Last modified: 2012-11-27 21:38:42 UTC
(In reply to comment 0) > Edits should add a VisualEditor tag (using MediaWiki's revision tagging > system) so they can be identified What's the use-case? I guess for the testing phase it won't be too bad, but as Terry points out on wikitech-l, the tagging system hasn't gotten much love since it was implemented, so it may not scale well (fire hose and all...). (Please don't file bugs with empty descriptions [comment 0]. Summaries often change and it makes replies with quoting more difficult.)
(In reply to comment #1) > I guess for the testing phase it won't be too bad, but as Terry points out on > wikitech-l, the tagging system hasn't gotten much love since it was > implemented, so it may not scale well (fire hose and all...). > I think we'll be fine for a while. According to https://en.wikipedia.org/wiki/Special:Tags , the most-used tag (references removed) has been used over half a million times. There are five tags used >100k times, and they add up to about 1.8 million. Overall the number of tagged revisions on enwiki is probably about 2 million. Based on a quick toolserver query, the number of edits by logged-in users in the main namespace in the hour between 18:15 UTC and 19:15 UTC today (Nov 21) was about 4,500. Even if all those edits used VE, it would take 18 days for VE-tagged edits to reach 2 million at that pace. That's kind of fast, but it's slow enough that we'll have time to respond. I made change tagging configurable and turned it on by default. If it causes problems, we can turn it off.
https://gerrit.wikimedia.org/r/34686
The use-case for adding this functionality/feature is still unclear to me. In theory, there should be no distinction between VisualEditor edits and edits made via the standard (the classic?) textarea editor.
Max, The reason for knowing which edits come from which source is so where the bugs are likely to be (and has the bonus of being possible to do statistical analysis). Longer-term, as we make editing through the API easier, we'd like to move every user of the API to be labelled as the source (this is the approach taken by Twitter, for example). And of course, for the edit box to go through the API too, though that's more for sanity and security model than for tracking down bugs given how stable that code is. It's entirely possible that Tags aren't the best venue for this data, but it's what we have for now, and we can move to a better one later.
