Non-editor reverting to stable version

[Avi]

I know I'm dropping in a bit late, and perhaps this was already handled, but
while I was testing this evening, it seems to be that when a non-editor
reverts a page back to the last sighted version, it still reads current.

Wouldn't it make sense that if the version reverted to is in and of itself
sighted, that that should be reflected, regardless of the person performing
the revision?

Or am I missing something?

Thanks,

--Avi

[Gregory Maxwell]

Post by Avi
I know I'm dropping in a bit late, and perhaps this was already handled, but
while I was testing this evening, it seems to be that when a non-editor
reverts a page back to the last sighted version, it still reads current.
Wouldn't it make sense that if the version reverted to is in and of itself
sighted, that that should be reflected, regardless of the person performing
the revision?
Or am I missing something?

Imagine that a user reverts to a year old sighted version and we mark
the new version as sighted. This would result in a decreased amount
of review of the edit and as a result this bad change may go unnoticed
for a longer span of time. Clearly that isn't good.

What if we only preserve the flagging if they revert to the most
recent? There too we may miss the chance to catch a reversion of good
material. And in this case if the default view were the sighted
revision it moving the pointer really wouldn't help.

Do these points convince you that the current behavior is better than
your proposal?

[Aaron Schulz]

Also, when you revert, all of the current versions of the images/templates become the permanent ones for that stable version. So a user reverting to a good version with bad templates would be reviewed.

-Aaron Schulz

Date: Tue, 9 Oct 2007 00:23:42 -0400
Subject: Re: [Wikiquality-l] Non-editor reverting to stable version

Post by Avi
I know I'm dropping in a bit late, and perhaps this was already handled, but
while I was testing this evening, it seems to be that when a non-editor
reverts a page back to the last sighted version, it still reads current.
Wouldn't it make sense that if the version reverted to is in and of itself
sighted, that that should be reflected, regardless of the person performing
the revision?
Or am I missing something?

Imagine that a user reverts to a year old sighted version and we mark
the new version as sighted. This would result in a decreased amount
of review of the edit and as a result this bad change may go unnoticed
for a longer span of time. Clearly that isn't good.
What if we only preserve the flagging if they revert to the most
recent? There too we may miss the chance to catch a reversion of good
material. And in this case if the default view were the sighted
revision it moving the pointer really wouldn't help.
Do these points convince you that the current behavior is better than
your proposal?
_______________________________________________
Wikiquality-l mailing list
http://lists.wikimedia.org/mailman/listinfo/wikiquality-l

_________________________________________________________________
Boo! Scare away worms, viruses and so much more! Try Windows Live OneCare!
http://onecare.live.com/standard/en-us/purchase/trial.aspx?s_cid=wl_hotmailnews

[Avi]

Good points, Gregory.

Regarding reverting back to anything other than the most recent sighted
version, yes, I agree your point is absolutely convincing.

Regarding the possibility of missing potentially good edits when reverting
back to the most recent sighted version, I would suggest that it depends on
how far along the continuum between accuracy and completeness we are going
to choose to live.

If we wanted the best, most complete picture, then we should never show a
sighted version due to the risk of missing some good information (ala wiki
now). Of course, we do _not_ want that, otherwise we would not be having
this project. So, how much potential lost information is allowable and how
much is not? I am not certain that the risk of losing some of the most
recent information, which is still in the history, and which may be pure
vandalism, outweighs the apparent need to have a large selection of stable
articles.

Thanks,

--Avi

Post by Avi

Post by Avi
I know I'm dropping in a bit late, and perhaps this was already handled,

but

Post by Avi
while I was testing this evening, it seems to be that when a non-editor
reverts a page back to the last sighted version, it still reads current.
Wouldn't it make sense that if the version reverted to is in and of

itself

Post by Avi
sighted, that that should be reflected, regardless of the person

performing

Post by Avi
the revision?
Or am I missing something?

Imagine that a user reverts to a year old sighted version and we mark
the new version as sighted. This would result in a decreased amount
of review of the edit and as a result this bad change may go unnoticed
for a longer span of time. Clearly that isn't good.
What if we only preserve the flagging if they revert to the most
recent? There too we may miss the chance to catch a reversion of good
material. And in this case if the default view were the sighted
revision it moving the pointer really wouldn't help.
Do these points convince you that the current behavior is better than
your proposal?

[Erik Moeller]

Post by Avi
Regarding the possibility of missing potentially good edits when reverting
back to the most recent sighted version, I would suggest that it depends on
how far along the continuum between accuracy and completeness we are going
to choose to live.

Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change. Let's not
overestimate the significance of the flag; article histories & editing
capabilities continue to exist as before. And the cost of having to
re-screen after a simple revert (one of the most common operations) is
unacceptably high.

[Gregory Maxwell]

Post by Erik Moeller
Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change.

Reversion to an old version can be vandalism as much as the insertion
of new text. ... it depends on the context.

[Erik Moeller]

Post by Gregory Maxwell

Post by Erik Moeller
Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change.

Reversion to an old version can be vandalism as much as the insertion
of new text. ... it depends on the context.

The situation right now:

1) Trusted user A makes an edit.
2) Untrusted user B vandalizes.
3) Trusted user A reverts.
4) Trusted user A has to re-review after save, because the revert is
counted the same as any other change to an untrusted version.

This doesn't make sense; when a trusted user performs a revert to the
most recently screened version, the newly created version should be
sighted.

[Erik Moeller]

Post by Erik Moeller

Post by Gregory Maxwell

Post by Erik Moeller
Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change.

Reversion to an old version can be vandalism as much as the insertion
of new text. ... it depends on the context.

1) Trusted user A makes an edit.
2) Untrusted user B vandalizes.
3) Trusted user A reverts.
4) Trusted user A has to re-review after save, because the revert is
counted the same as any other change to an untrusted version.

Ah, I see now that Avi was actually talking about a different scenario
- sorry for not reading carefully. I agree that in the _non-editor_
scenario, the newly created version should not have the "sighted" flag
since we don't know anything about the true nature of the change.

[Aaron Schulz]

The untrusted users edit would not have been reviewed. So a good revert is all they need. When the next reviewer comes, it will prompt to review on edit with a diff. Since the bad stuff was reverted, their change would be the only think to review (cake).

-Aaron Schulz

Date: Tue, 9 Oct 2007 23:41:36 +0200
Subject: Re: [Wikiquality-l] Non-editor reverting to stable version

Post by Gregory Maxwell

Post by Erik Moeller
Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change.

Reversion to an old version can be vandalism as much as the insertion
of new text. ... it depends on the context.

1) Trusted user A makes an edit.
2) Untrusted user B vandalizes.
3) Trusted user A reverts.
4) Trusted user A has to re-review after save, because the revert is
counted the same as any other change to an untrusted version.
This doesn't make sense; when a trusted user performs a revert to the
most recently screened version, the newly created version should be
sighted.
--
Erik
DISCLAIMER: This message does not represent an official position of
the Wikimedia Foundation or its Board of Trustees.
_______________________________________________
Wikiquality-l mailing list
http://lists.wikimedia.org/mailman/listinfo/wikiquality-l

_________________________________________________________________
Windows Live Hotmail and Microsoft Office Outlook – together at last.  Get it now.
http://office.microsoft.com/en-us/outlook/HA102225181033.aspx?pid=CL100626971033

[Magnus Manske]

Post by Erik Moeller

Post by Gregory Maxwell

Post by Erik Moeller
Indeed. All that "sighted" says is that it's believed to be free of
vandalism, not that there might not be a useful change.

Reversion to an old version can be vandalism as much as the insertion
of new text. ... it depends on the context.

1) Trusted user A makes an edit.

To a reviewed version, I assume.

Post by Erik Moeller
2) Untrusted user B vandalizes.
3) Trusted user A reverts.
4) Trusted user A has to re-review after save, because the revert is
counted the same as any other change to an untrusted version.
This doesn't make sense; when a trusted user performs a revert to the
most recently screened version, the newly created version should be
sighted.

So, how about:
* An edit to a reviewed version by a trusted user automatically becomes reviewed
* A revert by a trusted user to a reviewed version becomes the latest
reviewed version

Problem solved. Or did I miss something?

Magnus

[P. Birken]

Post by Magnus Manske
* An edit to a reviewed version by a trusted user automatically becomes reviewed
* A revert by a trusted user to a reviewed version becomes the latest
reviewed version
Problem solved. Or did I miss something?

Mh, maybe ;-) The first thing is already implemented. The second thing
has the problem that the revert does not screen the templates.
However, I think the added efficiency is worth the risk and therefore,
this should still be implemented for rollbacks.

Bye,

Philipp

[Platonides]

Post by P. Birken

Post by Magnus Manske
* An edit to a reviewed version by a trusted user automatically becomes reviewed
* A revert by a trusted user to a reviewed version becomes the latest
reviewed version
Problem solved. Or did I miss something?

Mh, maybe ;-) The first thing is already implemented. The second thing
has the problem that the revert does not screen the templates.
However, I think the added efficiency is worth the risk and therefore,
this should still be implemented for rollbacks.
Bye,
Philipp

Why not auto-sight it but with the previous-sighted template-expansion?