I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of "new user", "new article" (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab le_to_create_citation_with_a_live_external_link_in_it
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling "real articles". Then on the "real articles" I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on "popular articles"). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It's hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don't think the new articles would have any "might be dodgy" status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It's just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Thanks for the explanation about the opt-in. It wasn’t a problem – just a surprise for me.
However, the Captcha problem isn’t the issue of whether or not there are Captchas. I was expecting Captchas on a new user account. But I wasn’t expecting some bug that prevented the user from Saving the Page despite being presented with Captcha after Captcha after Captcha … without end. There’s a software bug which is either 1) incorrectly determining the user response to the Captcha challenge is in error when it is correct or 2) some control flow error in relation on what to do in the event of getting it right.
Kerry
From: WereSpielChequers [mailto:werespielchequers@gmail.com] Sent: Monday, 17 August 2015 3:41 PM To: kerry.raymond@gmail.com; Research into Wikimedia content and communities wiki-research-l@lists.wikimedia.org Subject: Re: [Wiki-research-l] Visual Editor experiment might have a problem ...
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond <kerry.raymond@gmail.com mailto:kerry.raymond@gmail.com > wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
_______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org mailto:Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine On Aug 16, 2015 10:41 PM, "WereSpielChequers" werespielchequers@gmail.com wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
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
No, I'm not aware of any ongoing CAPTCHA work. There was a long thread on wikitech-l starting last December ("Our CAPTCHA is very unfriendly") that resulted in some Phabricator tasks and a wikipage. But I don't know of any active development plans.
By the way: Aaron's in transition (of the timezone variety) right now, so it may be a day or so before he's able to respond to this thread.
- J
On Sun, Aug 16, 2015 at 10:53 PM, Pine W wiki.pine@gmail.com wrote:
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine On Aug 16, 2015 10:41 PM, "WereSpielChequers" werespielchequers@gmail.com wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
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
No, you don't need CAPTCHA. You can use Honeypot. I think that would fix it.
On Mon, Aug 17, 2015 at 9:44 AM, Jonathan Morgan jmorgan@wikimedia.org wrote:
No, I'm not aware of any ongoing CAPTCHA work. There was a long thread on wikitech-l starting last December ("Our CAPTCHA is very unfriendly") that resulted in some Phabricator tasks and a wikipage. But I don't know of any active development plans.
By the way: Aaron's in transition (of the timezone variety) right now, so it may be a day or so before he's able to respond to this thread.
- J
On Sun, Aug 16, 2015 at 10:53 PM, Pine W wiki.pine@gmail.com wrote:
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine On Aug 16, 2015 10:41 PM, "WereSpielChequers" < werespielchequers@gmail.com> wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
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 User:Jmorgan (WMF) https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)
Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Hi Kerry, IIUC, the original bug was https://phabricator.wikimedia.org/T109113 ("Can't successfully complete a second CAPTCHA in VisualEditor's save dialog") which started last Thursday, and was fixed this morning (and the fix should now already be deployed to the production wikis). It was only an issue if the first CAPTCHA attempt failed (but I know how common that can be!). Sorry it affected your training session. :( I see that Whatamidoing already replied to your post at WP:VEF
Quiddity
On Mon, Aug 17, 2015 at 10:58 AM, Sam Katz smkatz@gmail.com wrote:
No, you don't need CAPTCHA. You can use Honeypot. I think that would fix it.
On Mon, Aug 17, 2015 at 9:44 AM, Jonathan Morgan jmorgan@wikimedia.org wrote:
No, I'm not aware of any ongoing CAPTCHA work. There was a long thread on wikitech-l starting last December ("Our CAPTCHA is very unfriendly") that resulted in some Phabricator tasks and a wikipage. But I don't know of any active development plans.
By the way: Aaron's in transition (of the timezone variety) right now, so it may be a day or so before he's able to respond to this thread.
- J
On Sun, Aug 16, 2015 at 10:53 PM, Pine W wiki.pine@gmail.com wrote:
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine On Aug 16, 2015 10:41 PM, "WereSpielChequers" < werespielchequers@gmail.com> wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
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 User:Jmorgan (WMF) https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)
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
Thanks, Nick! I will pass on the news!
Kerry
From: wiki-research-l-bounces@lists.wikimedia.org [mailto:wiki-research-l-bounces@lists.wikimedia.org] On Behalf Of Nick Wilson (Quiddity) Sent: Tuesday, 18 August 2015 8:47 AM To: Research into Wikimedia content and communities wiki-research-l@lists.wikimedia.org Subject: Re: [Wiki-research-l] Visual Editor experiment might have a problem ...
Hi Kerry, IIUC, the original bug was https://phabricator.wikimedia.org/T109113 ("Can't successfully complete a second CAPTCHA in VisualEditor's save dialog") which started last Thursday, and was fixed this morning (and the fix should now already be deployed to the production wikis). It was only an issue if the first CAPTCHA attempt failed (but I know how common that can be!). Sorry it affected your training session. :(
I see that Whatamidoing already replied to your post at WP:VEF
Quiddity
On Mon, Aug 17, 2015 at 10:58 AM, Sam Katz <smkatz@gmail.com mailto:smkatz@gmail.com > wrote:
No, you don't need CAPTCHA. You can use Honeypot. I think that would fix it.
On Mon, Aug 17, 2015 at 9:44 AM, Jonathan Morgan <jmorgan@wikimedia.org mailto:jmorgan@wikimedia.org > wrote:
No, I'm not aware of any ongoing CAPTCHA work. There was a long thread on wikitech-l starting last December ("Our CAPTCHA is very unfriendly") that resulted in some Phabricator tasks and a wikipage. But I don't know of any active development plans.
By the way: Aaron's in transition (of the timezone variety) right now, so it may be a day or so before he's able to respond to this thread.
- J
On Sun, Aug 16, 2015 at 10:53 PM, Pine W <wiki.pine@gmail.com mailto:wiki.pine@gmail.com > wrote:
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine
On Aug 16, 2015 10:41 PM, "WereSpielChequers" <werespielchequers@gmail.com mailto:werespielchequers@gmail.com > wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond <kerry.raymond@gmail.com mailto:kerry.raymond@gmail.com > wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
_______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org mailto: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 mailto:Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Hey folks!
I'm not running any experiments with VE right now or I would have announced it on the village pump. However, the VE team is doing a staged deploy of VE to new accounts at the moment -- slowly ramping up deployments from 5% to 100%. That looks a lot like an experiment because its similar to the strategy we used to bucket users during my experiments.
Thanks to J-Mo for pointing out my travels yesterday! I'm traveling again through the rest of the week for OpenSym, so I may go silent again for a few days.
-Aaron
On Mon, Aug 17, 2015 at 6:40 PM, Kerry Raymond kerry.raymond@gmail.com wrote:
Thanks, Nick! I will pass on the news!
Kerry
*From:* wiki-research-l-bounces@lists.wikimedia.org [mailto: wiki-research-l-bounces@lists.wikimedia.org] *On Behalf Of *Nick Wilson (Quiddity) *Sent:* Tuesday, 18 August 2015 8:47 AM *To:* Research into Wikimedia content and communities < wiki-research-l@lists.wikimedia.org> *Subject:* Re: [Wiki-research-l] Visual Editor experiment might have a problem ...
Hi Kerry,
IIUC, the original bug was https://phabricator.wikimedia.org/T109113 ("Can't successfully complete a second CAPTCHA in VisualEditor's save dialog") which started last Thursday, and was fixed this morning (and the fix should now already be deployed to the production wikis). It was only an issue if the first CAPTCHA attempt failed (but I know how common that can be!). Sorry it affected your training session. :(
I see that Whatamidoing already replied to your post at WP:VEF
Quiddity
On Mon, Aug 17, 2015 at 10:58 AM, Sam Katz smkatz@gmail.com wrote:
No, you don't need CAPTCHA. You can use Honeypot. I think that would fix it.
On Mon, Aug 17, 2015 at 9:44 AM, Jonathan Morgan jmorgan@wikimedia.org wrote:
No, I'm not aware of any ongoing CAPTCHA work. There was a long thread on wikitech-l starting last December ("Our CAPTCHA is very unfriendly") that resulted in some Phabricator tasks and a wikipage. But I don't know of any active development plans.
By the way: Aaron's in transition (of the timezone variety) right now, so it may be a day or so before he's able to respond to this thread.
- J
On Sun, Aug 16, 2015 at 10:53 PM, Pine W wiki.pine@gmail.com wrote:
I thought that someone was already working on CAPTCHA improvents, but I can't remember who, and I haven't heard anything recently which makes me wonder if this was de-prioritized.
J-mo, do you happen to know the status of the CAPTCHA work is, and who, if anyone, is active on that project?
Pine
On Aug 16, 2015 10:41 PM, "WereSpielChequers" werespielchequers@gmail.com wrote:
Hi Kerry,
there is an experiment going on that randomly opts half if new users into V/E and leaves half using the classic editor. That should account for why one of your newbies had been opted in but not the other.
Captcha when adding citations is a longstanding problem, we need Captcha on account creation to keep the spam bots at bay, but somehow it also applies to newbies adding external links as cites, so we have a software feature that doesn't effect the vandals but instead targets the best of our newbies. My suspicion is that if we could work out when that was introduced and then compare it to subsequent recruitment and retention we would find that this was one of the most damaging mistakes we've made.
Regards
Jonathan
On 17 Aug 2015, at 04:56, Kerry Raymond kerry.raymond@gmail.com wrote:
I ran my first training session using the Visual Editor this morning and hit what appeared to be a show-shopping bug. It appeared that the two new users (thankfully I had only 2) could not create a citation. They found themselves in an infinite loop of Save Page with Capcha when they tried to create a citation.
By the end of the session, I managed to refine the bug to a combination of “new user”, “new article” (although created by me, not the new users), and citations involving a live URL, duly reported at https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#New_users_unab...
Ironically it first happened on their newly created User Pages where we were practising our new Wikipedia skills because tackling “real articles”. Then on the “real articles” I had created earlier for them to use (a training approach that has the benefit of not unleashing a horde of angry watchlisters when they make some silly mistake, which occurs if you let new people make their early edits on “popular articles”). (Spot the pattern, both were new articles!).
Now if this had happened to a new user sitting at home, they would have been stymied. Because I was there to hold their hands in a training setting, I found a way around the problem by logging them in as me and we continued the training session on that basis (but not an option to the user sitting at home frustratingly typing in Capcha responses until they got frustrated and walked away).
So, Aaron, it may be that your research on the impact of the VE was impacted by this bug. I imagine that users affected would have eventually aborted the edit as they were unable to save, unless by chance they were able to realise that the problem was caused by their citation and either removed the citation and just saved the text changes. It’s hard to say what the likelihood of a new user being affected is, as the problem seemed to relate to the age of the article (I am autopatrolled so I don’t think the new articles would have any “might be dodgy” status flags on them, but I am not familiar with how that side of things works).
Also, is this experiment (or one similar) currently running? It’s just that when we went into the Preferences of the two new user accounts to enable the VE, one of them already had it enabled (yet I had seen both new user accounts created in front of me a couple of minutes earlier), so there was no possibility that this was anything other than a default setting for one of the two users. I thought enabling the VE was normally strictly opt-in?
Kerry
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
User:Jmorgan (WMF) https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)
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
--
Nick Wilson (Quiddity) Community Liaison 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@lists.wikimedia.org