cluttered, which is a tool issue I think. > >