Personal tools
  • We're Hiring!

You are here: Home Community Minutes Meetings December 2005 Developers Meeting

December 2005 Developers Meeting

Meeting held in Dundee on Dec 1st and 2nd.

General Info

The purpose of this meeting is to bring together developers, largely based in Europe, who are using the OME data model and/or data management software or other software tools to discuss status of project ad future directions for all of our projects. Participants are drawn from both academic and commercial sectors.

We hope that a major outcome of this meeting is the definition a group of individuals and labs dedicated to development of software tools for data management. In the future, this group will drive tool and stadard development, and can seek funding for larger meetings.

Programme will include presentations by developers from the OME project and other data management projects (email to JasonSwedlow if you want to present)

Meeting Programme

Location: University of Dundee, CIR Building, M Floor Meeting Room (CIR is Bldg 34 on the map).

December 1-- Status of Projects

We will have a slightly late start to allow those flying in on early flights to arrive

   9-10:00 Arrive, meet, etc.
   10:00 - 10:15  Welcome and Introduction
   10:15 - 11:00  OME-- Current Status
   11:00 - 11:30  OMERO Server Project
   11:30 - 12:15  OME Shoola UI Project
   12:15 - 1:15 Lunch (Catered)
   1:30 - 2:00 OME Project: Process and Management
   2:00 - 3:30 Other Data Management Projects (20 - 30 mins each)
   3:30 - 4:00 Coffee (WTB Cafe)
   4:00 - 5:30 Requirements for Data Management-- External Presentations (20 - 30 mins each)
   5:30 - Adjourn to pub for further discussion/dinner.
   

December 2-- The Way(s) Forward

We will have an early start so that people can head off to tarins/planes by 2:30 - 3:00 pm

   9:00 - 10:30 Integrated Plans I
   10:30 - 11:00 Coffee (WTB Cafe)
   11:00 - 12:30  Integrated Plans II
   12:30 - 1:30 Lunch (Catered)
   1:30 - 2:30 Summary, Wrap-up
   

Accomodation, etc.

Unfortunately, we have no resources to cover accomodation, at least for this first meeting. The University of Dundee has a good [http://www.dundee.ac.uk/admissions/undergrads/hotelug.htm|list] of accomodations in Dundee.

Notes

The purpose of this meeting was to bring together developers, largely based in Europe, who are using the OME data model and/or data management software or other software tools to discuss status of project ad future directions for all of our projects. Participants were drawn from both academic and commercial sectors.

Participants

Peter Lars Claude Sean Barry Ilan Graham Patrick Jason Chris Brian Jean Russell Richard

Presentations-- December 1

Jason: Introduction and Project Status

PDF

Desired Outcomes of the Meeting

  • Clearer vision of the goals and work being done by the OME team and everyone present
  • Describe the links that are present between the various groups using OME
  • Review roadmaps for OME
  • Review funding options for OME
  • Understanding the extents of the potential community

Comments from presentation:

  • OME meets the need to attach annotations to any point or time reference in a 5d microscopy image structure
  • Shoola needs a 'most recently accessed' side panel - like in tiki
  • Internal client tutorials on how to use the tools
  • Large user activation energy (Graham)
  • as a technical supporter, what do I do? (Graham)
  • better documentation (Graham)
  • better user interface (Patrick)
  • comfort level from the user - make it easy for the user to start (Ilan)
  • automatic imports with a few features automatically set up - a dummy version (Ilan)
  • from documentation: needing to know what its limitations and features are (Graham)
  • Deployment issue, it needs to go through a evaluation phase which needs to be easy (Graham)
  • It has to be easy to kick the tires (Jason)
  • under performance (Jason)
  • Needs the "DVD" install level of simplicity
  • Needs an I-photo simplicity for organization to start - just to get folks going with it (Ilan)

Current Actions

  • Fundamentally fix the import problem (need large number of used formats properly and thoroughly supported)
  • Fundamentally fix the install problem

Chris' Presentation solicited the following input:

  • Client side caching would be nice (Ilan)
  • Offline client use - offline database chucks sent down to the client (Ilan)
  • What is the statue of the data model extensibility? (Graham) A: The system is now adapted to a permanent use case change (as opposed to the dynamic one)
  • Do we have written use cases for various OME uses? (Graham) A: Yes, various ones written to varying digress. These are now at (Use Cases)[http://cvs.openmicroscopy.org.uk/tiki/tiki-index.php?page=Use+Cases)
  • How does OMERO3 relate to the older versions? A: OMERO3 does not integrate with the older version. Our goal is to integrate the OMERO3 with the java client fully instead. Eventually, web browser UI integration will be supported in some format, however Web UI has certain limitations which we don't expect to ever overcome, so integration will be limited.

Lunch Discussion

  • When a project is set up, decide what secondary fields are set to "required", "strongly recommended", and fully optional.
  • suggestion: provide a project template system so people can just 'pick' a template. Alternately, a project start wizard.
  • from Brian: Jason, I would love to see a project to integrate remotely distributed cluster computer for OME, ala the Seti@Home project

Jean Marie's Presentation

  • How much does migrating the rendering from client to server increase server demand? (Peter) A: In our testing, it was not much.

Graham's Presentation (Oxford)

PDF

  • Important for scientist to be able to make their annotations at the microscope.

Oxford's Experience

  • OME can be a difficult system to get into
  • Complex, monolithic system
  • Documentation seems to assume audience already "on message". Tell users what OME is good at and bad at.
  • Limited and awkwared in the area of user annotation
  • Steep learning curse for the even simple project-specific annotations
  • Problems with export of annotations
  • Data model is not easy to understand (2 forms?)

Word of the day: "Desiderata"

  • They would like "Small pieces, loosely coupled" - Small ontology - lots of extensibility
  • Suggest that OME becomes 'nodes' on a web-based bio-'google' system.

After Presentation Discussion

  • The data management problem for bio-imaging data needs a solution!
  • We have now improved our QA process. Closed CVS with vetting of developers, Unit Testing, better docs, better process
  • When did we write documentation that was good enough for you? (a possible discussion for some bbs?)
  • OME Importer: Will its plugins be extensible? IE: Can a plugin have plugins?

December 2nd OME Meeting Minutes

Claude, Peter, Lars, Graham, Patrick, Jean, Brian, Chris, Jason

Agenda Topics

  • Tour
  • Specs discussion (Peter, Lars)
  • Dynamic Data Types
  • Use cases (Patrick)
  • Funding
  • Software Tools
  • Longer Term Roadmap

Action Items

*Jason: check information on 'old hardware tender' for Peter and Claude [DONE]

Dynamic Data Types

  • Ilya's original vision was 'all dynamics all the time'.
  • This is what is in the original documentation
  • In OMERO the structure is much more locked down.
  • The real question is "What is the requirement (what are the use cases) for dynamic typing?"
  • Updates must support new types
  • Suggest a domain registration style respository of 'data types' to prevent type conflict (Peter)
  • We should avoid this for a number of reasons (Chris)
  • Problem: Just adding a data type to the SERVER means updating the API and Rich Client. (chris)
  • look at SPRQL (or other simple means for interoperability XPATH, XQUERY, Reflection of the DB)for something that might help dealing with dynamic types. (Graham)
  • Suggest that the data types be reflected in the OMERO API, thus allowing developers to add their own utility interfaces to their own database systems (thus allowing them to add their own dynamic data types to a separate repository, or whatever). (Graham)
  • Jason sees 3 kinds of use cases for local dynamic types:
  • User driven
    • (Give very limited abilities to do this.)
  • Specific Interface
    • This becomes an engineering problem (modify Server, API, and Client).
  • Project Driven
    • How is this going to be handled (if at all)?
  • The solution seems to be an extensible API model (which is in the works).

Use Cases

  • Jason is going to review and revise our use cases
  • First, we need to enumerate the uses cases
  • These use cases should be driving our test cases
  • Do we need a template we can send out to collect use cases from other labs?
  • NOTE: see [http://cvs.openmicroscopy.org.uk/tiki/tiki-index.php?page=Use+Cases|Use Cases|nocache] page.

Funding

  • Renewal of the Welcome Trust Grant in the new year. 3 full time positions for 5 years. NOTE: Submitted Feb 2006
  • NIH-NII Grant submitting in June/06 with Dundee, Baltimore, MIT, Madison. Somewhere in the 3 to 5 year range.
  • Pending EPSRC grant (Peter Gregor). Expect to hear about it soon. 1 full time, 1 grad student for 3 years NOTE: Awarded, start June 1, 2006.
  • Lars and Peter are not bioinformatics experts, and are instead concentrating on the DB side. Looking at an oracle DB expert to work on it (performance and integration testing?). There is some interesting by the German industrial sector.
  • Patrick knows of now part of a project to develop new imaging equipment, but not storing it. He suggests using the Network of Excellence as a financial source.
  • JISC is funding virtual research environments.
  • Claude suggests the Swiss Institute of Technology's grad students may be a resource.

Roadmap

Ultimately, philosophic goal for OME is to data management screening fun and easy! This goal is obtainable within the 5 year scheme.

  • 2006
  • Omero3 released Q2-Q3
  • Web UI? Needs exploration - Java Applet?
  • Java Client - Interfaces
    • ImageJ (Probably looked at in 2007 instead)
  • Data model integration (not really 'complete' until some time in 2007)
  • 2007
  • 2008
Document Actions