2009/8/31 Brion Vibber brion@wikimedia.org:
On 8/31/09 7:35 AM, Michael Peel wrote: We've been planning to get a test setup together since conversations at the Berlin developer meetup in April, but actual implementation of it is pending coordination with Luca and his team.
My understanding is that work has proceeded pretty well on setting it up to be able to fetch page history data more cleanly internally, which was a prerequisite, so we're hoping to get that going this fall.
To add to what Brion said: The author of the Wired story, Hadley Leggett, scheduled a call with me earlier this month, but she missed the call. I didn't have time to follow up with her after that, and she filed the story without it. This is why there's no WMF quote in the story.
The gist of it is that:
We're very interested in WikiTrust, primarily for two reasons:
- it allows us to create blamemaps for history pages, so that you can quickly see who added a specific piece of text. This is very interesting for anyone who's ever tried to navigate a long version history to find out who added something.
- it potentially allows us to come up with an algorithmic "best recent revision" guess. This is very useful for offline exports.
The trust coloring is clearly the most controversial part of the technology. However, it's also integral to it, and we think it could be valuable. If we do integrate it, it would likely be initially as a user preference. (And of course no view of the article would have it toggled on by default.) There may also be additional community consultation required.
Any integration is contingent on the readiness of the technology. It seems to have matured over the last couple of years, and we're planning to meet with Luca soon to review the current state of things. There's no fixed deployment roadmap yet, and the deployment of FlaggedRevs is our #1 priority.
On Mon, Aug 31, 2009 at 6:13 PM, Erik Moeller erik@wikimedia.org wrote:
- it allows us to create blamemaps for history pages, so that you can
quickly see who added a specific piece of text. This is very interesting for anyone who's ever tried to navigate a long version history to find out who added something.
Yes. Incredibly useful. What I'd like would be when colors are shown, if you hover over some text it pops up a hover of the user who wrote it and when it was written (the revision). This would be a bit like the way Google Translate pops up the source text. I'd position the popup far left or far right of the window though so it doesn't obscure the text and annoy one so much (or have positioning be an option).
- it potentially allows us to come up with an algorithmic "best recent
revision" guess. This is very useful for offline exports.
Makes sense. Also good for anti-vandalism work.
The trust coloring is clearly the most controversial part of the technology. However, it's also integral to it, and we think it could be valuable. If we do integrate it, it would likely be initially as a user preference. (And of course no view of the article would have it toggled on by default.) There may also be additional community consultation required.
A show/hide button on the screen, with "default status" in preferences, please. And maybe an interface issue to consider, having a narrow top bar that doesn't scroll, where status, flagged revision etc info can be put that will always be visible no matter where you are in the article.
FT2
2009/8/31 FT2 ft2.wiki@gmail.com:
Yes. Incredibly useful. What I'd like would be when colors are shown, if you hover over some text it pops up a hover of the user who wrote it and when it was written (the revision).
A simple version of that is already implemented. Go to
http://wikitrust.soe.ucsc.edu/index.php/Main_Page
and click the "check text" tab to see it, hover over a piece of text, and click it. The hover shows the username, and by clicking it, you'll get a diff. (This may not be the latest code.)
A show/hide button on the screen, with "default status" in preferences, please. And maybe an interface issue to consider, having a narrow top bar that doesn't scroll, where status, flagged revision etc info can be put that will always be visible no matter where you are in the article.
There's definitely a need to consolidate the FlaggedRevs revision tag indicator with any WikiTrust UI elements.
On Mon, Aug 31, 2009 at 12:33, Erik Moeller erik@wikimedia.org wrote:
A simple version of that is already implemented. Go to
http://wikitrust.soe.ucsc.edu/index.php/Main_Page
and click the "check text" tab to see it, hover over a piece of text, and click it. The hover shows the username, and by clicking it, you'll get a diff. (This may not be the latest code.)
According to their Wikimania presentation (hopefully available soon on Commons), they've also prepared a Firefox add-on, WikiTrust, which adds a new "trust info" tab to the top of mainspace articles. The trust info database is still being populated, though, so the trust info itself may be a little skewed; at the presentation they estimated that the English Wikipedia trust info database would be finished in about a month. (Their existing algorithm is language-independent, so presumably the add-on will work for non-English wikis as well.)
The Firefox add-on is still classified as "experimental", but adventurous persons can still get it at < https://addons.mozilla.org/en-US/firefox/addon/11087%3E.
On Mon, Aug 31, 2009 at 6:33 PM, Erik Moeller erik@wikimedia.org wrote:
2009/8/31 FT2 ft2.wiki@gmail.com:
Yes. Incredibly useful. What I'd like would be when colors are shown, if
you
hover over some text it pops up a hover of the user who wrote it and when
it
was written (the revision).
A simple version of that is already implemented. Go to
http://wikitrust.soe.ucsc.edu/index.php/Main_Page
and click the "check text" tab to see it, hover over a piece of text, and click it. The hover shows the username, and by clicking it, you'll get a diff. (This may not be the latest code.)
A show/hide button on the screen, with "default status" in preferences, please. And maybe an interface issue to consider, having a narrow top bar that doesn't scroll, where status, flagged revision etc info can be put
that
will always be visible no matter where you are in the article.
There's definitely a need to consolidate the FlaggedRevs revision tag indicator with any WikiTrust UI elements. -- Erik Möller Deputy Director, Wikimedia Foundation
Added a note on it here: < http://usability.wikimedia.org/wiki/Talk:Releases/Acai#Mini_toolbar_idea%3E
FT2
I am a little concerned that we are adopting a metric into our interface without adequate testing. Quality or trust in an article is not a simple numerical matter, much less a rough scale of a few categories. it will take a lot of experimentation with it until the rest of us can decide if its valid enough to be part of our actual interface--this is a decision that needs to be made by each community, and I hope it will be made carefully, before we commit to it.
'What people may want to use as an add on is their affair--what we offer to them as a gadget is something else. I'm not sure we have any formal method for approving them, but we ought to. The WMF should not be prescribing it for us.
That this should be done at the same time as the flagged revisions test is yet another complication.
David Goodman, Ph.D, M.L.S. http://en.wikipedia.org/wiki/User_talk:DGG
On Mon, Aug 31, 2009 at 2:15 PM, FT2ft2.wiki@gmail.com wrote:
On Mon, Aug 31, 2009 at 6:33 PM, Erik Moeller erik@wikimedia.org wrote:
2009/8/31 FT2 ft2.wiki@gmail.com:
Yes. Incredibly useful. What I'd like would be when colors are shown, if
you
hover over some text it pops up a hover of the user who wrote it and when
it
was written (the revision).
A simple version of that is already implemented. Go to
http://wikitrust.soe.ucsc.edu/index.php/Main_Page
and click the "check text" tab to see it, hover over a piece of text, and click it. The hover shows the username, and by clicking it, you'll get a diff. (This may not be the latest code.)
A show/hide button on the screen, with "default status" in preferences, please. And maybe an interface issue to consider, having a narrow top bar that doesn't scroll, where status, flagged revision etc info can be put
that
will always be visible no matter where you are in the article.
There's definitely a need to consolidate the FlaggedRevs revision tag indicator with any WikiTrust UI elements. -- Erik Möller Deputy Director, Wikimedia Foundation
Added a note on it here: < http://usability.wikimedia.org/wiki/Talk:Releases/Acai#Mini_toolbar_idea%3E
FT2 _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
True. The moment you give people a tool, many people will simplistically assume what it does or rely unthinkingly on it.
- WikiTrust might be described as "a way to see how long an edit endured and how much trust it seems to have"; in most users' hands it'll be "its colored red/blue so its right/wrong." - People won't think, they'll assume and rely.
If it is introduced, then I would suggest introducing it as a gadget for admins and experienced users, a limited number at first. Communally, it shouldn't be available to all, but to those who request it and seem to understand what it shows and how to interpret it (perhaps package it with rollback or something that gets a little scrutiny of their cluefulness?)
Thats for the future, but no harm thinking ahead. Very wary of what people will assume it means, and that we're clear it is a tool that needs considerable experienced interpretation and is *misleading *without it.
FT2
On Mon, Aug 31, 2009 at 7:23 PM, David Goodman dgoodmanny@gmail.com wrote:
I am a little concerned that we are adopting a metric into our interface without adequate testing. Quality or trust in an article is not a simple numerical matter, much less a rough scale of a few categories. it will take a lot of experimentation with it until the rest of us can decide if its valid enough to be part of our actual interface--this is a decision that needs to be made by each community, and I hope it will be made carefully, before we commit to it.
'What people may want to use as an add on is their affair--what we offer to them as a gadget is something else. I'm not sure we have any formal method for approving them, but we ought to. The WMF should not be prescribing it for us.
That this should be done at the same time as the flagged revisions test is yet another complication.
David Goodman, Ph.D, M.L.S. http://en.wikipedia.org/wiki/User_talk:DGG
On Mon, Aug 31, 2009 at 2:15 PM, FT2ft2.wiki@gmail.com wrote:
On Mon, Aug 31, 2009 at 6:33 PM, Erik Moeller erik@wikimedia.org
wrote:
2009/8/31 FT2 ft2.wiki@gmail.com:
Yes. Incredibly useful. What I'd like would be when colors are shown,
if
you
hover over some text it pops up a hover of the user who wrote it and
when
it
was written (the revision).
A simple version of that is already implemented. Go to
http://wikitrust.soe.ucsc.edu/index.php/Main_Page
and click the "check text" tab to see it, hover over a piece of text, and click it. The hover shows the username, and by clicking it, you'll get a diff. (This may not be the latest code.)
A show/hide button on the screen, with "default status" in
preferences,
please. And maybe an interface issue to consider, having a narrow top
bar
that doesn't scroll, where status, flagged revision etc info can be
put
that
will always be visible no matter where you are in the article.
There's definitely a need to consolidate the FlaggedRevs revision tag indicator with any WikiTrust UI elements. -- Erik Möller Deputy Director, Wikimedia Foundation
Added a note on it here: < http://usability.wikimedia.org/wiki/Talk:Releases/Acai#Mini_toolbar_idea
FT2 _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On Mon, Aug 31, 2009 at 7:46 PM, FT2ft2.wiki@gmail.com wrote:
<snip>
If it is introduced, then I would suggest introducing it as a gadget for admins and experienced users, a limited number at first. Communally, it shouldn't be available to all, but to those who request it and seem to understand what it shows and how to interpret it (perhaps package it with rollback or something that gets a little scrutiny of their cluefulness?)
I'm not sure why this would be restricted to admins and experienced users. Does *anyone* else support this view? It would be more logical, in my view, to either have it or not have it, not some halfway house.
Thats for the future, but no harm thinking ahead. Very wary of what people will assume it means, and that we're clear it is a tool that needs considerable experienced interpretation and is *misleading *without it.
Admins and experienced users would be just as capable of misinterpreting it.
Carcharoth
Not saying I disagree with you, but with that in mind and looking at the test example, I'd say that the more useful concept isn't the ability to rate editors - which I could do without, it's a little too anti-AGF imho - but its usefulness as a metric of how many people have edited a particular section. That would give every sentence some measure of "Wiki-ness," how much it had been edited mercilessly. I for one take dislike seeing masses of paragraphs written by one or two people and would be far more likely to comb through and copyedit or look for things within that section, no matter who wrote it, than one with 10 or 20 editors contributing. That's the part that would be worthwhile.
~A
On Mon, Aug 31, 2009 at 14:46, FT2 ft2.wiki@gmail.com wrote:
Very wary of what people will assume it means, and that we're clear it is a tool that needs considerable experienced interpretation and is *misleading *without it.
FT2
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
2009/9/1 FT2 ft2.wiki@gmail.com:
I think there's a terminology issue. We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much. Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
Call it "Wikidrama" or "wikimyspace" instead? ;-)
Seriously, you need to propose the name change to Luca and team. The Wired article is nice publicity for them, but should show them what an epic disaster the name could be.
- d.
That's a very good idea.
Carcharoth
On Tue, Sep 1, 2009 at 11:36 AM, FT2ft2.wiki@gmail.com wrote:
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On Tue, Sep 1, 2009 at 7:26 AM, Carcharothcarcharothwp@googlemail.com wrote:
That's a very good idea.
+1
The name strikes me as the biggest drawback of the current system.
Carcharoth
On Tue, Sep 1, 2009 at 11:36 AM, FT2ft2.wiki@gmail.com wrote:
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On Sat, Sep 5, 2009 at 12:33 AM, Samuel Kleinmeta.sj@gmail.com wrote:
The name strikes me as the biggest drawback of the current system.
I think de Alfaro put it well himself in his quote from Information Week:
'Despite its name, WikiTrust can't directly measure whether text is trustworthy. "It can only measure user agreement," said de Alfaro. "That's what it does." '
http://www.informationweek.com/news/internet/security/showArticle.jhtml?arti...
-Sage
I think there's a real risk here, to be even more blunt.
Calling it a trust system risks someone looking at a piece of text and saying "oh, look, this is trusted, so i can -rely on this as advice before doing something dangerous/in making a medical decision/etc" -use this as my sole source in writing my college paper" -take for granted the claim this text makes that a living person cheated on his spouse (or worse possibilities" -assume this means WP as a group/the foundation itself makes the claim that *I* cheated on someone" ... and then, when the claim proves to be false, become angry and go after the Foundation? Not necessarily legally, though.... I fear that if they make an assumption "this text is highlighted as high trust, so it can be trusted", and are told that this is the meaning on a help page, we could be liable.
Nathan
On Tue, Sep 1, 2009 at 6:36 AM, FT2ft2.wiki@gmail.com wrote:
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
... and then, when the claim proves to be false, become angry and go after the Foundation? Not necessarily legally, though.... I fear that if they make an assumption "this text is highlighted as high trust, so it can be trusted", and are told that this is the meaning on a help page, we could be liable.
Yet another one of my fears.
Emily On Sep 1, 2009, at 8:25 AM, Nathan Russell wrote:
I think there's a real risk here, to be even more blunt.
Calling it a trust system risks someone looking at a piece of text and saying "oh, look, this is trusted, so i can -rely on this as advice before doing something dangerous/in making a medical decision/etc" -use this as my sole source in writing my college paper" -take for granted the claim this text makes that a living person cheated on his spouse (or worse possibilities" -assume this means WP as a group/the foundation itself makes the claim that *I* cheated on someone" ... and then, when the claim proves to be false, become angry and go after the Foundation? Not necessarily legally, though.... I fear that if they make an assumption "this text is highlighted as high trust, so it can be trusted", and are told that this is the meaning on a help page, we could be liable.
Nathan
On Tue, Sep 1, 2009 at 6:36 AM, FT2ft2.wiki@gmail.com wrote:
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
The problem is that while "long-standing" and "apparently reputable author" correlate with "trust", they are not the same.
The perception that a measure of text source and historicity is in any way a measure of trust, is a misconception we have to kill at root, burn, salt over, mercilessly counter, and also impale all those who defile it. And generally destroy it with prejudice.
Because we dare not allow that gadget to be misinterpreted that way (even if in knowing hands it can indeed indicate trust or doubt). It's very tempting, so people will, and they'll read it is in the media... so we have to bludgeon home it ISN'T.
(There would have been a "graphic imagery spoiler", but we deleted spoilers ages ago)
FT2
On Tue, Sep 1, 2009 at 5:39 PM, Emily Monroebluecaliocean@me.com wrote:
... and then, when the claim proves to be false, become angry and go after the Foundation? Not necessarily legally, though.... I fear that if they make an assumption "this text is highlighted as high trust, so it can be trusted", and are told that this is the meaning on a help page, we could be liable.
Yet another one of my fears.
Emily On Sep 1, 2009, at 8:25 AM, Nathan Russell wrote:
I think there's a real risk here, to be even more blunt.
Calling it a trust system risks someone looking at a piece of text and saying "oh, look, this is trusted, so i can -rely on this as advice before doing something dangerous/in making a medical decision/etc" -use this as my sole source in writing my college paper" -take for granted the claim this text makes that a living person cheated on his spouse (or worse possibilities" -assume this means WP as a group/the foundation itself makes the claim that *I* cheated on someone" ... and then, when the claim proves to be false, become angry and go after the Foundation? Not necessarily legally, though.... I fear that if they make an assumption "this text is highlighted as high trust, so it can be trusted", and are told that this is the meaning on a help page, we could be liable.
Nathan
On Tue, Sep 1, 2009 at 6:36 AM, FT2ft2.wiki@gmail.com wrote:
I think there's a terminology issue.
We cannot refer to this as a "trust" system, however "Wikitrust" brands it. We just can't. It misleads too many, and implies too much.
Call it a "text tracing system" or "a gadget to highlight text origins" instead. It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
FT2
On Mon, Aug 31, 2009 at 8:37 PM, James Alexander jamesofur@gmail.comwrote:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
On Mon, Aug 31, 2009 at 3:12 PM, David Gerard dgerard@gmail.com wrote:
2009/8/31 David Goodman dgoodmanny@gmail.com:
I am a little concerned that we are adopting a metric into our interface without adequate testing.
It appears we're not and Wired completely jumped the gun. There is no timeframe for release of this thing even as an optional extra.
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
-- James Alexander http://en.wikipedia.org/wiki/User:Jamesofur _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
I found an article recently that had a glaring bit of vandalism that had gone uncorrected for a long time. Does anyone here know if Wikitrust (or whatever it should be called instead) would have spotted this?
http://en.wikipedia.org/w/index.php?title=Thomas_Stevenson&diff=31269588...
The point here is that this sentence stayed in the article for a very long time. If that fools people into thinking that it is "OK", then that is not good. This sort of thing is rare, but if you look through the history starting from this edit forwards, you will see what happened there:
http://en.wikipedia.org/w/index.php?title=Thomas_Stevenson&diff=13239820...
A long series of vandal edits, and eventually someone only partially reverted.
Carcharoth
On Tue, Sep 1, 2009 at 8:36 PM, FT2ft2.wiki@gmail.com wrote:
It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
To be honest, what exactly is the point of this thing? I've seen this kind of thing a couple of times when academics have been doing research. But what's the use case? What are users supposed to do with the knowledge? Is it important? Should end-users care?
All I can see is a moderately handy tool for editors who do a lot of patrolling, to save them a bit of time. Other than that, it just makes the page text hard to read, imho.
Or have I missed some radical advancement in the tech?
Steve
I'd use it in a flash. I often find it helpful when examining an article (for edit warriors and vandals, or dodgy editorship), to trace back where a given wording was introduced.
I can also see it would be immensely useful to me, to be able to see which wordings were being warred over or changed recently and which were more stable or historically unchanged.
As I also know a number of users, it may further help me in evaluating a text, to have a quick way ("hover" information) to say "okay, these are texts introduced by users I know and consider decent responsible editors, so I don't have to spend time on them and can focus on these sections".
However I would be relying on my own experience and using it as a tool to assist and help me shortcut doing things I do already, not as a bible of reliability, a substitute for reliable sources, or as a measure of implicit trust.
FT2
On Tue, Sep 1, 2009 at 3:09 PM, Steve Bennett stevagewp@gmail.com wrote:
On Tue, Sep 1, 2009 at 8:36 PM, FT2ft2.wiki@gmail.com wrote:
It's a lot less glamorous, sounds alot less dramatic, doesn't get the dollars - but it's got zero capability of misleading.
To be honest, what exactly is the point of this thing? I've seen this kind of thing a couple of times when academics have been doing research. But what's the use case? What are users supposed to do with the knowledge? Is it important? Should end-users care?
All I can see is a moderately handy tool for editors who do a lot of patrolling, to save them a bit of time. Other than that, it just makes the page text hard to read, imho.
Or have I missed some radical advancement in the tech?
Steve
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
2009/8/31 James Alexander jamesofur@gmail.com:
How would the blame maps work with people editing around vandalism? For example someone either blanks the page or does extensive vandalism to it (especially over the course of a couple days or a couple users). I would imagine it would be fairly easy if the bad contributions just got rolledback but would the old blamemaps still be reinstated if someone went in and manually copy/pasted the old version (or something very close) in or would the system count it as a new contribution?
I haven't been following this in any particular detail - I have no intent of using the system! - but this is certainly an issue they have thought of and planned for; they currently describe the system as "robust to cut-and-pase, delete-and-reinsert, and most type of attacks..."
I suspect the papers linked at the bottom here - http://wikitrust.soe.ucsc.edu/index.php/Main_Page - go into a bit more detail about the algorithm.
- it allows us to create blamemaps for history pages, so that you
can quickly see who added a specific piece of text. This is very interesting for anyone who's ever tried to navigate a long version history to find out who added something.
I have to admit, I'd find this incredibly useful myself.
What I'd like would be when colors are shown, if you hover over some text it pops up a hover of the user who wrote it and when it was written (the revision).
I'd also find this useful.
Emily On Aug 31, 2009, at 12:26 PM, FT2 wrote:
On Mon, Aug 31, 2009 at 6:13 PM, Erik Moeller erik@wikimedia.org wrote:
- it allows us to create blamemaps for history pages, so that you can
quickly see who added a specific piece of text. This is very interesting for anyone who's ever tried to navigate a long version history to find out who added something.
Yes. Incredibly useful. What I'd like would be when colors are shown, if you hover over some text it pops up a hover of the user who wrote it and when it was written (the revision). This would be a bit like the way Google Translate pops up the source text. I'd position the popup far left or far right of the window though so it doesn't obscure the text and annoy one so much (or have positioning be an option).
- it potentially allows us to come up with an algorithmic "best
recent revision" guess. This is very useful for offline exports.
Makes sense. Also good for anti-vandalism work.
The trust coloring is clearly the most controversial part of the technology. However, it's also integral to it, and we think it could be valuable. If we do integrate it, it would likely be initially as a user preference. (And of course no view of the article would have it toggled on by default.) There may also be additional community consultation required.
A show/hide button on the screen, with "default status" in preferences, please. And maybe an interface issue to consider, having a narrow top bar that doesn't scroll, where status, flagged revision etc info can be put that will always be visible no matter where you are in the article.
FT2 _______________________________________________ WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
On 31/08/2009, Erik Moeller erik@wikimedia.org wrote:
The trust coloring is clearly the most controversial part of the technology. However, it's also integral to it, and we think it could be valuable. If we do integrate it, it would likely be initially as a user preference. (And of course no view of the article would have it toggled on by default.) There may also be additional community consultation required.
If I understand this correctly, wouldn't trust coloring inevitably mark all new users and anonymous IPs as untrustworthy?
So, basically, wouldn't trust coloring be a way of failing to assume good faith for all anonymous IPs and new users, and institutionalising this in the software?
-- Erik Möller Deputy Director, Wikimedia Foundation