Move Zest to a new home

GEF, Nebula and Mylar developers,


After a few months of work, the Zest API [1] is much better
and much 
more stable. Zest is a set of Widgets and Viewers build on
Draw2D. The 
widgets are currently limited to a Directed Graph Widget,
although a 
graphical tree widget and nested graph widget are in early
stages, and a 
sequence chart viewer (which may be released) is also
currently being 
developed.

[1] http://www.ecli
pse.org/mylar/zest.php

Zest was originally developed to provide visualization
support for Mylar 
task contexts, although the generic nature of these views
has made them 
applicable to many other domains. Currently the graph view
is being used 
for plug-in dependency visualizations [2]. A few months ago,
Zest was 
approved as component of Nebula project, although I wanted
to solidify 
the API by providing both an SWT and JFace style interface.
I have now 
completed this. As well, I have created several code
snippets to provide 
examples of how to make use of this technology [3].

[2] 
http://www.eclipse.org/mylar/images/zest_screen1.JPG
[3] http://www.eclipse.org/mylar/sandbox/zest/zest_snippe
ts.php

However, at EclipseCon I was approached by the GEF team to
readdress the 
possibility of Zest moving as a subcomponent of GEF. This
has left me a 
little conflicted as I don’t want to step on any toes. The
Zest widgets 
act like widgets and I have tried to follow both the SWT and
JFace 
guidelines (although there is some work still to do here).
On the other 
hand, Zest depends on Draw2D, and therefore will never
really graduate 
into SWT. Finally, Zest provides a set of Graph Layouts
which are really 
only appropriate for GEF / Draw2D based viewers and
editors.

I think everyone agrees that while Mylar gave Zest a great
home, it is 
time to make the viewers more generally available. I am
happy to move 
Zest to wherever the community feels is the best place.

I have created a bug to track this issue [4]. I have
e-mailed this out 
to a number of lists, and I’m sure this does not affect
everyone, so if 
you have comments please respond on the bug report.

[4] 
https://bugs.eclipse.org/bugs/show_bug.cgi?id=182621

Best Regards,
Ian

_______________________________________________
nebula-dev mailing list
nebula-deveclipse.org
h
ttps://dev.eclipse.org/mailman/listinfo/nebula-dev

Ian,

Thanks for allowing the community to comment on this
decision.

As a UI designer (and Nebula user), my primary concern would
be to keep the
dependencies of Zest as light as possible, no matter what
project officially
hosts the components.  Currently, GEF requires a full RCP
environment with
its associated complexity, while Draw2D/JFace/SWT can be
used with only a
few jar dependencies.  This makes adoption much easier for
small, focused
projects and for limited use embedded in Swing applications.
 To put this
another way, I think it would be best to keep Zest separate
from the full
GEF distribution, so that Zest does not inherit the full
dependency burden
of GEF.

Otherwise, I can see the logic in placing Zest as a part of
the GEF project.
There is a strong conceptual distinction between native or
native-like
controls vs. controls that operate on a large custom canvas.
 It seems more
likely that a designer looking for Zest functionality would
start with GEF
and Draw2D than JFace and Nebula.  Although the API may
resemble JFace, the
functionality and core purpose resembles GEF.

No matter what you decide, thanks for contributing!  Working
in the Eclipse
community is exciting because of the new creative options
that these
projects make possible.

--
Peter


_______________________________________________
nebula-dev mailing list
nebula-deveclipse.org
h
ttps://dev.eclipse.org/mailman/listinfo/nebula-dev