Double Workshop Meeting Notes
POSC Interoperability:
Application Interaction RFC Workshop
OpenSpirit 2d and 3d View Specification Workshop
June 24, 1999
IFP office, Rueil-Malmaison (Paris), France
Attendees
CGG:
Jean-Paul Marbeau, Jean-Gérard Boidot, Nabil
Tnacheri
Chevron: Fred Patrick, Clay Harter
Elf: Gérard Huard
IFP: Jean-François Rainaud, Jean-Louis
Pajon, Pierre Fery-Fogues
NITG-TNO: Rob Versseput
Paras: Alan Smith
PDS: Dave Foreman
POSC: Alan Doniger
PrismTech: Steve Osselton
Shell: Howard Chan, Ineke Reijsmeijer, Ben
Weltevrede
Statoil: Eirik Time
The Data Room: Neil McNaughton
Meeting
Agenda
- Introduction
- OpenSpirit
Viewer Specification Project: Technical
Presentation (Howard Chan)
- Application
Interaction RFC:
- Overview
- Demonstration
- Attendee Input
Notes
OpenSpirit
Viewer Specification Project: Technical
Presentation
Howard
Chan (Shell) gave a short presentation (PowerPoint file) summarizing the
project and then gave a detailed explanation of
the specification document.
- Objective: define standard interfaces for
visualizing OpenSpirit data based on OMG
Corba and OpenSpirit 2d graphics.
- Team: 9 people
- Duration: 6 working sessions lasting a
total of 7 work days.
- Experience: Most have 3d viewer
experience. Most do not have prior CORBA
experience.
- Status: A first draft has been completed.
There is room for many improvements.
- Open Issues: software license
requirements for viewer developers,
completeness, stability, review within
OpenSpirit, submissions to POSC.
- Future Items: more visuals, more event
handling devices, data duplication /
sharing, cross-vendor interoperability,
remove collaboration support, visual
capabilities using CORBA properties, 3d
utility classes, sub-menu support, and
more.
General Q&A:
- How difficult will it be to implement a
viewer? Answer: Two vendors with CORBA
experience (but no OpenSpirit experience)
plan to have prototypes ready at SEG 99.
- Comment: Some viewer developers have expressed
concern about the expense (money and resources)
needed to acquire and install vendor data store
development kits (e.g., for OpenWorks and GeoFrame)
in order to test their OpenSpirit compatible viewer
products. They would like there to be OpenSpirit
data servers that run over simple (ASCII) files so
the vendor data store development kits would not be
needed at all.
- Is there a separate requirements
document? Answer: No. The requirements
can be discerned from the working meeting
minutes. Comment: The requirements and
non-requirements must be extracted to
highlight the context, dependencies,
basic boundaries of capabilities, etc.
- Comment: The OpenSpirit eRoom should be
considered to host discussion on the
draft specification.
Viewer Q&A:
- Has stereo capability been addressed?
Answer: Yes. There is a stereo viewer
property. Stereo-capable hardware is also
required, though.
- Has screen splitting into four parts been
addressed? Answer: Multi-pipe
capabilities were not addressed. One
screen / window can be handled.
- A life cycle approach to viewer creation
should be considered.
Layout Q&A:
- Comment: We should be able to emulate the
style of a vendor product by controlling
aspects of layout.
- Has the attachment property been defined?
Answer: No. This can be added with an
extended layout manager. In general,
vendors can define additional properties.
We did not define how such properties can
be rolled into the "standard"
set.
- Comment: The dividing line between things
that are defined and things that are
delegated for implementations to define
should be defined.
- Is there a capability for an applet to
determine the features supported by the
available hardware? Answer: A viewer can
be asked about the features it supports.
The availability of a specific visual,
e.g. for voxels, indicates that the
associated feature is supported.
- How many visuals are involved if a
stacking chart is placed across a seismic
section? Answer: Two. There is a stacking
chart visual in 2d and a section visual
in 3d.
Properties Q&A:
- Is the "view type" a short-cut
for a list of supported visuals? Answer:
Yes. A query can be issued for the
available view types, e.g. you may ask
whether there is a viewer from the ABC
company with a 3d view.
- Comment: The direct use of the OMG Corba
Property service should be considered. A
subset of the properties can be made
available in the Trader service for
query.
View Q&A:
- What coordinate system is used with view
attributes? Answer: The coordinate system
used is specific to each attribute.
- Has output to hardcopy devices been
considered? Answer: No, not explicitly as
far as defining hardcopy-specific
attributes and parameters, print scaling,
etc.
- Is "polyline" defined as a
visual? Answer: We debated whether to
define only a small number of primitive
visuals to which others can be mapped. We
decided not to do that, but we anticipate
that some implementations will factor out
primitive visuals.
- Is the OpenSpirit "sampled
data" (i.e. Shape, etc.) supported?
Answer: No, but the document says that it
can be added.
Event Type Q&A:
- Howard: This will be covered in greater
detail in the application interaction
initiative. What is defined in the
document is very limited and there is a
lot of room for improvement. More
requirements must be considered.
- Was XML considered? Answer: XML was
discussed late in the project, but it was
not addressed.
Menu Q&A:
- Howard: This will be covered in greater
detail in the application interaction
initiative. Only one level of menu is
supported. This should be extended.
Closing General Q&A:
- Howard: It is important to identify
anything in the document which will
inhibit making extensions to the
specification where needed. Backward
compatibility is important once
implementations are begun using the
specification.
- Comment: There should be a separate
requirements document.
- Comment: Some functionality documented as
distributed interfaces in OMG IDL should
be converted to local functionality
specified in another manner.
- Can new visuals be added? Answer: The
lower level protocols needed to ensure
the ability to mix new visuals with
existing ones have not been defined
adequately.
- Is the specification limited to live user
sessions? Answer: This is its
orientation, but it is not an absolute
limitation. Data and property change
events, for example, need not be related
to live user interaction.
- Comment: This project was a good assembly
of qualified 3d graphic experts. Another
cycle of work is needed to complete and
integrate this work.
Application
Interaction RFC Overview
Alan
Doniger (POSC) began this session by giving a
presentation (PowerPoint file) explaining the
purpose and content of the Request for Comments
(RFC). Nabil Tnacheri (CGG Petrosystems)
continued by describing the examples presented in
the RFC. Alan concluded the session by explaining
the RFC process and by describing how a project
to develop open, application interaction
specifications could be organized.
Application
Interaction RFC Demonstration
Nabil
gave a detailed, live demonstration of the
current CGG Petrosystems implementation of the
concepts described in the examples of the RFC
document.
Application
Interaction RFC Attendee Input: Comments
- This
would should be based on existing work,
e.g. OpenSpirit, POSC interoperability,
etc.
- The
value and usefulness of each aspect of
this work should be clearly explained.
- The
formal description of the application
interaction model should be defined.
- Existing
applications should be mapped into
components that follow the proposed
interaction model as verification and as
a source of test cases.
- Past
experience has shown that users find too
much flexibility to be daunting.
- The
definition of a common style guide for
application interaction will be very
useful.
- The
mechanisms and protocols for assembling
virtual applications should be addressed.
- The
use of communication channels and
gateways should be studied.
Application
Interaction RFC Attendee Input: Follow-up
Intentions
- CGG
Petrosystems
- We will
respond to the RFC.
- We will
contribute to the project.
- We want to
understand the requirements of
other companies.
- The
participation of major vendors is
important.
- Chevron
- This
subject must be addressed.
- We will
respond to the RFC.
- The
process that should be followed
requires some thinking.
- IFP
- We are
pleased with this initiative. It
is interesting.
- We will
respond to the RFC.
- We will
participate in the project.
- NITG-TNO
- This is a
positive initiative.
- We will
respond to the RFC.
- Paras
- This is
interesting. The initiative
should proceed.
- We will
respond to the RFC.
- There are
commercial issues to consider,
e.g. to avoid limiting the
participation of some vendors.
- Shell
- This is an
important initiative. This is a
good start, and it is essential
that this initiative goes
forward.
- We will
answer the RFC.
- We would
like to contribute, but timing /
resource availability is a
question.
Workshop
Closing
With
respect to the Application Interaction RFC, Alan
restated how important it is for attendees to
respond to the RFC and that it is important for
attendees to encourage their partners, customers,
and suppliers to respond. This work will proceed
only if there is support within the POSC
membership. Given the positive response at
today's workshop, this initiative should go
forward and should be successful to the benefit
of the industry as a whole.
With
respect to the OS Viewer draft specification,
planning will proceed by OpenSpirit and POSC to
organize a meaningful technical review of the
specification and to propose an appropriate way
for this valuable work to continue in a
meaningful and timely manner.
Alan
Doniger closed the workshop by thanking the
presenters, Howard and Nabil, by thanking the
attendees for their contributions, and by
thanking IFP for hosting the workshops.
Comments and Requests for Further Information
|