You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current developer core is uncomfortable with iris.analysis.calculus.curl remaining in Iris' API.
curl was introduced in the earliest days when we were speculating about where Iris would prove most valuable, but complex calculus did not become a 'growth area', putting curl in an unfortunate position of requiring a high level of esoteric knowledge to maintain, and being unlike most other Iris operations. This leaves our developers unfamiliar and out of practice, hence our difficulty answering questions, and the fact that several elements (e.g. 3-dimensional curl) have never functioned as advertised.
Compare this to something like regridding, which also requires a challenging level of specific knowledge, but is a heavily used Iris feature and therefore has dedicated resource/expertise at all times.
There are several ways forward:
Seek a volunteer with the right expertise to maintain curl in Iris. There is even potential for this to expand into a larger satellite project with more calculus conveniences. (see Plugins).
Deprecate the curl function and convert the code into a detailed documentation example. This avoids the current implication that this is a maintained part of Iris, and leaves users to customise in whichever way works for them.
Deprecate the curl function and recommend people use MetPy's curl instead. MetPy would need to be included in environments at the UK Met Office.
The text was updated successfully, but these errors were encountered:
After posting this internally at the UK Met Office on 20th November, so far no volunteers have come forward. So it's currently looking like Options 2 or 3 above.
📰 Custom Issue
The current developer core is uncomfortable with
iris.analysis.calculus.curl
remaining in Iris' API.curl was introduced in the earliest days when we were speculating about where Iris would prove most valuable, but complex calculus did not become a 'growth area', putting curl in an unfortunate position of requiring a high level of esoteric knowledge to maintain, and being unlike most other Iris operations. This leaves our developers unfamiliar and out of practice, hence our difficulty answering questions, and the fact that several elements (e.g. 3-dimensional curl) have never functioned as advertised.
Compare this to something like regridding, which also requires a challenging level of specific knowledge, but is a heavily used Iris feature and therefore has dedicated resource/expertise at all times.
There are several ways forward:
The text was updated successfully, but these errors were encountered: