Hi
I propose to add a new feature, to improve our fight again vandalism. I mean a diff, for watchlist articles, between the current version and the last read one.
I need opinion about the feature ? and about an elegant technical solution too ?
Emmanuel Engelhart [Kelson]
On Thu, 14 Oct 2004 17:00:54 +0200 Emmanuel Engelhart emmanuel@engelhart.org wrote:
Hi
I propose to add a new feature, to improve our fight again vandalism. I mean a diff, for watchlist articles, between the current version and the last read one.
I need opinion about the feature ? and about an elegant technical solution too ?
Nobody again... ?
Follows my technical proposition: - add a field "wl_validate" (containing an article version id) - add a tab to validate an article (shown only if the article belongs to the user watchlist) - add a link for each modified article in the watchlist to compare the current version with the last validated one.
Kelson
On Fri, 15 Oct 2004 23:40:52 +0200, Emmanuel Engelhart emmanuel@engelhart.org wrote:
On Thu, 14 Oct 2004 17:00:54 +0200 Emmanuel Engelhart emmanuel@engelhart.org wrote:
Hi
I propose to add a new feature, to improve our fight again vandalism. I mean a diff, for watchlist articles, between the current version and the last read one.
I need opinion about the feature ? and about an elegant technical solution too ?
Nobody again... ?
Follows my technical proposition:
- add a field "wl_validate" (containing an article version id)
- add a tab to validate an article (shown only if the article belongs to the user watchlist)
- add a link for each modified article in the watchlist to compare the current version with the last validated one.
Kelson
Kelson, this sounds like a fine idea; the lack of response may be because a different review system is already being tested out, and trying to implement more than one at a time would be a pain. Of course there are details to work out :
* once one person validates a version, do others still have the option to validate it as well? what if a vandal comes, changes a page on her watchlist, and immediately validates it?
* what do you do with versions that have been validated by N users? show them differently in RC or watchlists?
* Can users 'unvalidate' a version, or do they have to revert the last changes?
L'informatique, en tant que discipline, ne traite pas plus des ordinateurs que l'astronomie le fait les téléscopes. -- E. W. Dijkstra
On Sat, 16 Oct 2004 03:41:02 -0400 Sj 2.718281828@gmail.com wrote:
On Fri, 15 Oct 2004 23:40:52 +0200, Emmanuel Engelhart emmanuel@engelhart.org wrote:
On Thu, 14 Oct 2004 17:00:54 +0200 Emmanuel Engelhart emmanuel@engelhart.org wrote:
Hi
I propose to add a new feature, to improve our fight again vandalism. I mean a diff, for watchlist articles, between the current version and the last read one.
I need opinion about the feature ? and about an elegant technical solution too ?
Nobody again... ?
Follows my technical proposition:
- add a field "wl_validate" (containing an article version id)
- add a tab to validate an article (shown only if the article belongs to the user watchlist)
- add a link for each modified article in the watchlist to compare the current version with the last validated one.
Kelson
Kelson, this sounds like a fine idea; the lack of response may be because a different review system is already being tested out, and trying to implement more than one at a time would be a pain. Of course there are details to work out :
Ok, I guess you speak about the Magnus special page, coded last July. I didn't know anything about this feature, so I take a look at this feature. Magnus goal was to provide a voting system for each article. I just want to improve the tracking of modifications.
I think, it would be good for me to work with the same data/framework.
But, this tool takes vote for one article. I need to have a validation for one version. And that's a weakness of the current tool.
It would be great to store the article revision id in the validate table.
- once one person validates a version, do others still have the option
to validate it as well? what if a vandal comes, changes a page on her watchlist, and immediately validates it?
A user first validates an article for himself, to know which version was entirely reviewed by him. So in a far future, he can review the entire article again by showing simply the diff between the current version and the last validate one.
- what do you do with versions that have been validated by N users?
show them differently in RC or watchlists?
- Can users 'unvalidate' a version, or do they have to revert the last changes?
I guess the last answer avoids this question ?? I hope so ;-)
Best regards
Emmanuel Engelhart
On Sat, Oct 16, 2004 at 03:41:02AM -0400, Sj wrote:
On Fri, 15 Oct 2004 23:40:52 +0200, Emmanuel Engelhart Kelson, this sounds like a fine idea; the lack of response may be because a different review system is already being tested out, and trying to implement more than one at a time would be a pain. Of course there are details to work out :
If both are written as options it should be possible to have both.
- once one person validates a version, do others still have the option
to validate it as well? what if a vandal comes, changes a page on her watchlist, and immediately validates it?
don't validate your own edits, simple as that.
- what do you do with versions that have been validated by N users?
show them differently in RC or watchlists?
a validate_counter sounds good.
- Can users 'unvalidate' a version, or do they have to revert the last changes?
If a revert comes (only sysop can do that, no?) then the validate_counter is reset to 0
ciao, tom
On Saturday 16 October 2004 11:23, Thomas R. Koll wrote:
- once one person validates a version, do others still have the option
to validate it as well? what if a vandal comes, changes a page on her watchlist, and immediately validates it?
don't validate your own edits, simple as that.
it's not that difficult for a vandal to create any number of user-accounts..
daniel
I think people are misunderstanding Kelson's proposal here (unless I am): perhaps the terminology chosen triggered the wrong associations. As I understand it, what is being suggested is a per-article, per-user label, recording when that page was last looked at - like a kind of "bookmark this version" feature.
Some online forums do something similar, keeping track in some way of where you have read up to ("There are 45 new posts since you last replied, and 15 since you last viewed this thread"). So this isn't for ''assessing'' articles, it's just to say "this is the version I've read".
I actually think this is a really useful idea: currently, the watchlist gives you two links: diff of most recent change, or history page to manually select a larger slice of the history. If more than one change has been made since you last looked, you'll miss all but the last one if you click "diff"; I often end up loading the history and looking for an edit I know I've already seen, or one I made, and diffing back to that. With this feature, you could have a third option "diff to last bookmarked/validated/viewed/whatever", which shows a cumulative diff comparing the current version to how it was when you 'bookmarked'/'validated'/viewed it.
As you can see, I'm not sure what the least confusing name for this button would be, but I would certainly like to see it. Alternatively, of course, we could reset the version every time the user loads the page, but I think loading doesn't necessarily imply reading, so this might be a bit annoying. We could, otoh, reset it when the user *edits* a page: you're unlikely to want to check on edits further back than your own last change.
On Fri, 15 Oct 2004 23:40:52 +0200, Emmanuel Engelhart emmanuel@engelhart.org wrote:
On Thu, 14 Oct 2004 17:00:54 +0200 Emmanuel Engelhart emmanuel@engelhart.org wrote:
Hi
I propose to add a new feature, to improve our fight again vandalism. I mean a diff, for watchlist articles, between the current version and the last read one.
I need opinion about the feature ? and about an elegant technical solution too ?
Nobody again... ?
Follows my technical proposition:
- add a field "wl_validate" (containing an article version id)
- add a tab to validate an article (shown only if the article belongs to the user watchlist)
- add a link for each modified article in the watchlist to compare the current version with the last validated one.
Kelson
On Sun, 17 Oct 2004 18:04:46 +0100 Rowan Collins rowan.collins@gmail.com wrote:
I think people are misunderstanding Kelson's proposal here (unless I am): perhaps the terminology chosen triggered the wrong associations. As I understand it, what is being suggested is a per-article, per-user label, recording when that page was last looked at - like a kind of "bookmark this version" feature.
Some online forums do something similar, keeping track in some way of where you have read up to ("There are 45 new posts since you last replied, and 15 since you last viewed this thread"). So this isn't for ''assessing'' articles, it's just to say "this is the version I've read".
I actually think this is a really useful idea: currently, the watchlist gives you two links: diff of most recent change, or history page to manually select a larger slice of the history. If more than one change has been made since you last looked, you'll miss all but the last one if you click "diff"; I often end up loading the history and looking for an edit I know I've already seen, or one I made, and diffing back to that. With this feature, you could have a third option "diff to last bookmarked/validated/viewed/whatever", which shows a cumulative diff comparing the current version to how it was when you 'bookmarked'/'validated'/viewed it.
Exactly.
Enotify patch from Thomas Gries should provide this feature or something like that. So I'm now waiting on the integration of enotify into the CVS tree.
As you can see, I'm not sure what the least confusing name for this button would be, but I would certainly like to see it. Alternatively, of course, we could reset the version every time the user loads the page, but I think loading doesn't necessarily imply reading, so this might be a bit annoying. We could, otoh, reset it when the user *edits* a page: you're unlikely to want to check on edits further back than your own last change.
Yes, or a new "reviewed" button/tab ...
My opinion is that the "validate" term, as used now, is not correct. "Vote" would be better for a system which is more a voting system. If users want to validate some articles, basing on the voting system results that's an other problem.
But I want to see Enotify integreded, to speak more about the workflow details.
Emmanuel
I know you have mentioned my validation feature, but perhaps you didn't realize that is, among other things, does *exactly* all that. It even has a "validate" tab. It doesn't do the "last viewed" thing, though.
Magnus
Emmanuel Engelhart wrote:
On Sun, 17 Oct 2004 18:04:46 +0100 Rowan Collins rowan.collins@gmail.com wrote:
I think people are misunderstanding Kelson's proposal here (unless I am): perhaps the terminology chosen triggered the wrong associations. As I understand it, what is being suggested is a per-article, per-user label, recording when that page was last looked at - like a kind of "bookmark this version" feature.
Some online forums do something similar, keeping track in some way of where you have read up to ("There are 45 new posts since you last replied, and 15 since you last viewed this thread"). So this isn't for ''assessing'' articles, it's just to say "this is the version I've read".
I actually think this is a really useful idea: currently, the watchlist gives you two links: diff of most recent change, or history page to manually select a larger slice of the history. If more than one change has been made since you last looked, you'll miss all but the last one if you click "diff"; I often end up loading the history and looking for an edit I know I've already seen, or one I made, and diffing back to that. With this feature, you could have a third option "diff to last bookmarked/validated/viewed/whatever", which shows a cumulative diff comparing the current version to how it was when you 'bookmarked'/'validated'/viewed it.
Exactly.
Enotify patch from Thomas Gries should provide this feature or something like that. So I'm now waiting on the integration of enotify into the CVS tree.
As you can see, I'm not sure what the least confusing name for this button would be, but I would certainly like to see it. Alternatively, of course, we could reset the version every time the user loads the page, but I think loading doesn't necessarily imply reading, so this might be a bit annoying. We could, otoh, reset it when the user *edits* a page: you're unlikely to want to check on edits further back than your own last change.
Yes, or a new "reviewed" button/tab ...
My opinion is that the "validate" term, as used now, is not correct. "Vote" would be better for a system which is more a voting system. If users want to validate some articles, basing on the voting system results that's an other problem.
But I want to see Enotify integreded, to speak more about the workflow details.
Emmanuel
On Sun, 17 Oct 2004 20:25:42 +0200, Magnus Manske magnus.manske@web.de wrote:
I know you have mentioned my validation feature, but perhaps you didn't realize that is, among other things, does *exactly* all that. It even has a "validate" tab. It doesn't do the "last viewed" thing, though.
Sorry, but I'm lost: which bits does and doesn't your system do? I had thought (without delving far into it) that yours was a 'global' validation feature: a way of the community communicating to each other their opinion of the quality of an article. Whereas the feature proposed here was an *individual* validation feature, to say "I have read this version" and then be told "this is what's changed since the version you read".
It seems something similar to the latter is bundled in with the 'eNotif' system, but I'm just confused at what "... does *exactly* all that ... doesn't do the "last viewed" thing ..." was referring to - it seems a bit of a contradiction.
Rowan Collins wrote:
On Sun, 17 Oct 2004 20:25:42 +0200, Magnus Manske magnus.manske@web.de wrote:
I know you have mentioned my validation feature, but perhaps you didn't realize that is, among other things, does *exactly* all that. It even has a "validate" tab. It doesn't do the "last viewed" thing, though.
Sorry, but I'm lost: which bits does and doesn't your system do? I had thought (without delving far into it) that yours was a 'global' validation feature: a way of the community communicating to each other their opinion of the quality of an article. Whereas the feature proposed here was an *individual* validation feature, to say "I have read this version" and then be told "this is what's changed since the version you read".
OK, brief summary: Every user can vote for every article version in a number of "categories". When you click the "vote" tab for any article version, you see your old votes as well, and the system can merge them with your new vote, if you want.
It does *not* automatically register which version of an article you viewed, though, only the ones you voted for.
Magnus
On Sun, 17 Oct 2004 20:25:42 +0200 Magnus Manske magnus.manske@web.de wrote:
I know you have mentioned my validation feature, but perhaps you didn't realize that is, among other things, does *exactly* all that. It even has a "validate" tab. It doesn't do the "last viewed" thing, though.
It's right, I apologize, a user validates an article version, not just an article.
Emmanuel
wikitech-l@lists.wikimedia.org