SDIWG:Meeting Minutes 20050218
Minutes for Software and Data Integration teleconference (working group and Centers liaisons) Friday 20050218
Attendees: (phone) Bill Lorensen, Don Harrington, Peter Highnam, Valentina Di Francesca, Jennie Larkin, and Larry Clarke; (Dem I) Carol Bean, Shira Katseff, Peter Good, Mike Marron, and Peter Covitz.
Lyster: Briefly discuss the document he circulated for the teleconference. This contains the SDI Mission Statement, and two planned activities (i) software engineering meeting and wiki site (hosted at NA-MIC site), (ii) interoperability demos.
Lorensen: Has established a page on NA-MIC wiki to host an ongoing NCBC SDI discussion. We urge all to view the NA-MIC wiki and become familiar with the features http://na-mic.org/Wiki/index.php/Main_Page. You can navigate to the SDI discussion through the “Community Portal” on the LHS of the main page, or go directly to http://na-mic.org/Wiki/index.php/SDIWG:Software_and_Data_Integration_Working_Group. We recommend that everyone sets up their own account (click on the top RH of the main page) so that they can actively edit the SDI page—that is the point (see Appendix below: bare bones instructions for using Kikinis’ Center (NA-MIC) wiki). This will be like a weblog. We hope that it will remain orderly and Bill Lorensen will help with that. Right now there are five threads: four of the threads will be discussions on issues relating to SDI for each separate Center, and the fifth is an integrative discussion that we’ve called “Software Engineering Body of Knowledge across NCBC Centers”. We hope that staff can enter vision/design/architecture documents for the software processes. We hope that this will be a living discussion on software engineering approaches across the centers and thus enable the best possible software and data integration.
Lorensen: Informal telecoms are good for SDI-like activity. Occasional face-to-face are useful. Involved NA-MIC staff have developed best practices over years, starting with experiences as GE. NCBC will need to develop (new) best practices because of (heterogeneous) applications.
Lorensen: Yearly interoperability demonstration should not be show and tell, but it can be very useful, e.g., at DARPA it works well. You don’t need to expect a full interoperation. Start out with simple interoperability and build on it.
Covitz: With CaBIG they found they need to define what we mean by interoperability. There are multiple categories and different degrees of strategies illustrated by a 2D matrix with ‘Category’ on one axis and ‘Stringency’ on another axis. We can structure this on Wiki: Here is the link to the caBIG Compatibility document. Disclaimer (from Peter) is that it is undergoing revision; a new version is expected in about a month. The substance won't be that different, but the revised version should be a bit clearer and more precise on some of the details. http://cabig.nci.nih.gov/guidelines_documentation It might be a useful starting point for the interoperability discussion in the NCBC program.
Lorensen: ITK has light-weight process on front end.
Marron: Interoperability means more than running software—there is a dissemination aspect. How to involve investigators outside of the immediate Centers effort using (as users and developers). Consider dissemination leading up to web services.
Covitz: Categories of caBIG—conforming to requests; programmer access to APIs. Doesn’t mean just access to database, but allowing for tools to be developed at remote sites.
Just a web service is not useful—canned information. You need to know what are you getting back—what is semantic content? What are standards?
Registries for site independent interoperable data and tools.
Covitz: Need to agree on three types of registries: (i) terminologies and vocabularies; (2) meta data registry; (3) Service registry. What has what service and where is it?
Marron: NCBC registry will have to reconcile heterogeneous domains efforts (clinical, basic science, molecular, imaging—lions and elephants).
Lorensen: NA-MIC has adopted Extreme Programming which is very anti-bureaucratic. Minimizes up front work, and doesn’t slow developers down. Users also don’t want to be gotten in the way of. Recommends a good book “Extreme Programming Explained” by Kent Beck.
Appendix: Bare-bones instructions for using the Kikinis Center (NA-MIC) wiki. Go to the main page and click on ‘Community Portal’. Follow the link to ‘Software and Data Integration Working Group’, and you will see that Bill Lorensen and I have set up five pages—one for each of the Centers and one titled ‘Software Engineering Body of Knowledge across NCBC Biocomputing Centers.’ There is nothing in those pages yet. Some of you will want to simply view the pages and watch the dialogue, and others will want to join the discussion by uploading your documents and modifying or adding to the text in the pages. If you want to join the discussion please click on the ‘create an account and log in’ link at the top right hand side of the page. To modify a page simply click on the ‘edit’ tab—there are buttons that give you help on how to edit or link to documents or pages--it’s pretty easy. The Main_Page has help for how to use the wiki.