Exception transitions move away when click on model - aqualogic.bpm.modeling(Archived)

Hi everyone,
I'm a bit annoyed about this. I have my project saved, and everytime I open the process so that Aqualogic displays the process' model, the exception transitions I have located at the bottom-right of my model, suddenly change place to the upper-left spot of the model.
It works perfectly, It's just a display problem, does anyone know why Aqualogic does this to me? ^_^ any ideas?
Thanks in advance


Viewing ADF components

plz help me with this issue.i am new to JSF.
I created a new project and created a JSF JSP page.
when i drag and drop the adf rich client components its all are coming in the proper way.By proper way i mean buttons in the shape of button, input text in its shape.all components are viewed properly.so that i can get an idea of the page in the design mode itself.
but if i take its source and then come back to the design mode or if i close the project and then open it again, all the components are there but they are not the proper shape.all the components are seen as just a box with its text value and the component name..
But in the output its all coming in a proper way.
the problem is that i cant view my page properly in my design view.so i have to run each time to get the correct appearance.
i don't know how to add the screen shot of my screen in the design mode in this post to give an idea what i actually meant.
if anyone understood my problem plz help me with this.this would be a great help.
plz help me with this. 
I know exactly what you mean, I've had the same issue, but the thing is I don't have any idea of what might happen. In my computer it happens occasionally, and it comes back to its correct look by itself or re-starting JDeveloper.
Are you using "windows vista" or "xp"? If you have many "visual" programs running at a time, maybe you don't have enough resources to "cover" them all.
You have this problem every time you switch views or just sometimes? How much RAM do you have? It happened to me more when I had 1Gb than now that I have 2Gb, it can be this either (Jdeveloper needs a lot)
Sorry I can't help more... I would try to open the program again or to close the others.
Hope you find a solution,
this effect happens usually if e.g. the faces-config.xml file has issues, or a dependency (e.g. a bean) referenced from the UI cannot be instantiated in designtime. Often bouncing JDeveloper solves the issue.
Thank you for the replies.
i using a windows NT system and i have only 512MB RAM.
is it due to this???

Sharing page flow scope between task flows

I have a use case where I would need two task flow to share the same task flow scope, is there any way to achieve that? I COULD work around that by passing a managed bean instance by reference to a called task flow, but such architecture contribute to break the back button as it prevent the page flow scope from correctly restoring itself from the scope token since the whole scope is not serialized between requests.
If anyone knows a different work around it would be most welcomed though.
~ Simon 
the way that taskflows are designed is that each of them has its own scope that is not shared. The idea is to pass information from the calling taskflow to a called taskflow, which I think is a sensible way of handling shared information.
... just wondering what the usecase is
Hi Frank,
The use case is quite complex, but here it is. We want to create an application back button, but that can go back only to some specific pages, most likely some task flow entry points. Note that the task flows use page fragments and are used in a dynamic region. The idea I had to do that was to use the following:
- Have the "container" task flow define a bean that would manage the navigation history.
- The history would be a stack of history entries that would be a serialization of the region's main managed bean's state and/or task flow parameters as well as the task flow id.
- When the user enter a task flow, use a switcher to determine if a bean state was passed as a parameter. If so, initialize the bean with the state, else initialize it from parameters.
- Push the bean state/parameter on the history stack. This action implies that the task flow has access to the container bean. I COULD pass by reference it as a task flow parameter, but doing so would potentially break the ability of the page flow scope to restore its state if the browser's back button is used.
- When the user click the application's back button, the parent region reset the dynamic region with the task flow id taken from the stack entry and push the bean state as a task flow parameter.
If you see any better way to achieve that, I would gladly hear it.
~ Simon 
Hi Frank,
Another way I see to achieve the same result would be to have a communication channel between a parent region and the contained region. One form such channel could take would be to expose region attributes in the RegionModel. Exposing the parameters themselves (as parameters or as attributes) would also be a good idea imho.
It would then be possible, using the RegionNavigationListener facility, to populate the navigation history stack correctly.
~ Simon 
You seem to have a very indepth understanding of TaskFlows...how did you achieve this with the current documentation. Is there other sources that I am missing?
The application I am working with has need of dynamic regions which we are accomplished with iframes. It seems like dynamic regions may work, but I am struggling through the documentation. Any help would be awesome. 
Hi Scott,
I got it from 5 sources mainly.
1. Apache Trinidad. It provides a decent chunk of the view layer behavior, the page flow scope among other things, event if it's a bit different from the one in 11g.
2. Documentation. It not perfect, but it contains a decent amount of information about task flows.
3. A LOT of decompilation. I decompile at least one of the following file once per day to figure out how something works: adf-controller.jar, adf-controller-api.jar, adf-controller-rt-common.jar, adfdt_common.jar, adf-pageflow-dtrt.jar, adf-pageflow-fwk.jar, adf-pageflow-impl.jar, adf-pageflow-rc.jar, adf-richclient-api-1.jar, adf-richclient-impl-11.jar, custComps.jar.
4. Blogs. Many people from Oracle produce useful blog entries every now and then, so Google can come to your help for that when a specific question arises.
5. Trial and error. I work on a quite big project that started a while back, so the team had to try quite a lot.
Working with a bleeding edge product come with its cost, and a lot of trials (and mistakes) are part of it.
More specifically for task flows, I suggest decompiling adf-controller.jar, adf-controller-api.jar and adf-pageflow-impl.jar, but not everyones are comfortable at using source code as documentation so it may or not be an option for you.
As for the documentation, it's not that bad. You should be using the Fusion Developer Guide (FusionDevGuide.pdf), is that the one you're using?
Sorry for not being more helpful...
~ Simon 
Thanks...I find that we have to decompile to find things out also. I was hoping there was a "magic" bullet out there.
If Oracle is really going to stick to the open-source road it would be much easier if they just gave the source out with the TP release...this would significantly cut the debugging time and help to better explain the issues back to Oracle.
A great example is EclipseLink...since I can get the source code, issue resolution time has dropped dramatically...ah wishful thinking. 
sorry, just got back to this thread today ;-(
one way of improving the documentation is to provide us with a list of topics that you guys like to be covered based on your business requirements.
The taskflow documentation is pretty good, but experience shows that the matrix of real life usecases goes beyond of what a doc writer can imagine to the time he or she drafts the docs. So whatever is missing should be provided in how-to, blog or documentation extensions.
.. just forgot ..
I'll keep a boomark of this thread. Though we have contextual events for inter region communication, in many cases this implementation seems to be too heavyweight.
In a hack, I got inter region communication working using the extended render kit service and JavaScript. While this works - I call it "contextual events light" - this shouldn't be the official answer to inter region communication. So I think its worth revisiting this topic based on usecases
Thanks Frank!
~ Simon

Transition vs. Connector

Hi all,
I'm looking for a clear definition of a BPM Connector vs. a Transition and under what circumstances each would be used. I can't seem to find/understand a clear role for each.
Thanks everyone!!
Effectively they are the same thing. The transition describes the direction of flow in the process, from one activity to the next. A connector is a 'jump' or 'shortcut' to a certain activity. The main purpose of a connector would be to help keep your diagram neat.
As a best practice, it is recommended to keep unconditional transitions horizontal and conditional transitions vertical. This then creates a horizontal 'happy path' that is very easy to follow. Another best practice is to not 'cross' lines. This creates a very 'sloppy' and unorganized diagram. To do this however, you may need to use connectors.
In some cases, you will however need to 'jump' back to a certain activity, or from one of the first activities jump to the end. This is where a connector would come in handy. It removes the need for that monster transition crossing the whole diagram. Beware, however, that too many of these can lead to some confusion, because the diagram loses its connect the dot flow, so use them sparingly.
Hope this helps!
Thanks Kevin. This helps quite a bit!! I thought they were pretty much the same, but you drew the distinction perfectly!!

OLM - Outline panel

In OLM, it happens very often that I need to assemble different learning objects into topic groups which are then displayed in an outline structure. However, I noticed that the outline structure always appear collapsed, either on the course page or when the player is launched, on its left side. The problem is that users are so focused on what has been launched by the player that they do not even notice 'Select to expand" button where they need to click in order to see the whole structure. Is there a way of setting it automatically in OLM so that the outline shows the learning object structure always expanded?
Hi Jasmina,
This is the way OLM works today, and you cannot change that. We already have an Enhancement Request for what you are asking for. Please contanct Oracle Support if you want your organization's name to be added to the Enhancement Request to add more weight to it.
Regards Anders 
Hi Anders,
Thank you for your reply'

How to add parameters to human task implementation in BPM process?

I tried the BPM process tutorial Tutorial
and at step "Implementing the User Task" 1. ("Click on the hello data object and drag it into the parameters panel") I encounter a problem.
I cannot drag anything onto the parameter pane. I can send you a screenshot if that helps. The data object window also does not have the buttons it should have according to the image in the tutorial.
I am using JDeveloper (the tutorial uses I don't know if thats a bug or if I missed something. I followed the tutorial step by step starting from "Creating the Process Model".
thanks in advance
I still need help with this one, any suggestions? 
Ok I finally figured it out by myself.
This is a bug that sometimes occurs, several re-installs solved the problem.
(However in the process also another bug appeared but in the end it installed without any of these two bugs)