On 10/07/2013 10:47 PM, Chuck Bridgham
wrote:
I know we do have
dependencies on PDE
from Data Tools, and OSGI development tools,
Ok,
The only bundle in datatools depending on PDE is
org.eclipse.datatools.connectivity.oda.template.ui which depends on
pde.ui. I'll check whether this is an absolutely necessary
dependency, as it introduce a lot of UI elements not related to the
general datatools use-case.
Also, does OSGi development tools help in this JEE package? I tend
to think that Java EE/Web/Mobile developers use the solutions
provided by framework, and AFAIK, none of the major ones relies on
OSGi. So it seems to me that OSGi Tools are not very relevant there.
I added a few cents to further discussion, but they are useless
until the two above points are not cleared.
but I would also be careful
before consulting the community / tools adopters that use this
package
as a starting point.
Makes sense.
Just beware of plugins developers who give their opinion on a
package that is not intended to them, and who want PDE in Kepler
package just to save them a few minutes every 4 monthes ;)
For JBoss Tools and JBoss Developer Studio in Eclipse, we recommend
using the Java EE package as a starting point, and it is certain
that our users are more annoyed/disturbed by these PDE UI entries
they won't use.
I agree with you the
majority of users
won't be developing plugins, but PDE is a must for the remaining
group.
They/we already have
a RCP/RAP package for this use-case.
It may be time to do some
analysis of
size, and popularity of the many features that make up this
package.
Not only popularity, we could try to evaluate
added_value_of_feature_X / cost_of_feature_X_in_UI , where
added_value_of_feature_X is the productivity benefit of feature X,
and cost_of_feature_X_in_UI is the additional cost of introdcing
such a tool in daily usage (useless UI elements, decrease of overall
performances, loss of stability...).
|