[Go-essp-tech] Minutes: 9/30/09 ESG Model Metadata Demonstration

[Sylvia Murphy]

Present:

Eric Guilyardi
Julien Chastang
Luca Cinquini
Serguei Nikonov
John Krasting
Gerry Devine
Chris Rehbein
Lori Steadman
Sophie Valcke
Marie-Pierre Moine
Aparna Radhakrishnan

On 9/30/09 the Curator team conducted its latest demonstration of  
the   developmental state of the model metadata display within ESG- 
CET.    Thanks to everyone who participated particularly those from  
GFDL, who   in addition to their participation provided us with our  
sample CMIP5   instance. The following items were covered:

* A general review of the trackback for those who had not seen it.
* A status report on the changes requested at our last demonstration.
* A demonstration of the enhancements made to the trackback tab    
structure including the introduction  of sub-tabs, a new experiments    
tab, and a back to search target feature.
* A demonstration of the new definitions pop-ups capability.
* An initial look at a CMIP5 example from GFDL.
* A quick examination of a CMIP3 simulation instance harvested from    
the ESG database.
* A description of the design for the new component navigation    
(completion is targeted for the Jan 2010 release).

PREVIOUS ACTION ITEMS:
The following action items from the June 2009 demonstrations have   
been  implemented:
* Definition pop-ups
* An "all properties" tab
* A link to GFDLs gridspec documentation on the grids tab
* A test of access to a controlled vocabulary server
The following item is carried over
* Connect the grid netCDF files to the grid tab and the grid metadata


ACTION ITEMS :
There was much helpful discussion about the structure and utility of   
the ESG model metadata trackback pages.  Below is a summary of the    
generated action items:

1) Move "Year released" from the technical properties tab to the   
basic  properties tab
2) Add a description of the date format to the definitions of    
X_StartDate & X_EndDate
3) Remove links to components in the grid tree
4) Change the simulation search on the search interface so that   
simulations are only associated with full models - sub-components of  
simulations won't be returned as search results (Luca)
5) Define the targets on the search interface somehow (Luca)
6) Fix synch issue with the esg.owl file.  The latest file with all   
corrections made to the definitions was accidently ovewritten.


The following is a summary of some of the general discussion:

* Several participants reiterated the need for a better component    
navigation mechanism.  Users are having difficulty determining where    
they are in the component tree.  Sylvia described the design for the    
new component navigation that was hashed out after the demonstration    
dry run.  The full fledged component navigation will be in place by    
the January release.  In the short term, a "back to most recent   
search  target" button has been added to the trackback pages.    
Everyone  seemed satisfied with this and how it functioned including   
the removal  of the link if the user returns to the initial search   
target.

* Sylvia reported that Metafor is developing a list of key properties   
for the various components.  These properties include things like the   
basic approximations used in the model and some geographical   
information.  Eric suggested that this information would be best   
displayed under the basic properties tab.  He warned that he thought   
there would be too many to display.  This will be examined once all   
the properties are in the system and a suitable example created.    
Marie-Pierre agreed to provide Sylvia with a sneak peak of these   
properties.

* Several questions arose about the names of the components in the    
tree and whether or not there was an ESG naming convention in  
place.    Sylvia indicated that the names of search targets  (models,  
model   components, simulations, datasets, and software tools) are up  
to the   users.  ESG's software is designed to display what it is  
given.  It  is  possible that Metafor may wish to establish a naming  
convention  for  simulations rather than let the users decide.

* Sylvia asked the Metafor representatives if they wanted to see  
URLs   and citations treated differently on the reference tab.  Eric  
looked   at the questionnaire and reported that it was likely that  
references   would come in as strings (e.g. a citation), as well as  
URLs.  Sylvia   reported that ESG could handle both.  Eric suggested  
that Sylvia look
carefully at references in the latest questionnaire (alpha 5).  He   
indicated there are other fields in the questionnaire that should not   
necessarily be displayed.

* Sylvia mentioned that another user of ESG was thinking about the    
need for references on individual controlled vocabulary terms.   To   
be  clear, these are the values of the attributes and not the names  
of  the  attributes. e.g. name="Mixed Layer Momentum Scheme", value =   
"TKE  turbulence closure").  The reference would exist on the value   
"TKE  turbulence closure". Marie-Pierre indicated that it is  
Metafor's  long- term goal to provide definitions for all controlled  
vocabulary   values.  She thought the logical place to add a reference  
was to this   definition.

* Sylvia reported that she had examined the questionnaire (alpha 4)   
to  better understand conformance, boundary conditions, and initial    
conditions.  She asked the Metafor representatives how they might   
like  to see these items displayed.  First, they thought that   
conformance  should only appear on the trackback for the root of the   
simulation  tree.  Sylvia indicated that if the output of the   
questionnaire comes  in that way, that would happen automatically.    
Second, they thought  that the initial conditions, boundary   
conditions, and conformances  should appear as subcategories on the   
experiment tab.  They anticipate  conformance referencing initial and   
boundary condition attributes  (e.g. "This simulation conforms to   
experiment X because of initial  condition Y").  Sylvia indicated  
that  ESG could handle strings of this  sort without difficulty and  
could  display them in a conformance  section.  Sylvia anticipates  
having  these items included in the ESG  display by the January release.

* Lori Steadman of GFDL agreed to work with Sylvia to identify some   
scientific properties that are well defined for the GFDL instance.    
Sylvia will then work with Metafor on the definitions and labels for   
these properties and get them into the ESG ontology.  Julien will   
modify the trackback display accordingly.  The goal is to have these   
into the system for the October release.  The remaining scientific   
properties are scheduled to be in by the January release.

* Sylvia, Cecelia, and Luca plan to examine the entire user  
experience  from datasets to simulations, various search scenarios  
etc.  The  purpose of this examination is to ensure the system is well  
defined  for a public release that contains limited metadata  
instances.  This  may result in changes to ESG's search interface.
-----------------------------------------------------
Sylvia Murphy
Project Manager

Earth System Modeling Project
http://www.esmf.ucar.edu
National Center for Atmospheric Research
1850 Table Mesa Drive
Boulder, CO 80305
303-497-1882
murphys at ucar.edu