Re: jython interface

Just a few general and historical comments:

m huang wrote:
> 
> I have a wrapper class prototype that can be called in Jython like
> 
> >>> plotVisAD2D([1,2,3],[[4,5,6],[7,8,9]],lineType=2)
> 
> it allows the X arrays to be more than a simple integer set,
> and multiple y arrays with settable individual looks. More
> importantly it is a package also for Java programmers who want to
> to imbed VisAD for plotting but don't want to understand all
> the underlying mechanism. I hope some part of the visad.Python
> package that provide casual-user-friendly interfaces can go
> to the visad core, though technically it is possible to use
> them at where they are now.

We'll keep it in the visad.python package. The core visad
package is already pretty large. And the visad.python package
is the focus of our easy-to-use efforts.

> > The real issue for me is that I think scientists should be using
> the
> > VisAD Data model to represent their data, because of the rich meta
> > data that goes along with it, and not just simple array of numbers.
> 
> The user group will fight teeth and nail against that statement :-)
> I know because we have people from the targeted users sitting
> in some of our IA meetings. They make it clear that they just
> want to get the number processed and plotted with minimum
> has-to-learn-another-system.

I appreciate your situation. But a little historical
perspective. When we were creating visualizations of
scientists' data back in the 1970's and 80's, we'd often
just plot an image. Then the scientists would ask for a
map overlay so they could see the geographical context,
and we'd ask "Where is the information describing the
map projection of your data?" The answer too aften was
something like "Written on a piece of paper in my top
left desk drawer."

We see something like this with our Matlab and IDL users,
who have the convenience of just plotting arrays of
numbers. But when they want to overlay data from multiple
sources, then they have to manage the spatial and temporal
registration themselves.

As you and Tom are demonstrating, it is quite easy for
VisAD to transparently transform arrays of numbers into
VisAD data objects with default RealType names and null
Units and CoordinateSystems.

If that's what they want, its easy to give it to them.
When the time comes that they need metadata, then they
can learn how VisAD can make that easier than rolling
their own.

> If I don't agree, they would just
> walk away from me. (This is not to say they can't be educated.
> But if tutorials and javaDocs are having low priorities,
> think who has the time to educate general users?)
> 
> Simple array of numbers is what every scientist who has
> learned programming can understand.
> 
> > the idea of automatically pasting successful commands into
> > the script is quite appealing [in Jython Editor / console].
> 
> what I wish is a a console like above plus a Jython IDE. One
> can run commands, edit scripts, and do all the debugging things
> like setting break-point and watches. But it will perhaps
> remain as a wish for a while because of lack of
> time/funding/users' voice level asking for such a thing.

The C versions of VisAD (1.0 and 1.1) actually did include
an IDE for its custom scripting language that supported
editting, break points, single stepping, and of course
interactive visualization of any data object. No one wanted
it. So now for Python all we support is an editor
(visad.python.JPythonEditor) with a simple execution
environment. Perhaps someone else in the Jython community
will develop a more advanced IDE for Jython that we can
use with VisAD.

Cheers,
Bill


  • 2002 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the visad archives: