<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">2018-03-27 10:37 GMT+02:00 Ilya Zverev <span dir="ltr"><<a href="mailto:ilya@zverev.info" target="_blank">ilya@zverev.info</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">First, it is not an improvement. For you, the mapper, it may be, because with the _current_ QA tools it is hard to review big changesets. </blockquote><div><br><br></div><div>improvements for the mappers are generally improvements for the project.<br><br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">But this essentially bans any big imports or QA.<br></blockquote><div><br><br></div><div>it could be, let us discuss this openly. I see that it will make big imports more difficult to happen, but on the other hand it is likely that it will improve overall quality, because mappers are more likely to look into local changes and locally limited changesets than into global ones.<br><br>Also changeset size should be optimized, when the cs are too small, their list becomes endless and it gets more difficult to get the big picture, if they are too big, many people refrain from looking into them because it takes too much time.<br><br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Second, Pierre suggest a rule that cannot be enforced. I see big QA changesets every day, and what I don't see are a similar stream of reverts or discussions on any of the mailing lists.</blockquote><div><br><br></div><div>it really depends what you mean by "QA changeset", if they don't create discussions, maybe it's because they don't create problems? Merging a third party dataset (import) is a completely different operation than finding intrinsic problems like highway=residental, because it means overwriting our data with that from another source, so we must have confidence that it is more accurate and up to date than what we have.<br><br><br><br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> This new rule would essentially block "good" imports or QA, authors of which are ready to discuss their work. But it won't stop people who do these without any prior discussion. Which means, more and more people will try to "sneak" their changes under radar.</blockquote></div><br><br><br></div><div class="gmail_extra">we can't have rules because people will try to avoid them? This could be said about any rule. I agree we should enforce the existing rules, and try to stop people who make imports ignoring the guidelines.<br><br></div><div class="gmail_extra">Cheers,<br></div><div class="gmail_extra">Martin<br></div></div>