Decisions:
D. Need to have a meeting to discuss common strategy between IS-ENES and Metafor with regards to portals.
D. We should regularise the mind maps as they are now and propagate up inputs to top level and see how it works. Maybe propagate them back down in a later version of the questionnaire. All the mind map get_data will be available at the realm level.
D. Mind maps should be finalised by the end of September (i.e. no more questions in them!)
D. In next fortnight there will be major revisions to parameter handling in the questionnaire. Next will be improving navigation around the questionnaire - needs to be more intuitive.

Issues:
I. How components will be displayed - what level to bundle information at. Bundling level introduces a new concept into the controlled vocabulary. Implication could be that in-bundle navigation is different from between-bundle navigation.
I. New CMIP5 website is now up - we need to discuss with Karl what text should be used to describe what's needed for the metadata collection.
I. It'll be very important to write stories about how users will use the questionnaire.
I. Should we have a link to mind maps from questionnaire?
I. Clarification needed - realm and discipline or realm and science type? Discussion carried on off-line.
I. Key messages to give to the WGCM meeting at end September.

Next general telco - Tues 22nd September, 3pm BST/4pm CET.

- Report from Brazil meeting
The main idea was to highlight potential collaborations between Brazilian and EU organisations on various e-infrastructure topics.

Currently writing draft proposals for future calls. Thanks to Gerry for represting Metafor at the meeting.

- CMIP5 Questionnaire news

Released version with coupling in it last week. Template now no longer easily editable. Can play with version on line and coupling all works as expected. Gone back to looking at issues about attributes. Decision needs to be made about input of files in component level.

We would like to get info from the mindmaps and propagate them to the realm level, even if they were originally at component level.

Should we ask for input at component level, or should we propagate them up to the realm level? If attribute is propagated up will we still have the information about what component it relates to.

Hard for user to understand when we want to collect configuration/namelist stuff. Not sure where/how we're asking about vectors.

Need to be consistent in questionnaire.

Need to go over mindmaps and find where we want to ask for input.

Regularise mind maps as they are now and propagate up inputs to top level and see how it works. Maybe propagate them back down in a later version of the questionnaire. All the mind map get_data will be available at the realm level.

Some syntax changes needed in the mind maps.

Questions in mind maps - they need to have no more questions in them by end September!

No feedback about the coupling. Will be important to write stories about how to use the questionnaire.

In next fortnight there will be major revisions to parameter handling in the questionnaire. Next will be improving navigation around the questionnaire - needs to be more intuitive.

Seven realm components, but 120 components in total. A lot of information about multiple components can be displayed on the same page. By default we're producing CIM instances which will have a hierarchy of components all the way down. We can do another transformation to concatenate them for ESG if needed. How do we want them bundled? Bundling up to the 7 realm components is probably a good way of doing this.

ESG tree controller is controlled by the fact that each leaf is its own identity and can be searched on etc.

Bundling level introduces a new concept into the controlled vocabulary. Implication could be that in-bundle navigation is different from between-bundle navigation. In-page navigation can be done with html and between-page navigation using whatever is already used.

- WP5 CIM tools news (Allyn)

Need to get CIM into rdf format. We need to standardise the UML and then use standard tools to turn the CIM into rdf. We've a lot of time in year 3 to take our initial tools and move them forward. Protoype query tool based on xml/database entries not rdf in year 2.

Services telco will map out what bits of services will map onto what EU deliverables.

Traffic on Metafor list about IS-ENES development of a portal - we're copied in to keep an eye on things. Need to make sure people don't work alone on this.

- CV news (Marie-Pierre)

No big news. Edits on OceanBioChemistry are ready to be committed. Need to decide what to do about interviewing Max Planck people.

Should we have a link to mind maps from questionnaire? Not a problem to do - need to keep mind maps and questionnaire coupled.

Potential confusion because would go from questionnaire to flattened mindmaps - trivial, can do day before launch so hold it as an idea.

Issue of prognostic variables still hanging - need to share it with someone.

Goal to close the mindmaps by end September. Want to have a good presentation on what we've done for WGCM meeting end Sept.

Some mind maps have typos and spelling mistakes - need to be corrected as they propagate into the questionnaire. Charlotte and Gerry to proofread the mindmaps.

- CIM news (Allyn)

CIM generated documentation will be included with the deliverable. Minor changes still being made. New version will be released soon.

- Misc:
  - second EU payment

 Arrived in Reading and Eric has mailed metafor_admin list with proposed shares of the payment. Payments need to be sent by the end of the month.

  - outcome of ESG dry run

 Summarised by Cecelia - all looks very nice!

 Need to agree extra search terms in advance and then they can get added without having to have them in the questionnaire.

Extra terms help to narrow down the search. Keyword in discipline category for CMIP5 can just be "climate". ESG will be used for more models than just climate models. People will want to search on realm (e.g. atmosphere), but beyond that users can find a realm in climate/forecasting models.

 Ticket 280 - column for which attributes need to get searched on. Master list of all attributes in ESG system which ESG anticipate searching on.

 Clarification needed - realm and discipline or realm and science type? Science type not right, yet discipline is too broad.

Need definition for keyword on the display. Will keep realm, but will discuss "discipline". We already have a definition for realm.

Currently don't have definitions for controlled vocabulary lists themselves. Want to provide definitions, but will be on a different timescale.

 Experiment is the search facet required here.

 Within the CIM nothing needs to be changed.

 Keyword hierarchy for types of experiments - would be good to do but we won't be able to get to it in the near future.

   - WGCM and GO-ESSP preparation

 Balaji presenting slides for us ~10-15 minutes. We can recycle Gerry's slides from Brazil but we need to discuss key messages to send to the meeting off-line. Details of how we link in metadata collection with CMIP5 protocol to be discussed.

 WGCM is not a talk about Metafor but instead is about the questionnaire and ESG portal.

 Go-ESSP - 1 Bryan on questionnaire. 1 general presentation on Metafor (under Eric's name but he can't go).

Attending: Gerry, Bryan, Marie-Pierre(maybe), Michael, Hans, Sylvia, Phil, Mark M, Sebastien.

 Bryan and Sylvia to coordinate their talks off-line to avoid duplication.

 ESG portal being moved to new machine at the moment. Talk to Luca about the portal when Sylvia's away.