| abian added a comment. |
Before a proper discussion about item selectors in general, I leave some thoughts about inline validation of constraints, which could bring different consequences depending on the use case.
First case: vandalism (bad faith)
- I am a vandal, i.e., someone who wants to damage Wikidata, probably just for fun, but maybe for other purposes.
- Without data validation:
- I load a page and write some nonsenses on it.
- I click the option for saving my changes and they are sucessfully saved.
- I do something else (e.g., vandalizing other statements/items or just leave the project). There's a chance p for my nonsenses to violate a constraint. p depends on how restrictive the constraints over that statement are.
- If there's a violation, this change will be easily detected by other editors and reverted. To achieve that, editors have to regularly spend some time.
- If there's no violation, it could take editors longer to detect this bad change.
- With data validation when pressing the Save option:
- I load a page and write some nonsenses on it.
- I try to click the option for saving my changes. If the data validation system detects my vandalism, I try with something else and repeat n times until I achieve my goal or get tired.
- If I haven't got tired, my changes are saved and I do something else (e.g., vandalizing other statements/items or just leave the project). My nonsenses don't violate any constraint so it could take editors longer to detect them.
- With complete, inline data validation:
- I load a page and write on it some of those nonsenses that the UI allows.
- I click the option for saving my changes.
- My changes are saved and I do something else (e.g., vandalizing other statements/items or just leave the project). My nonsenses don't violate any constraint so it could take editors longer to detect them.
Second case: mistakes (good faith)
- I am a good-faith contributor, i.e., someone who wants to help but can make mistakes.
- Without data validation:
- I load a page and accidentally write some nonsenses on it.
- I click the option for saving my changes and they are sucessfully saved.
- I do something else (e.g., editing other statements/items or just leave the project). There's a chance p for my nonsenses to violate a constraint. p depends on how restrictive the constraints over that statement are.
- If there's a violation, this change will be easily detected by myself or by other editors and reverted. Editors have to regularly spend some time on reviewing these cases.
- If there's no violation, it could take us much time to detect the mistake.
- With data validation:
- I load a page and accidentally write nonsenses on it.
- If there's a violation, data validation detects that my data are wrong. I realise and fix them. My changes are saved and I do something else (e.g., editing other statements/items or just leave the project). These changes don't violate any constraint and are probably correct. Other editors won't have to spend their time on this, they have nothing to fix.
- If there's no violation, my changes are saved. It could take us much time to detect the mistake.
- I load a page and accidentally write nonsenses on it.
In general:
- Vandalism (at least the one that is focused on "fun") is more likely to violate constraints than mistakes. But, at the same time, it tends to be more obvious when patrolling the recent changes or the watchlist.
- The more edits are made by a single vandal, the more likely it's for each of these edits to be detected (and reverted).
Opinion:
- Editors are the most valuable resource and shouldn't spend their day checking constraint violation reports or, in general, fixing mistakes that others made because of a lack of knowledge (knowledge which is not properly provided by the UI, or not provided at all). One of the main reasons I have read not to edit Wikidata is that it's hard to know what properties/items to use. So, please, consider inline validation and other measures to avoid as many mistakes as possible.
TASK DETAIL
EMAIL PREFERENCES
To: abian
Cc: Mbch331, Greta_Doci_WMDE, Jakob_WMDE, WMDE-leszek, Lexicographical data, KaMan, Lea_Lacroix_WMDE, abian, Aklapper, reosarevok, Lydia_Pintscher, Pablo-WMDE, Mringgaard, A.S.Kochergin, Lahi, Gq86, GoranSMilovanovic, QZanden, LawExplorer, JGirault, Wikidata-bugs, aude, Darkdadaah
Cc: Mbch331, Greta_Doci_WMDE, Jakob_WMDE, WMDE-leszek, Lexicographical data, KaMan, Lea_Lacroix_WMDE, abian, Aklapper, reosarevok, Lydia_Pintscher, Pablo-WMDE, Mringgaard, A.S.Kochergin, Lahi, Gq86, GoranSMilovanovic, QZanden, LawExplorer, JGirault, Wikidata-bugs, aude, Darkdadaah
_______________________________________________ Wikidata-bugs mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
