How to mature a proposal to changes? #261
Replies: 11 comments
-
Dear @blchoy
|
Beta Was this translation helpful? Give feedback.
-
I think contrasting what we have done in WAFS SIGWX forecast with those in response to amendments to Annex 3, there is clear benefits of moving away from following closely with the TAC templates. How and who to make a abstract of information requirements for modelling from the templates, however, is something this team and WG-MIE needs to follow up. |
Beta Was this translation helpful? Give feedback.
-
I am talking of the process that we need to follow and you are talking of spcific solutions. We need to define the process first and talk of solutions within the process. |
Beta Was this translation helpful? Give feedback.
-
I just want to mention the step we are facing great difficulties. With regard to your points:
We do have this here and here. I agree we have late comers (e.g. new extensions) that we have not critically review before considering a solution. This could be something to put into the workflow.
They are described in Issues, compared in branches for different packages and summarized in issues on "Readiness". I agree that we are not providing enough time for team members to review.
We also have a validation page to help members to validate the proposed solutions. None of these were available previously and I think we are already working on developing a process. Please see if these fits into the process or you think we need more or something else? |
Beta Was this translation helpful? Give feedback.
-
Speaking of Anna's workflow: What we are not doing well is the "Discuss Solution" process. We have a lot of discussions on the design (see example) but not until this is implemented to form a draft full set of schemas/rules we will not be able to arrive at a comprehensive conclusion, but then it will step into the validation phase. This could be a problem unique to schema development with UML models. |
Beta Was this translation helpful? Give feedback.
-
I agree! |
Beta Was this translation helpful? Give feedback.
-
To add to @choy comment about "discuss solution" I think we are missing the "discuss requirement" step. Speaking only for myself, it appears we have little say in the ICAO WGs when discussing new requirements. We'll continue to struggle with timeliness and following Anna's workflow unless we can get what is described in the Notes implemented. I have impressed upon my NWS colleagues who do participate in these ICAO WG meetings changes I like to see on behalf of IWXXM (and IWXXM-US) but never get any insight as to what was discussed and decisions made. |
Beta Was this translation helpful? Give feedback.
-
sorry I closed by accident |
Beta Was this translation helpful? Give feedback.
-
Glad to know it is. :) I am pleased to see team members are so willing to help me face my shortcomings. |
Beta Was this translation helpful? Give feedback.
-
Thank you for the effort. The only TAC template changes initiated by WG-MIE are those involving operational status (i.e. TEST or EXERCISE) and solidi to cover nil measurements. The rest of the template changes are initiated by WG-MOG. I have to say that we have some good collaborations with individual sub-groups of MOG like space weather advisory, WAFS SIGWX forecast and lately on volcanic ash advisory but not all. |
Beta Was this translation helpful? Give feedback.
-
I have a discussion with @amilan17 yesterday. Starting from next week or after, we will take a look into:
Speaking of specifications the current TAC templates provide only a little bit of information on the range/possible values of each elements with only a few examples. I know during development stage the respective teams should have analysed possible ranges/combinations. In case we cannot get involved in the development knowing what they have considered could be an alternative. |
Beta Was this translation helpful? Give feedback.
-
In a recent meeting participants raised their concerns on the maturity of certain proposals to changes. I am raising this issue to analyse what could be the contributing factors and hope that we could find a way to deal with them.
Case I: SIGMET
There are three types (WS, WC and WV) of SIGMET sharing a single TAC template for representation. While this leads to the impression that a single information model should be adequate for IWXXM SIGMET, a closer look at TAC SIGMET revealed that:
In the original design of IWXXM SIGMET, WC and WV SIGMETs are extensions to WS SIGMET. Changes in (4) and (5), however, had broken this simple relationship. In particular, the new use of the conjunction AND and the possibility of not reporting OBS/FCST (and FCST) as a snapshot across AND (i.e. sharing the same date/time in different OBS/FCST (and FCST)) needs further changing of the information model.
Other issues include: Some team members had strong believe that the 3 types of IWXXM SIGMET should be separated as they had different information models; whether we should follow closely with TAC SIGMET taking into account only a slight change in the TAC template (or event the notes) could induce a significant change in IWXXM SIGMET model; inadequate description of information requirements (we still don't know whether FCSTs can be different across AND).
Case II: Introduction of extensions for code list and enumeration classes
UML classes with stereotypes
<<FeatureType>>
,<<Type>>
and<<DataType>>
in IWXXM have an extension which allows users to add additional elements they want the instance to carry, provided that the elements involved are well defined (e.g. in IWXXM, GML or other external schemas the producers provide). In the original design extensions are supposed to be used nationally and should be removed when distributed internationally (similar to the REM section in TAC). Some States have mandatory requirements for aviation stakeholders to receive OPMET messages with REM section (TAC) and extensions (IWXXM). There are also States who need to provide information beyond Annex 3 provisions, even though they need to file a difference with ICAO.At the request of a State a technical study on how to introduce extension in
<<CodeList>>
and<<Enumeration>>
UML classes had been conducted. The instances compliant to Annex 3 requirements (i.e. without extensions) will be exactly the same before and after introduction of the extensions to code lists and enumeration. In accordance to Semantic Versioning 2.0.0 this is a minor change to the schemas.In either case members of the team "feel" that the proposed changes are not mature enough for public consultation. To me a proposed change is "mature" if it:
I would like to hear from others if they have other understandings in mind. In any case, both cases mentioned above have issues on these topics because:
For Case I even team members do not fully understand the information requirements, so there is no way to confirm the new design will meet the needs of producers and consumers.
For Case II we will need target producers and consumers to confirm the implementation fits their purpose, which needs another (public) consultation before the public consultation.
Apart from the first case which also needs better interaction with the ICAO teams, it seems to me that the TT-AvData's workflow is missing a validation step for broader audience to let team members say confidently the proposals are mature for public consultation. Putting this in (e.g. development of the WAFS SIGWX forecast), however, will definitely increase the development time and effort.
Happy to hear from anybody.
Beta Was this translation helpful? Give feedback.
All reactions