Reviewed the query from SC which questioned whether there should be an allowance for model developers not to fulfill the requirements.
@sno036: if clearly out of scope, yes; but not for 'optional' reasons: choice/lack of time/interest.
Guidelines should be seen as mandatory but exemptions should allowed where appropriate
Suggested RP to provide criteria of what is clearly out of scope
@HamishBrownPFR noted that not all developers will have time and resources
General consensus is that the functionality should exist but note that these may not be scientifically robust.
It was discussed that the interfaces need to exist (ie be implemented), but can throw an error if used. The content of this error could (eg) describe missing science or why an interface is being used in a nonsensical manner.
Provides for future use and consistency between models. E.g. there should be consistency on how the interfaces communicate
Checking the models fulfil these requirements is part of the review process.
Examples can be seen as demonstrations of of the functionality, yet not scientifically robust examples.
Motivation is to avoid the APSIM Classic issue, where models/interfaces are unable to communicate/work together.
Acknowledged when validation data isn't available the link to the interface should still be there for future use. If no validation, can have sensibility tests.
ACTION @APSIMInitiative/Reference-panel to work with @hol353 to provide a statement of objectives/intent of the guidelines. These can include the following:
Ensure consistency and realibility
Facilitiate future development
open up for model improvement/non developers
Discussion of where examples of interfaces should reside: if they're included in the examples folder, they would complicate an otherwise simple example, yet if included in validation simulations, they would require a separate download (validation data not installed by default). Some degree of compromise required.
ACTION @APSIMInitiative/Reference-panel to work with @hol353 to finalise draft
--- RP approved Anaerobic capability for nutrients & ponds proposal. Noted that the work should be broader than rice due to experience from issues in ponding in Classic. Also the team should plan for 100% online depending on the COVID situation.
AI SC impressed with the RP coordinating this type of meeting and encouraged this. ACTION KP to inform project team. @sarahcleary to inform AI SC.
--- Discussion on 7.10 'support' and communication to broader community.
General agreement that no AI funds should be used to support classic but note that APSIM Classic is still broadly used. No further action
--- Update on AI Software position. UQ request for continued 50% time post Dec 2020.
RP approved the request from UQ to retain @hol430 at 50%. Requested that CSIRO and @hol430 are consulted prior to final decision.
Noted that AI SC has budget for funding projects/additional support. ACTION: discussion at September RP meeting. @APSIMInitiative/Reference-Panel to bring ideas to the next meeting for discussion.
5 Training
Update on APSIM Week
@sarahcleary provided update. ACTION: @sarahcleary to update registration form; communicate new dates to attendees. Agreed to Jan 31 as 'decision' date - whether to proceed/cancel.
--- RP role includes building the APSIM community. RP to promote engagement with the community via both github & science discussions.
-- Recommend
--- Attempt to reach consensus but have clear processes to acknowlege dissenting opinions
--- Scheduling: empty meeting slots as per monthly RP meets with the expectation that if no agenda items proposed, then no meeting occurs.
--- Would require clear GitHub decision making processes and active involvement from the RP reps. At least weekly review of the GitHub issues and comment on RP relevant issues.
--- Include non-RP reps into meetings/discussions. RP reps provide linkages to parent organisations
--- Agreed that in general, more than one RP rep is required, but should not be a requirement
-- Intent - more flexible membership - discussions to include appropriate people, not just RP
7 Other Business
No further issues discussed due to time.
For next meeting
@sarchontoulis - to present details on ISU training workshops. @APSIMInitiative/reference-panel to invite additional interested parties to this part of the meeting.
@sno036 noted the effort Erik has made to APSIM. Discussion on how to acknowledge this effort given his retirement in the time of covid.
@peter-devoil Include the infrastructure replacement as a GitHub issue.
Date
Tuesday 4/8/20; 9:30 AEST
Members: Video/Teleconference:
@peter-devoil; @JulianneLilley; @yashvirchauhan; @sarchontoulis; @sarahcleary; @EnliWang; @MarkLieffering; @jbrider; @kchenu; @Keith-Pembleton; @sno036; @HamishBrownPFR;
Apologies/Not in attendance
@LouisAK
1.1 Welcome/Apologies
Welcome from @peter-devoil
1.2 Review Minutes
Review Minutes: 2020-05 Minutes Taken as read and correct.
Noted two outstanding issues from the last meeting: @hol353 to complete Guidelines for Development - https://github.com/APSIMInitiative/ReferencePanel/issues/61 @sarchontoulis - proposal for potential coordinated training 'update'.
Request to add https://github.com/APSIMInitiative/ReferencePanel/issues to Teams RP meeting invites
2. Science/Software
2.1 Check for new APSIM Major Improvements
GitHub MAJOR science issues
@Keith-Pembleton provided an update a project being undertaken which focusses on climate change work in South America. The project does not involve any developers but requires infrastructure additions (addition of CO2 sensitivity code) to the Sorghum model - https://github.com/APSIMInitiative/ApsimX/issues/572. @hol430 is adding functionality: https://github.com/APSIMInitiative/ApsimX/pull/5479
No further discussion on any issues
2.2 Updates from RP on models and reviews
Models under development
No new models under development
Models currently under review
Soybean Model @yashvirchauhan confirmed he will continue the review this month.
No other updates
Models requiring review
No new models requiring review
Models/Updates in release since last RP meeting
N/A
3 Software
-- Naming of APSIM Next Gen - https://github.com/APSIMInitiative/ApsimX/issues/4690
Keep on agenda, but on hold until 'quiet time'
-- Guidelines for Development
Reviewed the query from SC which questioned whether there should be an allowance for model developers not to fulfill the requirements.
@sno036: if clearly out of scope, yes; but not for 'optional' reasons: choice/lack of time/interest. Guidelines should be seen as mandatory but exemptions should allowed where appropriate Suggested RP to provide criteria of what is clearly out of scope
@HamishBrownPFR noted that not all developers will have time and resources
General consensus is that the functionality should exist but note that these may not be scientifically robust.
It was discussed that the interfaces need to exist (ie be implemented), but can throw an error if used. The content of this error could (eg) describe missing science or why an interface is being used in a nonsensical manner.
Provides for future use and consistency between models. E.g. there should be consistency on how the interfaces communicate
Checking the models fulfil these requirements is part of the review process.
Examples can be seen as demonstrations of of the functionality, yet not scientifically robust examples.
Motivation is to avoid the APSIM Classic issue, where models/interfaces are unable to communicate/work together.
Acknowledged when validation data isn't available the link to the interface should still be there for future use. If no validation, can have sensibility tests.
ACTION @APSIMInitiative/Reference-panel to work with @hol353 to provide a statement of objectives/intent of the guidelines. These can include the following:
Discussion of where examples of interfaces should reside: if they're included in the examples folder, they would complicate an otherwise simple example, yet if included in validation simulations, they would require a separate download (validation data not installed by default). Some degree of compromise required.
ACTION @APSIMInitiative/Reference-panel to work with @hol353 to finalise draft
Review of Visualisation - https://9-volt.github.io/bug-life/?repo=ApsimInitiative/ApsimX Interesting but no further action
4 Outstanding Actions
-- Concept Note - https://github.com/APSIMInitiative/ReferencePanel/issues/57 Went over the rationale & history;
ACTION: @sarahcleary follow up with key people Chickpea, Cotton, Grapevine, SoilTemp2, Peanut & Lentil.
Outstanding actions from AI SC meeting:
--- RP approved Anaerobic capability for nutrients & ponds proposal. Noted that the work should be broader than rice due to experience from issues in ponding in Classic. Also the team should plan for 100% online depending on the COVID situation.
AI SC impressed with the RP coordinating this type of meeting and encouraged this. ACTION KP to inform project team. @sarahcleary to inform AI SC.
--- Discussion on 7.10 'support' and communication to broader community.
General agreement that no AI funds should be used to support classic but note that APSIM Classic is still broadly used. No further action
--- Update on AI Software position. UQ request for continued 50% time post Dec 2020. RP approved the request from UQ to retain @hol430 at 50%. Requested that CSIRO and @hol430 are consulted prior to final decision.
Noted that AI SC has budget for funding projects/additional support. ACTION: discussion at September RP meeting. @APSIMInitiative/Reference-Panel to bring ideas to the next meeting for discussion.
5 Training
Update on APSIM Week
@sarahcleary provided update. ACTION: @sarahcleary to update registration form; communicate new dates to attendees. Agreed to Jan 31 as 'decision' date - whether to proceed/cancel.
6 RP Roles and Responsibilities
@peter-devoil recapped the previous discussions - https://github.com/APSIMInitiative/ReferencePanel/issues/65
Agreed on the following:
--- RP role includes building the APSIM community. RP to promote engagement with the community via both github & science discussions.
-- Recommend --- Attempt to reach consensus but have clear processes to acknowlege dissenting opinions --- Scheduling: empty meeting slots as per monthly RP meets with the expectation that if no agenda items proposed, then no meeting occurs. --- Would require clear GitHub decision making processes and active involvement from the RP reps. At least weekly review of the GitHub issues and comment on RP relevant issues. --- Include non-RP reps into meetings/discussions. RP reps provide linkages to parent organisations --- Agreed that in general, more than one RP rep is required, but should not be a requirement -- Intent - more flexible membership - discussions to include appropriate people, not just RP
7 Other Business
No further issues discussed due to time.
For next meeting
@sarchontoulis - to present details on ISU training workshops. @APSIMInitiative/reference-panel to invite additional interested parties to this part of the meeting.
@sno036 noted the effort Erik has made to APSIM. Discussion on how to acknowledge this effort given his retirement in the time of covid.
@peter-devoil Include the infrastructure replacement as a GitHub issue.
8 Next Meeting
1 September 2020 - 9:30am