Another thing: I am not an expert on state machines. You might want to 
post your state machine questions separately.

- Sascha

On 08/17/2010 04:33 PM, Sascha Zelzer wrote:
> Hi,
>
> some tough questions here... look below for some comments.
>
> - Sascha
>
> On 08/12/2010 05:30 PM, Wagner Sales wrote:
>    
>> Hi all,
>>
>> I'm thinking in start an application migration to MITK. I've read
>> about functionalities, did a cross-relationship requirements matrix,
>> tested this on ExtApp and I really see MITK a very pretty choice to do
>> this job.
>>
>>      
> Nice to hear :-)
>    
>> I've read the some documentation about BlueBerry, State Machine ( i've
>> installed the State Machine editor on my linux/eclipse box with
>> success ) and now I want to start my job but I'm a little bit
>> confused. Let me to explain my problem and questions:
>> 1) I'm developing an application that's are touchscreen based. Since
>> this, I have my own window management way, my controls are very
>> stylized to touch GUI ( by CSS on Qt, by theming and in some (hard)
>> cases by a paintEvent method reimplementation ).
>>       a) I want to reuse the registration, segmentation and may be other
>> modules presented in MITK. But I think this relies in BlueBlerry, and
>> if I use BlueBerry I think I'll not have the same level of freedom to
>> stylize my application in "fancy" ways. Are some tip here? For
>> example: "Use BlueBerry and reimplements some parts in this way, or
>> stylize in other way" or "Don't use BlueBerry and create your own
>> interface. See this example...". Other controls that's are present in
>> MITK like the window/level chooser will be probably reimplemented by
>> specialization in some way, to change appearance and size.
>>
>>      
> BlueBerry provides a plug-in mechanism and on top of that the
> "Workbench". This is the window which manages the "views" (i.e. the gui
> for the segmentation plug-in) and editors (i.e. the QmitkStdMultiWidget
> in the center). The workbench lays out these components and provides
> drag and drop capabilities.
>
> In your case, the layout mechanism (contained in org.blueberry.ui) is
> probably too restrictive for a touch based UI. Unfortunately, the views
> (like the Segmentation view) usually use a couple of Workbench services,
> hence you cannot isolate them from the workbench. You would have to
> write your own org.blueberry.ui plugin, which is a matter of months (if
> you know what you are doing).
>
> You could use all BlueBerry capabilities like services, extension
> points, preferences, etc. except for classes in the org.blueberry.ui
> plug-in. How to write applications completely without BlueBerry is
> covered in the tutorial: http://docs.mitk.org/nightly-qt4/TutorialPage.html
>
> Since the view classes usually contain only GUI code and call methods on
> classes in pure MITK, you could reuse the .ui file and implement your
> own segmentation view by looking at the original code and working around
> BlueBerry specific calls. Note the nicest option, I know...
>
>    
>> 2) Since my application are touch based, the interaction are very
>> different than mouse. I don't have a keyboard ( then, no way shortcuts
>> or use of CTRL, SHIFT, and so on ) and I don't have a mouse too, then
>> no wheel, third button and so on. If I need to interact in some way
>> that's are implemented in this way, I need to put a button to the user
>> to choose this interaction. For example, if the user haves to interact
>> with a slicer to changes the slice or to change the zoom, since aren't
>> a mouse, the user needs two buttons in the interface ( a dialog ) to
>> choose the desired interaction type.
>>       a) because that, are easy to think I have defined a lot of
>> "personalized" interactions. I've seen the state machine approach in
>> MITK and I really don't know if this way are the most suitable to me
>> in a way to define different type of interactions. Other way, off
>> course, are to specialize some classes to do this. Which can be a
>> better approach? In the first case I'll simple create and try the
>> interactions by using state machine definitions. But, in second case,
>> what's are the start point ( classes to start to read the code and
>> probably to change the interactions behaviour )?
>>
>> 3) I have a 3D mouse (from 3DConnexion, supporte by VTK 5.6)  and I'm
>> thinking in develop a virtual endoscopy module ( will be the last
>> module, and I think I can donate these to MITK ).
>>
>>      
> We would happily accept any donations!
>    
>>       a) another interaction issue. Here, I think I'll really need to
>> reimplement or create new classes to do this job and put this in a
>> plugin. Are some tips about a start point here?
>>
>>
>>      
> The 3DConnexion mouse was already supported by the Qt3 MITK version. We
> recently migrated the code to the new application infrastructure and it
> will be available soon (also without problems in a custom Qt4
> application not based on BlueBerry). So no work to be done here.
>    
>> Any help will be must appreciated,
>>
>> Best Regards,
>>
>> Wagner Sales
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by
>>
>> Make an app they can't live without
>> Enter the BlackBerry Developer Challenge
>> http://p.sf.net/sfu/RIM-dev2dev
>> _______________________________________________
>> mitk-users mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/mitk-users
>>
>>      
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> mitk-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/mitk-users
>    


------------------------------------------------------------------------------
This SF.net email is sponsored by 

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
mitk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mitk-users

Reply via email to