Another round of lies, damn lies and statistics...
Confronted with an angry mob, the WMF stooges squeal, "buh.. buh... everyone loves this tool!!"
When asked for the methodology that shows that some mythical 70% of the people love it, they begin backpedalling talking about how, "...only 0.34% of the users have disabled it..." which smells more than a bit funny on the face of it.
Enter the actual truth. (ObHint: It's as ugly as you'd suspect.)
From the
RFC
In case you are curious, this is the database query that was used:
SELECT up_value, COUNT(*)
FROM USER
LEFT JOIN user_properties ON user_id = up_user AND up_property = 'multimediaviewer-enable'
LEFT JOIN user_groups ON ug_user = user_id AND ug_group = 'bot'
WHERE ug_user IS NULL AND user_touched > '20140604000000'
GROUP BY up_value;
This counts users who have either edited the site or changed their preferences since 2014-06-04 midnight (the rollout date, loosely). Also, this is enwiki-specific. We will publish numbers for other wikis soon. --Tgr (WMF) (talk) 00:10, 24 June 2014 (UTC)
And the rejoinder
Comment user_touched is not about "users who have either edited the site or changed their preferences", but "made a change on the site, including logins, changes to pages (any namespace), watchlistings, and preference changes". It will mostly be logins; it's ridiculous to think 300k users could make an edit or change preferences in less than one month. --Nemo 15:05, 24 June 2014 (UTC)
So, to recap, the WMF rocket scientists take the number of "disabling of the media viewer" as the numerator and "logins plus other stuff" as the denominator and try to claim that this shows vast, unswerving support for the Media Viewer.
This brings on a storm of protest from being utterly lied to...
Insert : @Tgr (WMF): Could you keep this sort of cryptic presentation off the talk page? It fails to mesmerize me. It's absolutely meaningless to 99% of the readers. It's already been demonstrated that the numbers of users who haven't disabled are rather meaningless, given that many users haven't logged on in weeks and months while most users don't weigh in on discussions like this, which has become more tacky by the day, and I'm beginning to think that is why we're seeing computer code pop up into these discussions. The media viewer was introduced with no disable feature and now the disable feature remains hidden from view. Claims about the numbers who haven't disabled are meaningless. The question should be: how many 'informed' users have disabled media viewer? The last I checked it was close to 1000, and in only a couple of weeks. -- Gwillhickers (talk) 17:13, 25 June 2014 (UTC)
@159.53.110.140: Kevin, the tone is the result of an ongoing attempt to politely write-off the feedback here, on the MV talk page and their own statistics that reveal that most English and German Wikipeida editors don't approve of a slide show as a default viewer. The fact remains that the disable numbers mean little, as again, MV was introduced with no disable feature to begin with and it continues to hide this feature at the bottom of a popup menu which is also mostly hidden. As soon as it was made known how to disable some 1000 registered editors disabled it in only two weeks -- but they took this number and compared it to all users, most of whom didn't and still don't know about the disable feature in a rather transparent attempt to support a bogus conclusion that this supports their "approval" rating. So again, if we're going to heed the numbers of those who have disabled, it should be done from the perspective of how many informed users have disabled the viewer. -- RfC not a "popularity poll"? I have to disagree here. Isn't that what the MV crew have done with their approval rating, such that it was? As you seem to know by now the reasons why MV is not popular overall have been articulated by numerous users, so the "popularity", or lack thereof, actually has a basis to it. -- Gwillhickers (talk) 18:52, 25 June 2014 (UTC)
We should not be using opt-out ratios as a definite measure of support for a feature, you cannot imply that people who do not opt-out are supporters that are in favour of Media Viewer.
When Media Viewer first came about, and there was no opt-out feature within the user preferences (or at least, before I knew of the existence of any such thing), I created a rough workaround for myself in the form of a Greasemonkey userscript that circumvents Media Viewer. Since nothing on my end is broken yet, I'm too lazy to change things around even though I now know that I can disable Media Viewer in the user settings. Within your statistics, I'm probably considered a "supporter" of Media Viewer. I'm just a lazy person, but there are probably many other reasons why other people haven't opted-out yet, and you cannot infer that it's because they all support the feature. --benlisquareT•C•E 04:04, 25 June 2014 (UTC)
@Tgr: I have a few issues with your post. I think that several folks here and on the talk pages have absolutely implied that low disablement rate was an indicator of approval. Several folks have taken issue with that (including me), yet people continue to defend the idea. Second, while some opponents seem firmly in the Chicken Little camp, most are not trying to say that users are totally freaked out but that the consensus to make this a default feature was flawed and that the feature itself was not ready for Prime Time. Third, I fail to understand how a lack of complete freak-out is any argument whatsoever in favour of the making Media Viewer a default. As a Reductio ad Zombium, the fact that 99.25% of the people on earth are expected to recover from a Zombie Plague is a piss-poor reason to set the virus loose in Heathrow's International Departure Lounge. Lastly, the banner for this section relates to the less-than-honest use of statistics. I think the originator was incensed by the attempt to use a 0.34% (or 0.75%) "death rate" as an indicator that the tool is reaching acceptance. The same can be said of the (at best) sloppy use of approval ratings from account-holding, project-team-chosen, Hungarian and Catalan users as an indicator of worldwide acceptance. !Vote is all well and good, but it's baffling that an objection rate exceeding 60% is seen as anything other than an clear indicator that this needs to be rethought. You have been an island of sanity in this discussion, but this post is simply weird. 159.53.110.140 (talk) 16:48, 25 June 2014 (UTC) (Kevin)
I'm stunned.
Probably the single most dishonest thing I've ever seen the WMF engineering group try.
So the victims say, "Let's make it opt-in..."
A different proposal
@Fabrice Florin (WMF): -- In a section above, Fabrice Florin informs us that he intends to replace the present statistics with the percentage of users who disable the new feature, claiming that such metric gives us a better representation of the community's overall acceptance. Fine! Thus, let's do what he proposes, with a twist: disable the feature as the default viewer and count the users who turn it on. Fair enough? -- Alvesgaspar (talk) 20:59, 24 June 2014 (UTC)
Howls of protest from the other side.
Too funny