wiki:WP1/TelcoMinutes/Telco143_5Jan2012

Telco 143 5th January 2012

Attending: Sarah, Sophie, Sylvia, Hans, Martina, Kirstin, Frank, Gerry, Marie-Pierre, Bryan, Eric, Charlotte, Stephen, Allyn

Apologies: Mark M

Next telco:

to be arranged week of 7th February 2012 Doodle at  http://www.doodle.com/84ypgsckhq9twr9b

New Actions

  • A. Mark M - arrange recurring telco for priorities (people: Mark M, Gerry, Allyn, Sylvia, Luca, Hans, Sebastien, maybe Rupert and Paul)
  • A. Mark M - send summary of priorties telco to list
  • A. Sebastien and Gerry - deploy ensembles questionnaire
  • A. Gerry and Allyn - prod Rupert and Paul about who's in charge of the validation software.
  • A. Bryan - take a first pass over the GMD papers, then set up a telco to discuss after next telco.
  • A. Sophie and Marie-Pierre - look at re-structuring the CV paper.
  • A. Gerry and Charlotte - set up wiki page to list all question requests.
  • A. Eric and Bryan - discuss new questions issue with Karl and send him a draft email to questionnaire users asking what questions they'd like to ask
  • A. Bryan - talk with Karl about who has the ball for the governance committee, and then get it rolling.
  • A. Martina, Frank, Sylvia, Stephen, Charlotte, Gerry - continue to discuss DOI landing pages via ticket and arrange a telco if necessary and report to next telco
  • A. Charlotte - summarise DOI issues as discussed in the telco and send to the list.
  • A. Martina, Frank, Sylvia, Stephen, Charlotte, Gerry, Bryan - continue to discuss requirements for QC level 3 with Karl and Michael and arrange telco as needed.
  • A. Sylvia - send details of next ESG demo to list when available.

Previous Actions

  • A. Sarah - complete Metafor final report - ongoing, minor issues keep coming back...
  • A. Sarah - Put public deliverable documents on the wiki - ongoing
  • A. Sebastien, Hans, Mark - write specification for tracking id resolution service - ongoing

Completed actions

  • A. Sarah - add Martina to the Metafor list - done
  • A. Mark M, Gerry, Allyn, Sylvia, Luca, Hans, Sebastien - come up with a list of potential developments and the effort involved to discuss at telco for PIs to determine priorities - done
  • A. Eric - let Gerry know when to set up a doodle poll for telco for potential US-EU work - done
  • A. Gerry - put replies from questionnaire "how's it going" email on wiki page - done
  • A. Gerry - email Paco and Sebastien when the ensemble questionnaire is deployed - dropped
  • A. Gerry - give demo of the Ensembles questionnaire via dfn at telco next week - dropped
  • A. Charlotte - definitions for new properties from questionnaire scrub - done- ticket 326 - done atmosphere

Decisions

  • D. Focus on getting the 2 papers into GMD special issue by end Feb.
  • D. Every six months, issue page in the questionnaire with new questions and send to the modelling groups. Also canvas modelling groups for questions they'd like to have answered.
  • D. Monthly telcos dates should be arranged via doodle

Issues

  • I. Who's in charge of the validation procedure?
  • I. Long-term display of tables, including interactive tables.
  • I. Issues with assigning DOIs to decadal experiments

Happy new year everyone!

- Requested topics

- GMD special issue papers (Sophie)

Editors happy with end Feb target - Bryan aiming for that too. Marie-Pierre needing to discuss this with others, but won't be ready to work on this until Feb.

We need a bit of guidance on how to re-structure the papers for this special issue.

Better to have two papers, though it is possible to merge them into one.

Could we write another paper for this special issue on the tools? Could have one on the portal? Mark would need to decide.

GMD allow special issue to keep adding papers over time.

- IS-ENES GA (Sophie)

Dynamical core workshop following IS-ENES GA - trying to measure the performance of the different ESMs. How many years of simulation per day the ESM can achieve - no one knows how to measure this. Proposed to add this measure to the CIM.

To do this, need the composition - e.g. configuration of processors. Not in CIM 1.5, but in CIM 2.0. Only way to get this info is to add a free question to the end of the CIM and ring up people who have already filled the questionnaire in. (see next section)

- questionnaire news (Charlotte, Gerry)

- request for new information from modelling groups

We need to have a strategy for dealing with these requests. Every 6 months go back to the community with a list of extra questions that people want for CMIP5/from the CIM.

Should keep all extra information requests in the questionnaire. One page of additional questions that just builds up gives potential for a very disorganised questionnaire.

CIM 2.0 won't be out before the summer.

Keep the new information in the questionnaire database and build the questions and can move over everything when CIM 1.5 is ported to CIM 2.0

IPCC questions, phase 1. IPCC questions, phase 2, all on one page for users to complete. Info put in the questions can be stored as appropriate in the DB.

It's up to us to decide what questions should be asked, with the blessing of Karl. Time phases of extra questions with IPCC deadlines.

Can actively ask groups for what information they'd like to have collected.

How do we display these results? Will want to put the answers in the right place. Even though things are collected on one page, they'd be put in the right places in the DB (stick it in property values).

Extra questions mainly to create special tables - answers don't have to go through the trackback. Can bring these into CIM 2.0 in the future. Tables can be generated using a mix of CIM and new content.

Might be some sensitive issues. New questions won't be mandatory.

Will need a discussion about the long-term display of tables, including interactive tables.

- the missing stable key to connect data from the DRS world to CIM instances (Martina and Frank)

Want to link to the CIM metadata from the landing pages of the DOI, and are having some difficulty. No authoritative CIM portal to link to.

Right now the ESG federation is the one place displaying the metadata (until the Metafor portal comes on-line operationally).

ESG could change to use experiment name rather than simulation name. But multiple simulations per experiment - so simulation name is the unique id. Mismatch affecting DOI and search!

Simulation name is a key for other things in the DRS. From each simulation name, has associated info like DRS, rip etc.

DOI for an experiment, so more than one simulation for each each experiment. Simulation can have multiple ensembles.

Decadal is special - solved by appending the start year of the simulation to decadal e.g. decadal1962

One DOI pointing at multiple simulations for decadal - special case (TAMIP might be similar).

For most cases, one simulation per experiment.

Data author and metadata author should be able to tell you what the connections between the model and simulations are.

Can help users enter model names which are recognised by the DRS by using text recognition.

Have to access the xml from the atom feed to link to the DOI landing page. Atom feed has correct DRS names in the xml. UUID is needed to get xml from the atom feed, but DOI pages don't know what the UUID is.

- relaxation in the requirements for completed CIM records so that data is ready for assigning QCL3 by the *end of january*

(Stephen).

Suggestion that less complete should be released into the atom feeds by the end of the month. Need a higher filling level of the CIM instances - can agree on other measure. At end of month need to start handing out DOIs - need accessible content in CIM instances.

Charlotte and Gerry going to start a charm offensive and targeting groups which have data in the nodes, but not metadata.

Relaxing the validation rules will affect things in the software displays downstream.

DOI was always the carrot for filling out the questionnaire.

Deadline for DOI assignments is IPCC deadline in June.

Pressure should be put on the modelling groups to fill out the questionnaire, rather than compromising the questionnaire content completeness.

One experiment from the Norwegian group has a DOI. This group has no metadata. Data was ok, and had passed QC2.

Should keep the rule that experiments only get a DOI when they provide complete questionnaire metadata.

Metadata on the data can come directly from the groups (netcdf files, author conversations etc.).

Started to work on MPI experiments as well - have to work with many in parallel.

BADC - MIRP project planning on writing a paper about using the questionnaire to find information for the impacts communities.

- WP4/5/6 CIM tools and services news (Mark M et al.)

CIM Portal

  1. Version 1.0 (beta 2) was released during December and demoed at IS-ENES meeting in Lecce, Italy.
  2. Version 1.0 (release candidate 1) to be deployed either next week or the week after. If this release passes testing then official announcement can be sent out.
  3. Viewer to be refactored in version 1.1 to use new cim library being developed in the context of EU/US working group (see below).

Embedded CIM Viewer

  1. Prototype successfully integrated into new ESGF Peer-to-peer web front end.
  2. Requires CMIP5 questionniare feeds to be modified so as to inculde DRS compatible information in order to resolve CIM metadata from ESGF datasets names.
  3. Requires CIM Portal 1.0 (final) to be deployed.

EU / US Working Group

  1. First telco held in December.
  2. Currently setting up project management infrastructure such as wiki, ticketing system ...etc.
  3. Development of a meta-programming library that auto-generates code in various programming languages. The auto-generated code will allow a developer to working with a cim instances in both the programming language of their choice (e.g. python, java, javascript, c++, ... etc) and the encoding of their choice (e.g. XML, JSON, RDF ... etc).

Other

  1. Contacted Peter Fox and he has agreed to assist with developing a CIM linked data rendering engine that can be embedded in web-pages. Will also work on CIM metadata visualization.
  2. Idea of developing a CMIP5 IPad application has been mooted.

- ESG news (Sylvia)

ESG working on a new version that's very different from current software - v2.0. Demo next Tues 1600 UK time. Worth a few Metaforians attending.

Heads- up: politics with ESG and ESGF. PCMDI rumoured to not be upgrading to ESG 2.0, but will be using a new software stack that doesn't have metadata in it. Trackback may disappear in 1-2 months, depending on what software gets installed.

Not doing a lot of development on trackback now as a result of this.

- Dissemination news (Sarah)

Nothing to report.