Closed liamhuber closed 1 month ago
Check out this pull request on
See visual diffs & provide feedback on Jupyter Notebooks.
Powered by ReviewNB
Coverage variation | Diff coverage |
---|---|
:white_check_mark: -0.09% (target: -1.00%) | :white_check_mark: ∅ |
Codacy stopped sending the deprecated coverage status on June 5th, 2024. Learn more
Files with Coverage Reduction | New Missed Lines | % | ||
---|---|---|---|---|
nodes/static_io.py | 9 | 76.32% | ||
<!-- | Total: | 9 | --> |
Totals | |
---|---|
Change from base Build 10983970779: | -0.09% |
Covered Lines: | 2989 |
Relevant Lines: | 3278 |
This proposed change goes in the opposite direction of what I wanted to achieve. The main issue is that the current for_loop syntax was considered by many to be too complex and non-intuitive, and there was concern that such constructs would scare people away. The .iter statement, although only briefly discussed, was seen as an easy to use approach. In fact, when the benefits of the workflow approach were highlighted, this was one of the killer features that immediately won people over. So I would really like to keep it, or extend it based on the discussion we had in #456. Rather than removing such a feature, I would like to push the syntax in a way that feels comfortable and intuitive to users. I am fully aware that this is not an easy path, and will require several iterations to arrive at constructs that are easy to use without introducing unwanted or unexpected side effects. But to be successful, this is the path we must take.
and update quickstart accordingly.
@JNmpi, from #456 it sounds like not only would you like to have different behaviour for
node.iter(...)
, but also that the current behaviour was actively confusing at the hackathon. I would suggest then that we just remove the current behaviour; this avoids confusion immediately, and in the long run frees up.iter
and.zip
in the node namespace for behaviour that better represents what you want. By removing it now, re-introducing the improved behaviour will avoid breaking backwards compatibility.