Events
|
 |
|
|
Calendar
| APPLIGRAPH
Meeting |
Paderborn,
Germany |
September
5-6, 2000 |
| Graph
Drawing 2000 |
Colonial
Williamsburg, USA |
September
20-23, 2000 |
| CASCON
2000 |
Toronto,
Canada |
November
13-16, 2000 |
| WCRE
2000 |
Brisbane,
Australia |
November
23-25, 2000 |
| Dagstuhl
Seminar |
Dagstuhl,
Germany |
January
21-26, 2001 |
| APPLIGRAPH
Meeting |
Bremen,
Germany |
February
22-23, 2001
|
| ICSE
2001 |
Toronto,
Canada |
June
2-9, 2001 |
APPLIGRAPH
(Application of Graph Transformation) is a meeting of the graph transformation
community to discuss and exchange format for graphs. A brief
trip report is available below.
Graph
Drawing 2000 is the annual meeting for that community. The Steering
Committee is initiating an effort to define a common exchange format for
graph/network data. They held an informal kick-off meeting at GD2000.
A brief trip report is available below.
CASCON
is a Canadian software conference for researchers, industry, and government.
Registration is free and there will be two half-day workshops on GXL. CSER
(Consortium for Software Engineering Research) will meet November 11-12.
At WCRE,
there will be a panel on exchange formats.
The Dagstuhl seminar is
entitled "Interoperability
of Reengineering Tools" and attendance is by invitation only. The organizers
are Jürgen Ebert, Kostas Kontogiannis, and John Mylopoulos.
It may be possible to hold
another WoSEF at ICSE 2000,
but it will not be an official workshop since the deadline for submitting
proposals has passed.
Past Meeting Reports
APPLIGRAPH
Meeting, September 5-6, 2000
Andreas Winter and Andy Schürr
attended this meeting in Paderborn. There were 18 attendees from
11 universities. The following is excerpted from Andreas' trip report.
The workshop started with
a comparison of the proposed DTDs by Daniel Varro. He presented his understanding
of these approaches on a conceptual level using class diagrams. This enabled
us to discuss what to exchange on a conceptual level, independent of implementation
in XML-DTD. They identified three areas where they needed to achieve
agreement:
-
graphs
-
schemas (with a lot of associations
to the class diagram for graphs)
-
graph transformation systems
(which includes graphs and schemas)
During the meeting, they discussed
a number of different approaches to the different components. Some
participants were assigned "work packages" that are due on December 15,
2000. They planned to discuss the proposals and resuts electronically
in December to February and to have a further meeting at the end
of February. The work packages were:
| W1: UML Diagram to DTD |
Ralph Depke, Reiko Heckel |
| W2: UML Diagram to XML
Schema |
Daniel Varro |
| W3: XML Schema |
Daniel Varro |
| W4: UML Diagrams for graph
and graph schema |
Andy Schürr, Andreas
Winter |
| W5: UML Diagrams for control |
Albert Zündorf |
| W6: UML Diagrams for GTS |
Michael Matz, Gabi Taentzer |
GD2000
Kick-off Meeting, September 20, 2000
Susan Sim attended this meeting
in Colonial Williamsburg. The following information is excerpted
from my trip report. More information is available in the official
meeting minutes.
There were approximately
thirty people at the meeting. The meeting began with a couple of
short presentations. Stephen North (AT&T's dot) gave a
very brief introduction to XML and Scott Marshall (GraphXML) gave a very
brief survey of existing
exchange formats. We had a short break and then a requirements gathering
session. Ulrik
Brandes sat at the front
of the room and took down a list of issues that needed consideration.
This exercise actually went very well and a number good points were raised.
In general terms, they decided
to break down the problem of representing drawings of graphs into four
levels: structure, topology, geometry, and rendering. Structure is
simply the entities of the graph, such as nodes and edges. I'm not
really sure what is the difference between topology and geometry.
I asked a couple people afterwards and they weren't sure either.
I think topology is graph drawing characteristics, such as planar
and maximum path length, while geometry is layout. Rendering is the
final depiction step, e.g. exactly what does a dashed line look like?
They decided to attack the
structural level first, and called this "core level." They decide
that they wanted:
-
nodes with unique IDs
-
edges with optional IDs
-
edge direction attribute, with
possible default at graph level (They want to allow both directed and undirected
edges in the same graph.)
-
attributes on entities
Some questions still to be
determined at the core level are:
-
other identification mechanisms
-
hyperedges
-
special edges
-
other edge types
After this discussion, I was
given a chance to talk about GXL. There is clearly a great deal of
overlap between GD requirements and GXL features. As a community,
they have some work to do to determine what direction they
want to go-- using GXL,
using some other existing standard, or creating something on their own. |
Back to WoSEF
home page. |