results victoria meeting and roadmap
Gerard
gerard.lemson at mpe.mpg.de
Thu Jun 1 05:24:19 PDT 2006
Dear collegues
To all of you who participated during the interop in Victoria:
thank you very much for your contributions during a rather busy
week. I think we accomplished quite a lot, on which we can now
build during the time till the next interop in Moscow.
In particular, I want to list here the projects that we started,
the time scales on which we want to see results, as well as
the people who volunteered to participate.
NB.
My memory of precisely who volunteered during the meeting is patchy.
I am sure there may have been more hands raised than I remember.
Please mail me if you feel left out.
----------
Semantics:
----------
Following early work by Laurie on UCDs for simulations,
we started compiling a list of concepts on the meeting's TWiki page.
This list will be open for editing by everyone for two weeks starting today.
After this period a number of us will translate this list
into a request to the Semantics WG for additions to the list of UCDs and
words to the Semantic Vocabulary.
This we want to be finished in about a month (i.e. mid July) so that we
can send it to the Semantics group and have it discusssed by Moscow Interop.
Everyone is invited to add to the list of words on the TWiki page.
Please see that page for how to proceed to add these.
The following persons volunteered for adding structure to this list.
We should start discussing how to do so via the email list.
Following practice in other groups I suggest to prefix the subject with
[SEMANTICS] :
Laurie Shaw
Dave De Young
Peter Teuben
Claudio Gheller
Herve Wozniak
Frank LePetit
GL
.
-----
SNAP:
-----
We decided to move ahead with the Simple Numerical Access Protocol (SNAP)
that was initiated
during the workshop in Cambridge, early this year. There was discussion
about whether the
protocol should incorporate both N-Body and mesh simulations. The majority
opinion was that we
should do this.
We decided to follow the SSA approach, not the (old) SIAP approach that SNAP
was originally based on. This implies that apart from a specification of the
query protocol and
data formats, we need a data model for the metadata of the data products.
This model will determine
what is returned by the query protocol, the protocol itself, and likely also
the metadata
accompanying the final delivered data products. The project for this is
described separately below.
Also the tasks surrounding data formats is described separately.
We discussed the functionality described by the protocol. It was decided
that apart from (in principle)
allowing retrieval of complete data sets, there should be options for
retrieving subsets. These
can be sub volumes/slices etc, but possibly also random sampling.
A problem is how we can let a user decide which subset they want.
Possibilities such as querying
for a "thumbnail" small version of the complete set were shown
by Claudio in his demo of VisIVO which already has SNAP like functionality.
This will need
to be worked out further.
Volunteers for working on the SNAP protocol are
Claudio Gheller
Laurie Shaw
Herve Wozniak
Peter Teuben (?)
GL
.
Let's use [SNAP] as prefix for emails about this subject on the theory
mailing list.
Mails could be Cc-ed to the dal mailing list.
We want to have a version 0.1 document containing descriptions of the
functionality for SNAP
as well as the data formats we propose for the results to be ready for
discussion in Moscow.
After this we should start incorporating the results of the data modeling
project and
progress to a version 0.5 or further by the China interop, 2007.
-------------
Data Formats:
-------------
We discussed the formats for data products sent from theory specific
services,
such as the SNAP discussed above.
The concensus was that we'd for now stick to VOTable where appropriate,
likely using
the BINARY form for the data itself.
It was realised this may require some selling to the community, as well as
further polling
into what data formats are actually being used and how translators can be
written to
the proposed standard.
Dave Zurek volunteered to start this, Dave De Young will participate, GL in
Europe.
Results of this poll will be communicated by the next interop.
Claudio Gheller volunteered to write example translator modules for some of
the data formats
he deals with in VisIVO.
GL will do so for some of the Gadget postprocessed data sets.
These will be shown next interop.
Discussion about this could be prefixed with [DATAFORMATS]
Mails could be Cc-ed to the dal mailing list.
--------------
Data modeling:
--------------
We had a special session about data modeling for simulation data.
It was decided we would start working on this according to the following
plan:
- Agree on the problem domain (focus on discovery as use case)
- Define conceptual or analysis model of the problem domain(s)
(Physics, characterisation, computational details)
- Define a fully detailed logical model for the particular problem area
(SNAP)
- Define binding to particular languages: XML, Java/C#, relational model
See the PPT on "Data modeling for simulations" at the meeting's TWiki page
for some more details.
Volunteers for working further on this subject were:
Gerard Lemson
Herve Woznaik
Dave De Young
Frank LePetit
Laurie Shaw
Mireille Louys
Brian Thomas
.
Time scale is that by Moscow we should have the conceptual model ready for
discussion.
The model should then be tied in with SNAP, with a logical model for SNAP
incorporated in the SNAP 0.5 version.
When mailing the theory list use something like [SIMULATION DM] as prefix.
I guess the DM list could be included in such mails.
-----------------
Summary, roadmap:
-----------------
Summarising these developments is the following road map for the theory
interest group. Note that some of these tasks really fall under the
responsibility
of particular working groups, in particular DAL and DM. Their chairmen have
agreed
with our plans and we will keep them in the loop.
----------------------------------------------------------------------------
-----------------
Description
When finished
----------------------------------------------------------------------------
-----------------
Semantics:
Collect unstructured list of words and concepts for use in -
June 16 2006
registration of non-standard theory services/datasets
and data model.
Result: TWiki page
Add structure to the list, divide in UCD and Semantic
- July 14 2006
Vocabulary, cross check with VOEvent and other such lists
Result: Note
Discuss Note within IG, update and
- July 31 2006
send to Semanticsa working group as request
Data modeling:
Create analysis model for simulations containing:
- Interop Moscow 2006
characterisation
physics
numerical algorithm
experiment
Result: WD
Logcial model included in SNAP
- Interop China 2007
Result: WD
Data formats:
Poll theorists about the data formats they use in their -
Moscow Interop 2006
simulations, and their wilingness to assist in translating
these to VOTable.
Result: Note
Example code for translating some popular data formats
- Moscow Interop 2006
into binary votable
SNAP:
Write WD about SNAP, focusing on functionality and data formats
- Moscow Interop 2006
of different types of simulations: N-Body, regular mesh, AMR
Result: WD version 0.1
Incorporate data models
- China Interop 2007
Result: WD version 0.5
Prototpye applications using following possible serialisations
- After China Interop, before Autumn Interop
of data model
XML
Java, C#
relational model
Result: applications that "do" something with the model
Propose as PR
- Autumn interop 2007
Various:
Outreach to theorists to register their services
- June 5 2006
Result: email to representatives of national projects
This mail will be added as a Projects page on the TWiki of the theory group.
Please send comments to the theory mailing list.
Best regards
Gerard Lemson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/theory/attachments/20060601/43c99cc2/attachment-0001.html>
More information about the theory
mailing list