IBM Engineering Rhapsody Tip #99 - Making life easier with some proxy port automation

In this short silent video, I wanted to showcase a couple of helpers I wrote for a client. Basically, if you expect people to follow steps that are repeatable, then why not get the computer to do it? (Steve Jobs: I'd rather than a lazy engineer than a good one). I want my Rhapsody modeling to be like my iPhone, in that it just does stuff I expect without me having to think about it. In this video I show a couple of helpers that I did to speed up modeling for a client who wanted to use Proxy Ports. Essentially, I just roll the things that I would do manually into a profile plugin, and Bob's your uncle. The interesting thing about automation is that the more you do, the more ideas to speed things up pop out. Also, the market for SysML has matured a bit hence there are a few more new users, with less training and patience. I want to make it as easy as pie to model, so that they stick to the fun and creative stuff, and are able to demonstrate business value first time, rather than flop around like busy bee's unable to find a flower. 

Here's the transcript for those that need it:

In this silent video I thought I'd share a snippet of some automation I've been writing to make it easier for a client to model with Proxy Ports in Rhapsody.

The first bit here, is to speed up creation. Since, I know modeler is expected to use proxy ports, I detect the add new for part->part connector, and create proxy ports automatically. My plugin also creates the interface block that will define the contract. The 'opposite port' has the fiddly stuff done, e.g., the conjugated flag is set.

Now we can define what is exchanged via the InterfaceBlock. This is usually done by creating an interaction model, i.e., using sequence diagrams. This is an event (asynchronous signal) from the Operator to the system. I'm using a condition mark here to imply the state of the Block, i.e. I'm inferring states for the subsystems and, importantly, the signals sent between the subsystems via the proxy ports.

In "Analysis" mode, operations/events in sequence diagrams are not created automatically in the browser, meaning that it's easy to change event names. This is standard behavior in Rhapsody. There is a separate "Realization" step that you perform once the sequence is finalized. The final piece of "speed up" automation I wanted to show, is a helper that no-only realizes the events/operations, but also updates the ports and interfaces.

This is similar to a really useful helper provided by the Harmony/SE toolkit, but I've added a user interface. Still some work to do but the aim is to always explain what you're going to do, before doing it, as this helps with training people in what is going to happen.

As well as realizing the operations and adding events and event receptions to the Blocks, it has updated the Interface Block. As the rest of this video shows, I can go on to create statecharts and build an animating model of them.

Ultimately, with "speed up" helpers like the ones I've shown, we can accelerate your modeling process and engineers can focus on the creative and fun stuff.

In lieu of TrustPilot, some recent training feedback

I don't really have the volume of deliveries or business to live in the world of trust pilot. However, I do appreciate that it's good to have real feedback from real people.

As such, I thought I'd share some feedback from recent 3 day Rhapsody trainings in the last 2 months. In the following survey, 5 = strongly agree and 1 = strongly disagree.

What did you like most about the training?

• The labs really helped enforce the theory and brought the material full circle.
• It's a new tool for me that can capture big picture ideas and decompose to manageable chunks.
• Labs are great. Good examples.
• Instructor & the labs
• Flexibility to demonstrate tasks on demand, make it easy in this environment.
• Lots of tips and short cuts for action users 
• I had no prior knowledge regarding Rhapsody however leaving the training I feel like I'm able to perform task via Rhapsody. The trainer was also very knowledgeable and kept the participants engaged. 

This week, I'm moving on to Version 12 of my material. I've added a simplified DOORS NG with Rhapsody half-day workshop which can be optionally used to finish things off with a fun "from scratch" example. I've also added even more differentiation to the labs to give me more options based on whether participants are experts who use Rhapsody a lot, or new to SysML/Rhapsody, so that I don't scare people with too much information, if I can help it.

IBM Engineering Rhapsody Tip #98 - Debugging user Java plugins running in Rhapsody

When you write a lot of add-ons for IBM Rhapsody, it becomes very helpful to be able to debug them. This can be a little more difficult when they're running inside the Rhapsody processes, rather than as an external Java application. There are lots of reasons for running in process though, not least because they run an order of magnitude faster, and you can tie them to a profile that adds things like menu commands and properties to control it. In this video I show how it's possible to debug them as remote java applications enabling you to set breakpoints and step through the code in Eclipse, even though the code is running in Rhapsody. You first need to add some changes to your Rhapsody ini file though. The video illustrates. Hopefully, it helps someone out there. p.s. This is a real life example.

Here's the transcript for those that need it:

This video demonstrates how it's possible to debug user-written Java plugins running in the Rhapsody process using Eclipse. 

Java add-ons for Rhapsody can be written to run as applications outside of Rhapsody, or as plugins running in the Rhapsody process (as here). Plugins can be loaded by a profile. 

A certain amount of debugging can be done using instrumentation. It doesn't always work though. In this case, I'm expecting my plugin listener to detect the event and automatically move it into an 'interfaces' package.

To determine what is happening, I want to use the Eclipse debugger to set breakpoints and step through the plugin as it runs inside of Rhapsody.

It's possible to do this by first adding some JVM settings to the rhapsody.ini file. As Rhapsody reads/writes the .ini file, always create a backup and make sure Rhapsody is closed before editing.

The JVM section is where the Java Virtual Machine that Rhapsody will use is specified.

I've added the Debug1, Debug2 and Debug3 options to my .ini file before, hence I just need to include them in Options list for the JVM.

[JVM]
Options=ClassPath,LibPath,Debug1,Debug2,Debug3
Debug1=-Xnoagent
Debug2=-Xdebug
Debug3=-Xrunjdwp:transport=dt_socket,address=6743,server=y,suspend=y

Now when I run Rhapsody and an attempt is made to load a plugin into the Rhapsody JVM, it will wait for a remote debug connection.

To get the Eclipse debugger to connect we need to run a Remote Java Application configuration and change the port number for the Remote Java Application to 6743.

Voilà, the debugger connects with the waiting plugin running in the Rhapsody JVM.

It's now possible to run the application as normal, but set breakpoints in the plugin code. E.g., I'll set a breakpoint in the Java plugin.

That's strange...

To reload the plugin code, I need to exit Rhapsody not just close the project.

Doh!! (an unnecessary space, but I can fix it)

Hope that helps. Interested in MBSE/UML or Rhapsody training or customizations then email me.

IBM Engineering Rhapsody Tip #97 - I can't get no Satisfactions (creating a browser view)

I'm keeping towards the magic 100! This IBM Rational/Engineering Rhapsody tips and tricks video shows the simple steps how to create a new browser filter that shows whether requirements have certain coverage in a model. I show that it works with both requirements in the model and remote requirements in IBM Engineering DOORS Next Jazz application. It's silent though hence if you fancy strutting your stuff then feel free.

The following is the transcript for those that need it:

The theme for this Rhapsody tips and tricks video is the Stones' "I can't get no Satisfaction". It's silent though, hence you have to sing it yourself. 

Rhapsody has a bunch of standard browser filters, e.g., filter browser to Requirements View. But look, there are some non-standard views listed here! E.g., which requirements are covered, and which requirements can't get no satisfaction?

If we add a Satisfaction to the model and save, then the requirements are removed to show they're now covered. Adding such views can speed up requirement tracing.

These custom views were added by one of my profiles. 

Let's have a go at creating the "no satisfaction" one from scratch ;-) First off, I'll do it in a profile, a special type of UML package for customizations. 

I'll create a Query element. Queries are like persisted model searches. We can define a search criteria, such as Element type = Requirement. If we double-click a Query then it runs a search in the model.

We can build a compound filter, e.g. Reference. Find elements that have exactly zero incoming relations of type Satisfaction. The compound filter is now looking for requirements that don't have satisfaction relations. Double-click enables us to test it.

What's more, the Model::Query::ShowInBrowserFilterList property is set. This means that our new Query - AintGotNoSatisfaction is listed as a browser.

Worth noting that these queries also work with remote requirements. These requirements are in IBM Engineering DOORS Next. With DOORS Next requirements, adding a standard dependency allows us to create an OSLC Satisfaction and the requirements has been removed from the filtered view of unsatisfied requirements, i.e., I can't get no satisfaction! ;-)

Interested in MBSE/UML or Rhapsody training/customizations? Email me.

IBM Engineering Rhapsody Tip #96 - The Good, The Bad and The Ugly of Label Mode

This video is a bit of a Rational Rhapsody aficionado's topic. The joy's of Rhapsody's Label Mode. Most effective use of this has been when presenting quite a complicated diagram, like an executable state machine or sequence diagram, to lay audience. I'm sure there are some other great examples of usage out there also ("The Good"). I do highlight some of the dangers, however. When you write things not knowing that label mode is enabled, for example.  The video is silent though, so you'll have to hum Ennio Morricone's fantastic spaghetti western theme song, if you get bored with reading my captions.

New IBM Engineering Lifecycle Management (ELM) User Group page

IBM have set a new User Group page for Engineering Lifecycle Management (aka Jazz Continuous Engineering platform). They're saying this will be the central point of all IBM ELM activities, hence you may want to check it out:

https://ibm.biz/elmusergroup

There will be features allowing IBM to communicate to the user groups and register to events.

It is open to the public. IBM encourages partners and clients to join. 

Academia is also welcome.

IBM Engineering Rhapsody Tip #95 - Why diagram ownership is important for Internal Block Diagrams

This video covers an area where new users of SysML often struggle, i.e. understanding the differences between Blocks and Parts, and the role played by the diagram frame of an internal block diagram. Fundamentally, in SysML the frame of the ibd is normally a block and the ibd represents an assembly of parts owned by that block. Within my old friend Rhapsody, unlike Cameo, there's not a context selection method on a diagram that defines the context of its frame. The context of the diagram is determined rather by the ownership of the diagram. This means that diagram ownership is critical. This is why I normally recommend that new users are careful to create internal block diagrams as children of blocks rather than at package level. Related to this is that if you drag an element on to an internal block diagram and it is out of context, then it will be displayed with a dashed line, something new users often struggle to understand. This is because it's not owned by the element that the frame represents. This is video #95 of my Rhapsody tips and tricks video channel that I started 6 years ago.

Here's the transcript:

Hi, in this short video I'm going to show why it's important to have Internal Block Diagrams owned by Blocks rather than being children of packages.

So, I'll start by creating a quick SysML project and I'll create a classic Block Definition Diagram with a structure. I'll create a System A which has a Subsystem B and Subsystem A. The way I create a structure is with Directed Compositions. I'll create a part A and a part B. 

By default, Rhapsody will name the part with the word "its" followed by the name of the Block but I can name it what I like. 

So what we have here is essentially an internal structure with parts. The block definition diagram defines the relationship between blocks but blocks are types of things and parts are named usages of those types, usually within the context of another block.

... although Rhapsody does allow you to also have parts which are global, i.e. owned by a package.

If I create an internal block diagram where I want to show these parts, then normally the frame of that diagram would represent the element, i.e. Block, that owns those parts. So, if I create an internal block diagram at package level, then we can see in the frame that this is a package diagram of a package called default.

I can put the parts in it and you can see that the dotted notation will tell me where in its nested structure it is.

Of course, I could put the part into the System A block here and that removes the dotted notation.

But the frame here, doesn't represent system A. It doesn't represent anything that's a classifier. Let me just add a port on the boundary of System A here, and another port with a connector.

Now, if I create an internal block diagram underneath the System A block then the frame tells us that this is an internal block diagram of a block called system A.

That means that the frame represents the block System A. This means that you can show a port of System A on the frame, and the things inside that frame then would normally represent parts that are owned by System A.

So the Block Definition Diagram is a diagram that shows relationships between Blocks and blocks are types of things.

The internal block diagram is a diagram that shows me the relationships between parts which are usages and the frame represents the block that owns these parts.

Let's see what happens if I have a nested hierarchy that goes deeper. Maybe component C is a child part typed by  Subsystem A and component D has a child part typed by subsystem B.

If I put part D here on the internal block diagram, then you can see that it's dashed because part D is not directly owned by subsystem A, rather part D is owned by subsystem B.

If I move it into the correct context then it's boundary is no longer dashed.

So, if I drag something on to an internal block diagram, for example a Block, then it's dashed because it's not owned by the element that the frame of the ibd is representing.

Internal block diagrams are diagrams of the internal structure of a block, so things that are not in that internal structure are going to have dashed lines on them.

There's two fundamental things here really. One is that blocks are types of things and that parts are named usages of blocks, usually within the context of another block. 

Secondly, an internal block diagram is a diagram of those usages and how they're connected within the context of a a block, and that the frame of the ibd normally represents that owning block.

In Rhapsody, there's no context setting on the internal block diagram. The internal block diagram frame is determined by the ownership of the diagram. So the internal block diagram here is owned by system A which means that its frame will represent system A. This is why I always create an internal block diagram underneath a block, rather than at package level (unless you have parts which are global - which I tend to avoid anyway).

I hope that helps. Don't forget that I have a business that can deliver customised training in Rhapsody that suits your needs and your budget, so feel free to email me if you want more information on ways of accelerating your adoption of successful model-based systems engineering with Rhapsody.