[kepler-dev] JAI actors

[Christopher Brooks]

Hi Nandita,

Having you work in the Ptolemy tree on the JAI actors is definitely
fine with me.  Please request a CVS read/write account, see
http://chess.eecs.berkeley.edu/ptexternal/#readwrite
When you are coding, please follow the Ptolemy coding style, see
$PTII/doc/coding/style.htm

With regard to the details of what you do in JAI:
Edward indicates that adding PortParameters can be done without 
compromising backward compatibility.  So, if you can avoid
creating backward compatibility issues, please do so.
That being said, I'm willing to put up with a certain about of
compatibility problems, especially if they are well thought out.
I don't think that many people are using the JAI actors.  Having
them suffer changes now is better than having more people suffer
later.  Note that we can use the MoML filter capability to
solve some compatibility issues such as name changes.

As you are working, please make sure the JAI demos work.
In the perfect world, the JAI demos would be unchanged.
However, if necessary, we could create filters that would
allow them to be updated automatically.  I can help here.

Be sure to give yourself and Kepler plenty of credit.
At the minimum, you should add yourself as a contributor to
any class you edit.  The @author tag might look like

@author James Yeh, Contributor: Nandita Mangal (SDSC/Kepler)

Feel free to change the (SDSC/Kepler) to something else.
Also, if you make substantive changes, you could add yourself
as a coauthor:
     
@author James Yeh, Nandita Mangal (SDSC/Kepler)

In this case, adding a PortParameter and otherwise cleaning up or
adding demos could be considered substantive.

Basically, I'd like you to be responsible for the JAI code, and the
author tag should reflect that.

About the double matrix affine/filter design choice, use your
judgement.  I'm willing to suffer backward compatibilty problems here
for a better design.


_Christopher


Edward wrote:
--------

    At 02:12 PM 12/1/2005 -0800, nmangal at sdsc.edu wrote:
    >However, most of the actors do not add the flexibility of putting such
    >actors in automated execution workflows with least level of
    >user-interactivity. Kepler Imaging actors keeping that flexibility in mind
    >have added PortParameters as inputs for image manipulations in various
    >actors. One can modify existing Ptolemy actors to add this flexibility or
    >make use of Kepler Actors depending on further discussion results.
    >On the other hand, some Unique Ptolemy actors can prove useful in future
    >scientific workflows. Image Scaler for example of Ptolemy gives a user
    >more options to modify the image as compared to ImageScaler of Kepler.
    
    Converting parameters to PortParameter is generally a good
    idea, I think. Moreover, this can be done without compromising
    backward compatibility (!).
    
    
    >Another feature distinct in Ptolemy actors, is that they are dependent on
    >specifying a double matrix for affine/filter operations. Kepler Actors
    >provide the user with the basic operations in affine /filter operations.
    >However improvement can be made by modifying either Kepler/Ptolemy actors
    >to incorporate both features of either using a commonly used operation or
    >custom operation in the same actor.
    
    I personally find the matrix specification rather hard to use...
    
    Edward
    
    
    
    ------------
    Edward A. Lee
    Professor, Chair of the EE Division, Associate Chair of EECS
    231 Cory Hall, UC Berkeley, Berkeley, CA 94720
    phone: 510-642-0253 or 510-642-0455, fax: 510-642-2845
    eal at eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal  
--------