- recommendation to improve the installation documentation, possibly with instructions for different target groups; @zklaus will bring up this topic at the next meeting of the technical lead development team
- science lead team will start a list of NCL diagnostics that would need to be converted to Python (or R) before NCL stops working (including priority); see ESMValGroup/ESMValTool#2617
- scientific long-term strategy will be discussed at the next ESMValTool workshop
@alistairsellar @axel-lauer @bettina-gier @hb326 @katjaweigel @ledm @schlunma @zklaus
- ESMValTool release v2.5
- Feedback, lessons learned, suggestions
- release went very smoothly
- communication on dates for feature freeze (tool) could be improved, in particular when release of core has been delayed
- as a side note: installation instructions could be improved by giving specific instructions for different target groups; @zklaus will bring up this topic at the next meeting of the technical lead development team
- big thanks to @remi-kazeroni and @schlunma for their work as release managers
- Feedback, lessons learned, suggestions
- Brief report on cross-team meeting (28 March 2022)
- Status strategy for backward compatibility (#7)
- discussion is still open until 15 April 2022
- afterwards, a small team of 4-5 people will work on a document to be discussed at the next ESMValTool workshop
- if interested in joining this team please contact @alistairsellar
- Status contributor license agreement (CLA)
- CLA almost finished
- technical lead team will start looking into implementing a CLA into the GitHub workflow
- Status ESMValTool governance
- document almost finished
- contact people from each participating institute need to be named soon
- Long-term scientific strategy regarding existing recipes and diagnostics
- see discussion item 3 below For details see also meeting notes of cross-team meeting from 28 March 2022 available at https://github.com/ESMValGroup/Community/tree/main/Cross%20Team
- Status strategy for backward compatibility (#7)
- Scientific long-term strategy
- Main aim of the tool (all recipes fully functional, all recipes can be applied to new datasets,
recipe archive, etc.) and strategy for contributions with no maintainer
(e.g. non-backward compatible changes)
- first thoughts of a long-term science strategy for the tool have been discussed including the main aim of the tool and a potential strategy for dealing with broken recipes that do not have an active maintainer (see slide attached below)
- a clarification of which tasks are expected from a recipe maintainer would be needed
- there is a link with the discussion on backward compatibility (see item 2)
- a suite of general diagnostics (such as map plots, time series, histograms, pdfs, zonal means, etc.) might be useful; existing diagnostics could be made more visible by adding extra documentation for a collection of such basic diagnostics
- this topic will be discussed in a dedicated session at the next ESMValTool workshop
- Future of NCL diagnostics, list of diagnostics to be converted from NCL to other language
- for technical reasons, it might not be feasible in the future to include NCL in the current ESMValTool environment; this might result in NCL diagnostics being no longer functional for most users
- the science team will create a list of essential NCL diagnostics (e.g. perfmetrics, IPCC diagnostics) that would need to be converted to e.g. Python; the corresponding GitHub discussion is available at ESMValGroup/ESMValTool#2617
- if possible, a priority should be assigned to each diagnostic to be converted based on e.g. scientific value, project deliverables, etc.
- Main aim of the tool (all recipes fully functional, all recipes can be applied to new datasets,
recipe archive, etc.) and strategy for contributions with no maintainer
(e.g. non-backward compatible changes)
- General questions and topics
- Updating diagnostics for new datasets (e.g. CMIP6 instead of CMIP5) can be difficult and time consuming
- Integration of IPCC AR6 diagnostics needs to be finished soon, in particular since some diagnostics are written in NCL (see item 3)