<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<div><br></div><div><br></div><div><br></div><div>Nov 16, 2023, 20:51 by talk@openstreetmap.org:<br></div><blockquote class="tutanota_quote" style="border-left: 1px solid #93A3B8; padding-left: 10px; margin-left: 5px;"><div dir="auto">In practice MR tasks are human-executed bit edits with no actual<br></div><div dir="auto">review happening (they could be useful where correct solution<br></div><div dir="auto">needs to be selected from some available, but here it would just wait<br></div><div dir="auto">until someone would manually but still robotically remove<br></div><div dir="auto">all name tags)<br></div></blockquote><div dir="auto">To clarify:<br></div><div dir="auto"><br></div><div dir="auto">- not all people active on MR skip verification or make bad edits,<br></div><div dir="auto">some are making fine edits<br></div><div dir="auto"><br></div><div dir="auto">- but there is enough of people doing edits without verification<br></div><div dir="auto">in case of tasks having single possible edit action (or not doing edit),<br></div><div dir="auto">to make MR useless for cases where verification is needed<br></div><div dir="auto">(especially with "press to edit" mode being activated)<br></div><div dir="auto">As there is high risk of such person appearing and making edits blindly<br></div><div dir="auto"><br></div><div dir="auto">- MR may be useful for cases where elements needs to be edited<br></div><div dir="auto">in multiple ways and decision needs to be taken before edit,<br></div><div dir="auto">though there were also many cases of bad editing here<br></div><div dir="auto">But it in general filters out people clicking "make edit"<br></div><div dir="auto">on repeat without any verification<br></div><div dir="auto"><br></div><div dir="auto">- MR is utterly pointless for cases where local knowledge<br></div><div dir="auto">would be needed and aerial imagery + already present tags<br></div><div dir="auto">are not sufficient for verification<br></div><div dir="auto"><br></div><div dir="auto">- MR has many edits that are actually undiscussed bot edits,<br></div><div dir="auto">just done manually by humans (still without actual review, some tasks<br></div><div dir="auto">are not even having fig-leaf request of human needing to review anything),<br></div><div dir="auto">done in violation of<br></div><div dir="auto"><a href="https://wiki.openstreetmap.org/wiki/Automated_Edits_code_of_conduct">https://wiki.openstreetmap.org/wiki/Automated_Edits_code_of_conduct</a></div><div dir="auto"><br></div><div dir="auto">- some MR tasks are undiscussed imports, breaking<br></div><div dir="auto"><a href="https://wiki.openstreetmap.org/wiki/Import/Guidelines">https://wiki.openstreetmap.org/wiki/Import/Guidelines</a><br></div><div dir="auto"><br></div><div dir="auto">- reporting challenges exists, but reports are not acted on by people <br></div><div dir="auto">maintaining MR ( even ones where people running bad challenges <br></div><div dir="auto">voluntarily closed them - still remain open at<br></div><div dir="auto"><a href="https://github.com/maproulette/challenge-reports/issues">https://github.com/maproulette/challenge-reports/issues</a> )<br></div><div dir="auto"><br></div><div dir="auto">-------------------------------<br></div><div dir="auto"><br></div><div dir="auto">Suggesting MR as way to verify objects before edit is pointless.<br></div><div dir="auto">If actual verification is needed before edits that would be done<br></div><div dir="auto">always or almost always in the same way, then MR is pointless<br></div><div dir="auto">for that and should not be suggested as a solution.<br></div><div dir="auto"><br></div><div dir="auto">If multiple types of edits can be done, depending on a case or<br></div><div dir="auto">someone has bot edit approval (but is unable to mass edit<br></div><div dir="auto">via script or JOSM or other tool) then it is a viable MR task.<br></div> </body>
</html>