Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Eclipse Projects » GEF » GEF incubator proposal discussion
GEF incubator proposal discussion [message #514278] Sat, 13 February 2010 13:47 Go to next message
Jens von Pilgrim is currently offline Jens von PilgrimFriend
Messages: 304
Registered: July 2009
Senior Member
I've already mailed to the gef-dev mailing list (as an answer to
Alexander Nyssen's mail), but I'm afraid that was the wrong place (as I
got not feedback), so I "move" that discussion here, where it belongs,
according to the proposal found at

The proposal proposes to create an GEF incubator project in order
develop new features in that incubator and move thesm to GEF once they
are matured. In general I like that idea, and this incubator may also
be a place for moving (or copying) things to GEF which are currently
found in other projects (also see bug ).

Alexander Nyssen mentioned an ongoing discussion about the incubator
proposal in his mail to the gef-dev mailing list, however I haven't
found any "traces" of that discussion. (Maybe the participants of that
discussion could provide the "rest of us" with some information about
that here?) So I maybe discuss issues which were already discussed

In the proposal, some "initial work areas" are listed, and Alexander
also added some more details about these in his mail. These areas
- Advanced Editing Concepts
- Lightweight Notation Model

Actually, there are many projects using GEF. In terms of being GEF
related projects, they provide GEF functionality as well. Besides all
these concrete editor projects (mainly subprojects of MDT, such as UML
Tools, Ecore Tools, Papyrus), there are some frameworks (projects or
proposals) as well (this list may be incomplete):

- GMF,
- GEF3D,
- Graphiti,

I'm wondering whether some of the features mentioned in the proposal
aren't already implemented by some of these frameworks?

For example:

- notation model:
GMF is not only a generator framework, it is also a runtime framework,
which can be used without the generator. One thing GMF runtime provides
is a common notation model. What is the difference between a
"lighweight notation model" and the GMF notation model? Why is a new
notation model needed?

- advanced Editing Concepts - multi-layered nodes:
GEF3D provides something like this. In GEF3D, multiple 2D diagrams can
be combined in a single 3D scene, each diagram is projected onto a
plane. In order to calculate the coordinates of connections correctly,
the connections are to be assigned to a specific plane. So, the planes
are some kind of multi-layered nodes, i.e. they serve as container for
connections. In order to make this work with GEF and existing editors
(which are embedded into a GEF3d multi-editor), a so called
DispatchingConnectionLayer is used as an adapater between the
"single-connection-layer" pattern used in GEF and the multi-layer nodes.

- advanced Editing Concepts - "decent support for rotation/flipping of
nodes on the level of EditParts/EditPolicies and Figures is needed"
(from Alexander's mail):
We have implemented full support for rotation on figure level in GEF3D
(i.e. in Draw3D), not only for rotation around z-axis (what you's need
in 2D) but full support for any kind of transformation. Adding
handles/requests/policies to support interaction is missing, but Madhu
Samuel will work on that (see GEF3D developer mailing list). However,
with rotation you get all these coordinate system problems. Our
solution is described in
A more user centered description (about the effects) can be found at -me/

- one more thing implemented in GEF3D which maybe is useful for 2D
editors as well is an OpenGL based implementation of the Graphics
interface (see ) -- would that be an
incubator candidate as well?

It would be great to learn more about the features mentioned in the
proposal. My list above certainly is a little bit GEF3D centered. Maybe
committers or users of other projects could provide information about
what features their projects provide interesting for others as well...

Frankly, I'm a little bit confused by these new graphical editing
related project proposals (namely Graphiti and GEF incubator). In terms
of providing new features I'm worried to have multiple implementations
of the same things (E.g., there exist like four different
implementations of the command pattern (in GEF, EMF, GMF, and Eclipse
core), maybe more. While there may be (good) reasons for having several
implementations, it is also confusing and it makes things more
complicated.). Besides, it would be nice if new features added in GEF
are compatible with existing solutions implemented and used in existing
projects (and editors).


Re: GEF incubator proposal discussion [message #514633 is a reply to message #514278] Tue, 16 February 2010 06:49 Go to previous message
Vijay RajFriend
Messages: 608
Registered: July 2009
Senior Member
Hi again Jens,

I Agree,
GEF incubator project would be the best way to merge all the features of GEF which were added in other projects....

Previous Topic:[Announce] GEF 3.5.2RC3 is available
Next Topic:[Zest] How to reveal a graph node
Goto Forum:

Current Time: Thu Mar 23 00:21:36 GMT 2017

Powered by FUDForum. Page generated in 0.02558 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software