SDIWG:Meeting Minutes 20050923

SDIWG Minutes for September 23rd Tcon/Breeze (first day of Fall)

Note taker: Henry Chueh

Attending: Bill Lorenson, Ivo Dinov, Peter Lyster, Karen Skinner, John Haller, Henry Chueh, Mike Sherman, Jennifer Couch, Bret Petersen, Jenny Larkin, German Cavelier

Next SDIWG teleconference will be at 2:30pm Oct 21st, 2005.

Three new NCBC Centers will be announced and if there are named liaisons and Lead Science Officers (LSOs) we will try to invite them.

Agenda:

Discussion about incorporation of three new PI/LSOs. Continued discussion on the NIH-forge and yellow pages connection—from the July 8 meeting minutes I quote the following:

Key elements of NIH-forge

1. Yellow pages of available tools to users (IATR model http://www.cma.mgh.harvard.edu/iatr/info.php, or for ontologies (see botton note for more on Ontologies) the OBO http://obo.sourceforge.net/) a. component, b. applications, c. combine components and applications to make pipelines/workflows
2. Restaurant rating a. user evaluation for ease of use etc, b. techie evaluation benchmarking testing against datasets.
3. Environment for users communicating needs to developers (NA-MIC programmers week model). The developers should probably retain as close to their native work environment as possible—i.e., federate!
4. Developer environment, e.g., based on programmer’s week.
5. Expansion of developer environment (Dashboard)—geographically separated developers working on the same software to avoid duplication. Probably federated.
6. Code bank (cf. Genbank model)—probably virtual.

(Key elements of NIH-forge: definition retrieved from http://www.na-mic.org/Wiki/index.php/SDI:_Minutes_for_July_8_meeting_in_Bethesda)

Minutes (taken by Henry Chueh)

Peter noted that he couldn’t yet talk about the 3 new Centers that will result in the incorporation of 6 new members of the SDIWG.

Upcoming meetings noted:

MICCAI 2005 (Oct 26-29, Palm Springs) http://www.miccai2005.org/

AMIA 2005 (Oct 22-24, Wash DC) http://www.amia.org/meetings/annual/current/index.html

Peter noted that there were some funds left in FY05, and had been asked if the SDIWG had recommendations on activities? Previous discussions had focused on NIH Forge / “Yellow Pages” as one potential feasible task. Open software development was another more resource intensive and complex task. However, logistics could not be worked out quickly enough to capture the funding.

Question about what the scope of such an NCBC Yellow Pages would be--Sherm suggested that the scope was the NCBCs, not necessarily collaborators initially. It would be more of a “branded” collection of NCBC software.

A discussion about software engineering in some related areas ensued. Recent announcements include:

• NSF04607 Interagency Opportunities in Multi-Scale Modeling in Biomedical, Biological, and Behavioral Systems

http://www.nsf.gov/publications/pub_summ.jsp?ods_key=nsf04607

• NSF05570 DDDAS: Dynamic Data Driven Applications Systems

http://www.nsf.gov/publications/pub_summ.jsp?ods_key=nsf05570

Sherm asked if these are Open Source -- the Multi-Scale Modeling announcement has the same requirements for 'software dissemination' as the NIH Roadmap National Centers for Biomedical Computing (NCBC) RM 04 022.

The group returned to a discussion about how best to handle being responsive to potential funding. Should funding be divided democratically, or what is the process for rapidly making decisions about common projects? General consensus was that for small amounts, dividing up the resources would not be helpful.

Karen mentioned that it may be worth prototyping a Yellow Pages so that there is a real point of reference for everyone in terms of what the SDIWG is thinking about relative to such a directory.

Peter noted that interoperability demos are on the back burner for now, so facilitating collaboration through places like a Yellow Pages is probably the best option for now. There was consensus agreement about this.

Bill then showed a software directory supported by the Insight Software Consortium that also supports user ratings and reviews. The ultimate goal is to test a model where Insight Journal http://www.itk.org/Wiki/ITK_The_Insight_Journal can accept software submissions for formal review, and then if accepted it would be brought into the ITK toolkit (if ITK software). Someone asked how many actual ITK developers there are in the community, and Bill noted that there are perhaps 20 active developers in the ITK community who have commit privileges.

Based on reviewing the Insight software directory, Peter suggested that we might devote part of the SDIWG meetings in the future to more “freewheeling” demonstrations of each center’s activities.

Ivo commented that we need at least some basic ontology to organize software submissions. There was agreement about this, but then this means the approach needs significant resources to execute. Several people noted that a simple solution might not present our group in the best light, so better to take the time to define what we would like to produce rather than try to put together a simplistic product without specific funding.

No specific action items.

Additional notes from Ivo Dinov (Liaison for UCLA NCBC "Center for Computational Biology")

Ivo suggested that we think ahead and plan to design an extensible NCBC “SW yellow-pages” architecture for the NCBC SW developments. Our plans should NOT exclude adding tools engineered outside the NCBC initiative, in the future. Of course, we’ll start with the tools from the 7 NCBCs.

Here are some things we may consider regarding the need and challenges in establishing an NCBC SWO (software ontology). Number (4) below may serve as a starting point (and can be extended, if need be):

1. Def: An NCBC SW ontology will be a formal explicit description of biomedical computational tools (classes/concepts), properties of each tool describing various features, inputs, outputs and attributes (slots/roles/properties)), and restrictions on these slots (facets/role-restrictions).
2. Def: An NCBC SW ontology together with a set of individual/Center instances of classes/tools will constitute an NCBC SW knowledge base. In reality, there is a fine line where the ontology ends and the knowledge base begins.
3. An Example of a General Ontology: http://www.daml.org/ontologies/.
4. An example of an Algorithm Ontology: http://www.daml.org/ontologies/67.
5. http://www.ksl.stanford.edu/people/dlm/papers/ontology101/ontology101-noy-mcguinness.html
6. http://www.xml.com/pub/a/2002/11/06/ontologies.html
7. http://www.w3.org/2001/sw/WebOnt/ (posted by Ivo Dinov, CCB)