René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[ENH]: Negative errors in errorbar()","articleBody":"**Update: See summary in https://github.com/matplotlib/matplotlib/issues/26187#issuecomment-1609533022**\r\n\r\n---\r\n\r\n\r\n### Problem\r\n\r\nIn https://github.com/matplotlib/matplotlib/pull/21266 (v3.6) we prohibited negative errors to bring the behavior in line with the documentation.\r\n\r\nI've got the request to again support negative errors, so that the error bar can be detached from the marker. There seem to be cases where users exactly want this behavior, e.g.\r\n\r\n![image](https://github.com/matplotlib/matplotlib/assets/2836374/1a3a0bd1-cce4-4383-9e05-a5c111757164)\r\n\r\n\r\n\r\n### Proposed solution\r\n\r\nI think we should support this somehow. Possibilities:\r\n\r\n- Revert to the old behavior and adjust the documentation accordingly. I'm not clear whether not allowing negative errors was really a conscious decision or whether we were only making the code behave according to the documentation. - Technically, negative errors are perfictly fine. The only justification I see for the positive error limitation is to enforce a more narrow semantics that errors are usually positive and we want to guard users against accidental negative errors. \r\n- Alternatively, we could introduce an explicit `allow_negative_errors` flag.\r\n\r\nChoosing between those, I'm inclined to say that the flag feels a bit cumbersome and we can assume that users are responsible enough to handle their data without us needing to enforce negative errors. After all, even if they are undesired, it should be relatively easy to spot and find out the reason.\r\n","author":{"url":"https://github.com/timhoffm","@type":"Person","name":"timhoffm"},"datePublished":"2023-06-26T07:20:18.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":22},"url":"https://github.com/26187/matplotlib/issues/26187"}