+1 Jonathan.  I also agree that the place where HHVM is likely to have an effect is in high-speed editing activities.  This was my conclusion when I had completed the experimental deployment to newcomers with Ori.  I think a good place to look would be edits that happen though the API.  I had a proposal of sorts that would work with the rollout to API (was separate from rollout to the rest of the site), but I wasn't able to get it picked up.  



On Wed, Aug 19, 2015 at 11:04 PM, WereSpielChequers <werespielchequers@gmail.com> wrote:
7 minutes is an average, yes?

I would agree that an editor whose hundred edits represents about 700 minutes per month would not achieve much more in the same amount of time. But the editors who do over a hundred edits a month are significantly skewed towards the gnomes and vandal fighters who's editing rate is more like one a minute, and at that point saving a couple of seconds per edit becomes more significant. So not surprising that this appears to be a power user phenomena and not something that your 5 edits per month editor would notice.

The other point is that not all time is equal. Time spent typing, searching is one thing, but time waiting for an edit to save is time the system is holding you back. So it makes total sense to me that speeding up the save time would improve the user experience for wiki gnomes and encourage them to do more. Content writers who might only save every half hour would barely notice the change unless they are working on larger articles where the speed up in save time is greater as it is proportionate to article size. Featured Articles do tend to be relatively large.

Regards

Jonathan Cardy


On 19 Aug 2015, at 23:15, Aaron Halfaker <aaron.halfaker@gmail.com> wrote:

I feel like I should expand on my skepticism of HHVM as a mechanism for the observed rise in active editors.  

The average edit takes 7 minutes[1,2].  HHVM reduces the time to *save* the edit by a couple seconds.  7 minutes - a couple seconds = ~7 minutes.  So, HHVM doesn't really help you edit substantially faster.

1. Geiger, R. S., & Halfaker, A. (2013, February). Using edit sessions to measure participation in Wikipedia. In Proceedings of the 2013 conference on Computer supported cooperative work (pp. 861-870). ACM.
2. Halfaker, A., Keyes, O., Kluver, D., Thebault-Spieker, J., Nguyen, T., Shores, K., ... & Warncke-Wang, M. (2015, May). User Session Identification Based on Strong Regularities in Inter-activity Time. In Proceedings of the 24th International Conference on World Wide Web (pp. 410-418). International World Wide Web Conferences Steering Committee.

On Wed, Aug 19, 2015 at 5:08 PM, Aaron Halfaker <ahalfaker@wikimedia.org> wrote:
So, I've been digging into this a bit.  Regretfully, I don't have my results written up in a nice, consumable format.  So, you'll need to deal with my worklogs.  See https://meta.wikimedia.org/wiki/Research_talk:Active_editor_spike_2015/Work_log/2015-07-09

TL;DR: It looks like there was a sudden burst in new registrations.  Work by Neil Quinn of the Editing Team suggests that these new registrations were largely the result of changes to the mobile app.  I didn't specifically look at 100+ monthly editors.  That seems like a fine extension of the study.  I'd be happy to support someone else to do that work.  I have some datasets that should make it relatively easy. 

> If the data is correct, then [HHVM] is likely to be one of the main reasons for the change.

Correlation is not causation.  There's no cause to arrive at this conclusion.  In my limited study of the effects of HHVM on newcomer engagement, I found no meaningful effect.  I think that, before we consider HHVM as a cause of this, we should at least propose a mechanism and look for evidence of that mechanism.  




On Wed, Aug 19, 2015 at 10:49 AM, WereSpielChequers <werespielchequers@gmail.com> wrote:
Most of those editors will have done 33 edits or less using V/E, and some, including me in 4th place, will have been having a look at V/E after the attention it has had recently at Wikimania, on the signpost and on mailing lists. I'm not sure that something that barely involves 10% of a group of editors could have had such a big effect.

More likely and just at the right time, late 2014, Erik Zachte has reminded me that we had a major speed-up with php parser change. 

If the data is correct, then that is likely to be one of the main reasons for the change.

Regards

Jonathan Cardy


On 17 Aug 2015, at 19:11, Jonathan Morgan <jmorgan@wikimedia.org> wrote:

It looks like about 10% of highly active Enwiki editors have used VE in the past month (across all namespaces): http://quarry.wmflabs.org/query/4795

On Mon, Aug 17, 2015 at 8:35 AM, WereSpielChequers <werespielchequers@gmail.com> wrote:
On a very non-scientific measure of how few editors currently use V/E, I took some snapshots of the most recent 500 mainspace edits yesterday and was getting circa 1% tagged as visual editor, I've just run two sample this afternoon and the first had not a single edit tagged Visual editor and the other only four, so unless some of those experienced users using V/e have opted out of having their edits tagged V/E, I'm assuming "gobs and gobs" are either on other language wikis, heavily skewed to a time of day I haven't sampled or big in number but still too small a proportion to account for the increase in the number of editors doing >100 edits per month.

On 17 August 2015 at 15:54, Jonathan Morgan <jmorgan@wikimedia.org> wrote:
There are gobs and gobs* of people using VE. Many of them are experienced editors. 

I'm also interested in looking at VE adoption over time (especially by veteran editors). I'll sniff around and let y'all know if I find anything.

No idea what might be causing the boost in active editor numbers. But it's exciting to see :)

Anyone else have data that bears on these questions? 

- J

*non-scientific estimate drawn from anecdata

On Sat, Aug 15, 2015 at 9:53 AM, WereSpielChequers <werespielchequers@gmail.com> wrote:
That's an interesting theory, but are there many people actually using V/E now?

I've just gone back through recent changes looking for people using it, and apart from half a dozen newbies I've welcomed I'm really not seeing many V/E edits.

Looking at the history of Wikipedia:VisualEditor/Feedback the last 500 edits go back three months. So apart from the Interior, you and I Kerry I'm not sure there is a huge number of people testing it, and I wasn't testing it in the first 6 months of this year. I did see some research where they were claiming that retention rates for V/E editors were now as good as for people using the classic editor, but I would be surprised if there were enough people using V/E to make a difference to these figures, especially as this is about the editors doing over 100 edits a month.

I agree it would be interesting to track the take-up of the VE (fully or partially) by editor by year of original signup. But I think the long awaited boost from V?E editing is yet to come, if the regulars have started to increase that is likely to be due to something else.

Jonathan

On 15 August 2015 at 15:11, Kerry Raymond <kerry.raymond@gmail.com> wrote:

Is there any way of telling what proportion of these 8% appear to be using the Visual Editor either exclusively or partially? It might be interesting to track the take-up of the VE (fully or partially) by editor by year of original signup.

 

Kerry

 

From: wiki-research-l-bounces@lists.wikimedia.org [mailto:wiki-research-l-bounces@lists.wikimedia.org] On Behalf Of WereSpielChequers
Sent: Saturday, 15 August 2015 11:12 PM
To: Research into Wikimedia content and communities <wiki-research-l@lists.wikimedia.org>; The Wikimedia Foundation Research Committee mailing list <rcom-l@lists.wikimedia.org>
Subject: [Wiki-research-l] Has the recent increase in English wikipedia's core community gone beyond a statistical blip?

 

Hi,

With 8% more editors contributing over 100 edits in June 2015 than in  June 2014, we have now had six consecutive months where this particular metric of the core community is looking positive. One or two months could easily be a statistical blip, especially when you compare calender months that may have 5 weekends in one year and four the next. But 6 months in a row does begin to look like a change in pattern.

As far as caveats go I'm aware of several of the reasons why raw edit count is a suspect measure, but I'm not aware of anything that has come in in this year that would have artificially inflated edit counts and brought more of the under  100 editors into the >100 group.

I know there was a recent speedup, which should increase subsequent edit rates, and one of the edit filters got disabled in June, but neither of those should be relevant to the Jan-May period.

Would anyone on this list be aware of something that would have otherwise thrown that statistic?

Otherwise I'm considering submitting something to the Signpost.

Regards

Jonathan

 


_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l




--
Jonathan T. Morgan
Senior Design Researcher
Wikimedia Foundation


_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l




--
Jonathan T. Morgan
Senior Design Researcher
Wikimedia Foundation

_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l