I would have put this in the MediaWiki software subforum, but nobody ever reads anything there, so here it shall go.

The basic unit of the MediaWiki platform is the "revision," each of which has several attributes, including a timestamp, an editor, a page, a summary, and all those things we all associate with an individual "edit".

I'm wondering if any of the current or former admins here know whether changes to a revision's attributes are among the tools available to users at any level (other than database admin).

If so, is a record kept?

To illustrate an example, as mentioned here in another thread, Jayjg has made several "oops" IP edits over the years, most of which were later Oversighted away. Some, on the other hand, were later changed to appear to have been made by Jayjg.

Here's a good example.

In the 1/25/2006 and 8/16/2006 database dumps, that revision is attributed to 209.47.33.228 (consistent with the signature). However between 8/17/2006 and the 11/30/2006 dump, the revision's editor attribute is changed to Jayjg, as you see it remains today.

I'm finding quite a few of these changes, by the way, and just as with this example, the article logs reveal nothing.

I guess my questions are: who has permission to do this sort of thing, is it at all common, and is there any transparency/accountability associated with it?

Maybe to put a finer point on it: could a random (non-Jayjg) editor request that unintentional (or intentional fore that matter) IP edits be retroactively altered to reflect his/her having been the editor?

Used to be possible at http://en.wikipedia.org/wiki/Wikipedia:Cha...ion_for_an_edit

But as I see they've deactivated that page.

There are tools for devs to do it.

A much more interesting question to ask — as I'm sure you realize, you sly dawg — is whether editors with similar tools could rewrite revision histories to attribute any revision they wish to any editor they wish.

And the answer is apparently — as I'm sure you realize, you sly dawg — «yes, why the hell not?».

Geobytes IpLocator puts 209.47.33.228 in Toronto, for whatever that's worth, «not much», says Jonny from Saint Louey.

Jonny

More than Toronto, it's the IP belonging to Gennum, which is where Jayjg appears to have been employed at the time. In fact, his first known edits to Wikipedia were made by that IP, which remains stable until November of 2004, and then changes to 216.94.22.2, which also belongs to Gennum.

Don't bother looking...they've all been Oversighted.

Only devs can do it.

So apparently, this was a service extended to Jayjg on a basis other than that available to other users.

Interesting.

Here's another scenario I need help understanding:

Examining Jayjg's first 32,414 edits, one finds that a few dozen pair have exactly the same timestamp (the database records the time of each edit to the second, while online you can only see them tracked to the nearest minute).

It turns out this is expected in the case of some special non-edit actions, such as moves, redirects and possibly some batch reverts. Applying the most liberal standard to Jayjg's identically timestamped edits, we can eliminate all but six pairs.

Six pairs represent just 0.04% of Jayjg's first 32,414 edits, but the fact that there are any identically timestamped pairs of edits seems impossible.

My question is, how likely is it that a database under heavy load could delay writing some edits such that they appear to have been made at exactly the same moment as others?

I've examined thousands of random edits, and while it's fairly common to find two or more edits with the same timestamp out there in the wild, I've yet to see more than one from the same editor (except for moves, etc, as described earlier).

The following six pairs of edits by Jayjg are the exceptions (all times UTC):

01/28/2005 15:54:08
01/28/2005 15:54:08

03/01/2005 16:01:43
03/01/2005 16:01:43

03/14/2005 03:01:45
03/14/2005 03:01:45

03/29/2005 18:24:50
03/29/2005 18:24:50

07/04/2005 17:36:29
07/04/2005 17:36:29

08/14/2005 21:00:03
08/14/2005 21:00:03 (article deleted)

What's going on?

Jayjg = Cecil Rhodes

Jonny

Wacky! If (like you say) the timestamps match down to the second, I guess he could have been using two computers simultaneously, and just fooling around to see what would happen if he clicked the "Save" button on both at exactly the same time... Or else there might have been two separate people using the account simultaneously on different machines.

I wonder if it's possible that two edits from the same user, posted within a minute or so of each other, might have been processed by two completely different database servers, operating in some sort of "clustered" environment, but with their system clocks set just slightly off from each other? That seems more far-fetched to me, but I'm not used to working with large databases requiring that kind of throughput.

Hey, I got to use the word "throughput" in a real sentence! Finally! Now I can die happy...

Jay did something weird with his edit history, I remember it changing and some early edits then appeared that were actually made much later.

My vote goes towards nothing nefarious. In the first pair I checked, Jay appears to have just quickly corrected a comment typo. Based on my own Wikidays, I'm guessing he submitted both edits shortly behind each other, and the first was held up by what I often saw when submitting posts, a delay in WP's acceptance of the submission due to technical difficulties.

Sometimes I'd submit an edit and watch the submission sit and spin for a minute or two while WP got its shit straight. I'm guessing the second was submitted in real time a few seconds behind the first, but the first was processed simultaneously when the WP bottleneck cleared.

Geez... WR has a reputation to uphold, which won't be served by anybody doing something so inappropriate as to say that anything whatsoever that ever happened on Wikipedia, particularly involving one of the disliked editors/admins, is "nothing nefarious"! Think of the children!

Do we know who the devs with this ability were during the first half of 2006? Would Danny Wool have been one, in theory (ignoring his technical ability)?

He isn't listed on the developers page on meta, and I'm not sure you can ignore his technical ability, or lack thereof... he doesn't strike me as the programmer type at all. Nor has he ever touched that page, according to the history on it.

Besides, why would he have to be a developer? Especially during most of 2006 - he was basically Jimbo's right-hand man for much of that year, and could have gotten somebody to do just about anything he wanted done, even under a questionable pretext.

Many strange things started happening after Danny arrived in October of 2005. Of course strange things continue (on steroids) even in his absence, but I'm noting enough oddness to propose a post hoc ergo propter hoc scenario.
Does anybody know anything else about how Danny came to be appointed to that spot?

What do you mean he arrived in October of 2005? Danny was on Wikipedia way before October 2005. His edits go all the way back to February 2002.

I think he meant when he became an employee of the Wikimedia foundation?

Could be. That was announced September 18, 2005. Close enough to October 2005.

Speaking of oversighting, it's still happening with SlimVirgin. I discovered tonight that the entry of Slim starting the stub for my bio on 2005-09-28 is now gone from her edit summaries. It's on the dump of her edit summaries that I made last August 6, and which is searchable here.

I thought that by making a record of Slim's edit history on August 6, they'd stop mucking around with oversighting her edits. Duh! No more "assuming good faith" from me! A small fire is still burning at that wonderful new Library of Alexandria, and it's being fed by a handful of spooky librarians huddled in the basement.

Actually it's because Doc glasgow deleted the redirect that contained the full edit history of the article.
*20:27, 1 December 2007 Doc glasgow (Talk | contribs) deleted "Daniel Brandt" ‎ (unneccessary redirect - privacy reasons. See me for details)

p.s. you might want to take another look at the PIR articles.

Okay, I knew Doc did this and I'm very happy about it. But I didn't realize that a user's contribution history was derived from the articles and pages themselves that a user edited. I thought it was in a separate database tied to that user, in which case it would be a more accurate and valuable record. (We can't have that, now, can we?)

Pretending for a moment that it was available to all people
wishing to protect their privacy (not for GFDL violation, Jonny),
would you want it to be transparent? It's sort of hard to
protect a person's privacy when people are staring at you.

I do not believe Jayjg's IP address or location is any of our
business. However, it is interesting that his privacy is protected,
while admins are permitted to violate random people's privacy in
the course of COI and sockpuppetry investigations.

Welcome to the world of Courtesy Blanking.

It's just their little way of adding insult to injury.

Jonny

Personally I don't find the presence or absence of transparency to be the issue. It is, as they increasingly say, what it is.

What I care about is having it, whatever it is, applied to everybody. No questions.

Through a comparison of my trusty database dumps, I've confirmed that after the practice of reassigning IP edits to user names was abandoned near the end of 2005, it happened exactly two more times: once in the case of a guy who begged and begged (and a record of his having done so exists on WP) and Jayjg, for whom no record exists.

The implications of this are sort of frightening, if you think about it. A database admin apparently finds it acceptable to alter the Wikipedia database for the benefit of a certain user -- the same one who Jimbo had previously appointed to the ArbCom after he was not elected -- in such a way that hides the act from the rest of us.

Very, very strange the way the usual rules fail to apply to one Wikipedian in particular, and how large the pile of discarded (banned) users at his feet is.

That, AB, is the issue.

And *this* is shocking and new? They failed to created any kind of mechanism for normal, customer service-related corrective action (as would a normal professional institution) and they "do" provide that service if you have an "in" or you can pull some strings at the top, or if you file a credible lawsuit.

This is how they operate.

Sorry to be a bit harsh. But I got over my shock and awe a long time ago with this gang of thugs.



I thought that IP edits aren't meant to "stick" and eventually get cleared out in general. But in principal, any login indicator could be given the "right to disappear" and renamed and refactored under another login indicator. It's just more complicated is all.

[quote name='WordBomb' date='Fri 9th November 2007, 7:29pm' post='59566']
I guess my questions are: who has permission to do this sort of thing, is it at all common, and is there any transparency/accountability associated with it?
[quote name='WordBomb' date='Fri 9th November 2007, 7:29pm' post='59566']
Maybe to put a finer point on it: could a random (non-Jayjg) editor request that unintentional (or intentional fore that matter) IP edits be retroactively altered to reflect his/her having been the editor?
[/quote]I have no idea, but why should this not be possible? Name changes are possible, so why not? Its just that they often don't respond to people, either because no one cares, or as in your case, they want to spite you.

As far as time-stamping is concerned, one of the modalities of cyberattackers or hackers, if you will, is to alter system files and restamp the time on them so it is not possible to detect easily the new file additions. One of the Microsoft security guys, I think it was Steve Reilly said that if you've been intruded upon, you can't even trust your own audit log, for precisely this same reason.

And if that's so, then what would prevent a systems person at Wikipedia from altering their own time settings, and making changes - if they really wanted to? If Jayg wanted to do this, the biggest obstacle would be to convince someone that it was worth the bother. Technically, changing time is always possible. I'm not a systems' person, but this seems evident.

I'm on the side of protecting people's privacy. The easiest way
is to just use proxies, so if you try to edit while logged out you
are either blocked because the IP is softblocked, or the edit is
attributed to the proxy if it is not blocked at all. Of course,
even in the latter case, one might still want one's GFDL
attribution under one's name or pseudonym, but at least one's
privacy is intact (depending on how well the proxy protects
ones privacy... which in the case of single-hop proxies is not
that well).

However, as WP hardblocks proxies, they really ought to do
this for people who accidentally edit while logged out (or
intentionally, for that matter, and then change their mind), as
opposed to drawing attention to the matter and screaming
COI based on nothing but geographic proximity.

I don't begrudge Jayjg his privacy, and I hope you or Somey
will consider deleting some of his personal information,
especially if he or one of his friends asks for that. However,
I also think the service should be available to everyone. If it
makes you feel better, think of it as a way of showing you
have more respect for privacy than WP.

You can get some measure of the discrepancy between actual contributions and contributions minus deleted edits by using the ContribCounter Tool.

For example, ContribCounter&titles=User:Jon Awbrey credits me with 13,224 edits while the Interiot Tool in my signature shows an ever decreasing number of edits (currently 13,050) due to deletions of articles that I wrote. Many of these deletions represent GFDL violations due to Internal Plagiarism (WP:IP).

Jon Awbrey

Copy Of Output

Ah, a WP:JZG initiated action.

No, that has no effect on the Jon Awbrey account, as JzG is doing Delete And Restore (WP:DAR) on altogether different accounts.

The Internal Plagiarism occurs when a Wikpedia Admin cuts-&-pastes content into another article and then deletes the plagiarized article along with its edit history. For examples, see this thread.

Jon Awbrey

Could we seriously redact Jayjg's IP address/physical location,
if for no other reason than to be better than WP is? Yes, I know
perfectly well that some people at WP expose other people's
personal information without remorse. And, so far as I know,
that isn't Jayjg's fault.

Anyway, if the point is just that WP likes Jayjg well enough to
hide that information, not that Jayjg did anything wrong, I think
we can do without those details?

You can always save the information somewhere private in case
you need it at a later date.

What is Jayjg's physical location? His IP address doesn't tell me, the average Joe, much of anything.

I don't think we're doing this because WP tries not to. We're doing it because it's just part of the conversation about what the heck is going on with that Jayjg "critter" (as Joseph would say). The mainstream media has reported on IP addresses often and unabashedly, thanks to WikiScanner. That tool was set up for little other reason than to embarrass "outsiders" who were editing Wikipedia. Then WikiScanner helped us to find things like Guy "JzG" Chapman's predilection for massive garbonzos, and I feel no ethical impulse to bury that "finding" any more than the New York Times felt it should bury Diebold's editing habits.

The problem is only accelerated every time we remind ourselves that Jimbo Wales promised the "community" months ago that Jayjg's habits were being privately reviewed, and that anything of substance that came of that discussion would be communicated publicly back to the community. Of course, that was just another Jim-bo Fib-bo, since we have heard nothing but chirping crickets in the Jayjg department from Jimbo's keyboard for months now.

Yet, Jayjg is back to work, asking his buddies to "cover his back" while he POV's another Judaism article.

That's not an environment where WE should be tiptoeing around protecting the identifying information of the "critter".

Joseph has given me a new word.

Greg

I won't repeat it, but Jayjg's general geographic location, assuming
that was not a proxy, is stated in this thread.



Did Jayjg or Jzg set up WikiScanner? Do they make a habit of
exposing people's personal information?

(You realise IP addresses do not correspond 1:1 to human
beings, right? So it is possible someone JzG is living with,
rather than JzG himself, is interested in certain models.)



Y'know what? I don't care about WP enough to care one
way or the other about the neutrality of the Judaism article.
I do, however, care about people, including people's
privacy.

Anyway, unless his IP is in Israel, I don't think it helps you
much.

At least move it off to one of the private forums, and redact
it here.

Actually, Guy goes out of his way to name "banned" users by both first and last name on Wikipedia, and on the mail lists. I used to very much like Guy, but his blind zealotry is hurting people.



Curiously, though, Wikipedia and it's editors DON'T care about privacy, unless it's the privacy of a Wikipedia editor in good standing. Why are Wikipedia editors more special or important than anyone else? Who appointed them? What gives any anonymous person more of a right to privacy, because they happen to edit a random website?



Please don't redact anything anymore.

Well, I am not WP or one of its editors.

I like the idea of complaining about them violating people's privacy, except
for the slight problem that if we linked to such situations, we'd be furthering
the privacy violations. : (

Anyway, not *all* WP editors care about WP more than people's privacy,
just some of them... including a number of the more powerful ones. ;_;



Fortunately for you, I'm not a moderator.

Yes, it's called Checkuser, Jayjg's own special wikiscanner that strangely he had access to before anyone else even knew about it.