All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
The idea crossed my mind after the past several days noticing quite a bit of information on Mediawiki.org is either poorly worded, outdated or just plain wrong. Now, MW.org doesn't suffer from most of the issues that are seen on other projects: we're small, we don't really have anything to edit war over, and we don't seem to get (as much :) spam and vandalism. I was curious as to whether we could use FlaggedRevs as a quality control over our documentation.
A lot of the docs have been written by people other than developers, and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Hopefully we can improve the overall quality of the docs on MW.org. I'm certainly open to other ideas too.
-Chad
On Fri, Jul 3, 2009 at 11:07 AM, Chadinnocentkiller@gmail.com wrote:
A lot of the docs have been written by people other than developers,
Ever met a developer who likes writing doc? :)
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
Steve
Steve Bennett wrote:
On Fri, Jul 3, 2009 at 11:07 AM, Chadinnocentkiller@gmail.com wrote:
A lot of the docs have been written by people other than developers,
Ever met a developer who likes writing doc? :)
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
That's true, but even if there is such a process, it would be helpful to communicate to users whether or not the revision that they're looking at has been reviewed.
I would support adding FlaggedRevs for the Manual: namespace.
On Thu, Jul 2, 2009 at 9:07 PM, Chad innocentkiller@gmail.com wrote:
All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
The idea crossed my mind after the past several days noticing quite a bit of information on Mediawiki.org is either poorly worded, outdated or just plain wrong. Now, MW.org doesn't suffer from most of the issues that are seen on other projects: we're small, we don't really have anything to edit war over, and we don't seem to get (as much :) spam and vandalism. I was curious as to whether we could use FlaggedRevs as a quality control over our documentation.
A lot of the docs have been written by people other than developers, and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Hopefully we can improve the overall quality of the docs on MW.org. I'm certainly open to other ideas too.
When I saw the topic title, I thought of FlaggedRevs at the more basic
level of reviewing edits. I try to go through the recent changes and review every edit that looks remotely suspicious or interesting, but I'm sure I miss plenty. When I finally catch up on my backlog I often see destructive edits from several days previous that have not been reverted. This means that other people aren't catching everything, and since I know I won't catch everything, this means that there's stuff going uncaught—and when that stuff includes changing a "true" to a "false," I get worried. The AbuseFilter could probably help in preventing such things, but I'd still like to have every edit reviewed. Given our high eyes–edits ratio, I think this is possible. With that aside, I like what you're proposing.
On Fri, Jul 3, 2009 at 12:03 AM, Steve Bennett stevagewp@gmail.com wrote:
On Fri, Jul 3, 2009 at 11:07 AM, Chadinnocentkiller@gmail.com wrote:
A lot of the docs have been written by people other than developers,
Ever met a developer who likes writing doc? :)
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
Steve
There's a class of people in between the people who write the code and the people who use it. We like documenting things and helping people with their problems, and even though we don't always know how something works, we (hopefully) have the savvy to either check the source or ask somebody who does know. (We have lots of wonderful and helpful devs, of course; I'm just saying that non-developers can still help, especially since we often have more spare focus and time.) Having developers review everything would be great, but if they don't have the time for all of it, we can leverage our base of knowledgeable MediaWiki hangers-on for a systematic review of the site docs.
On Fri, Jul 3, 2009 at 12:03 AM, Steve Bennettstevagewp@gmail.com wrote:
On Fri, Jul 3, 2009 at 11:07 AM, Chadinnocentkiller@gmail.com wrote:
A lot of the docs have been written by people other than developers,
Ever met a developer who likes writing doc? :)
Hehe, I've worked in IT for several years now. Nobody likes to :)
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
Steve
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
I try to ensure my documentation is thorough as I write it. That being said, there's still plenty out there that needs review. I'm just suggesting FlaggedRevs as a mechanism to aid in that.
To further expand on my original statements, I'm not suggesting the developers are the only ones with in-depth knowledge of how Mediawiki works. There are certainly other members of the community we can trust to handle this task as well. I'm looking at this primarily as tools to aid in fixing a problem. Of course without work on part of document writers/reviewers, this won't go anywhere.
-Chad
-Chad
Chad wrote:
On Fri, Jul 3, 2009 at 12:03 AM, Steve Bennettstevagewp@gmail.com wrote:
On Fri, Jul 3, 2009 at 11:07 AM, Chadinnocentkiller@gmail.com wrote:
A lot of the docs have been written by people other than developers,
Ever met a developer who likes writing doc? :)
Hehe, I've worked in IT for several years now. Nobody likes to :)
Guess what , I'm in IT for several years now also , and I like writing docs , no matter if it's a documenting my own code or documenting the functionality that I'm implementing , keeping an agenda of what I did and what I still have to do. It's called getting organized. If I don't do this it's going to get out of control sooner or later. I like to write docs because I am sure that after 6 months from now if I look at my stuff it would make no difference if it was foreign code or my own code. So maybe your question should be "Ever met a very good developer who likes writing doc?" The answer I suspect is in the affirmative.
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
Steve
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
I try to ensure my documentation is thorough as I write it. That being said, there's still plenty out there that needs review. I'm just suggesting FlaggedRevs as a mechanism to aid in that.
To further expand on my original statements, I'm not suggesting the developers are the only ones with in-depth knowledge of how Mediawiki works. There are certainly other members of the community we can trust to handle this task as well. I'm looking at this primarily as tools to aid in fixing a problem. Of course without work on part of document writers/reviewers, this won't go anywhere.
-Chad
-Chad
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 04/07/2009, at 5:21 AM, randomcoder1 wrote:
Guess what , I'm in IT for several years now also , and I like writing docs , no matter if it's a documenting my own code or documenting the functionality that I'm implementing , keeping an agenda of what I did and what I still have to do. It's called getting organized. If I don't do this it's going to get out of control sooner or later. I like to write docs because I am sure that after 6 months from now if I look at my stuff it would make no difference if it was foreign code or my own code. So maybe your question should be "Ever met a very good developer who likes writing doc?" The answer I suspect is in the affirmative.
Good for you, but that's not really what we're talking about.
and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Perhaps we could start by getting developers to thoroughly review documentation?
With no way of planning that review, nor of co-ordinating with other developers to make sure everything's been at least somewhat looked at, there's no point in doing so.
You're proposing a technical solution to a people problem. The problem is not that the site can't display the fact that a developer vouches for the quality of documentation. The problem is that there are no processes for getting developers to review documentation and vouch for it.
To further expand on my original statements, I'm not suggesting the developers are the only ones with in-depth knowledge of how Mediawiki works. There are certainly other members of the community we can trust to handle this task as well. I'm looking at this primarily as tools to aid in fixing a problem. Of course without work on part of document writers/reviewers, this won't go anywhere.
Giving some indication of what's been checked and is authoritative and what isn't is generally a good idea. I don't document much on MediaWiki.org, but I sure don't mind checking over existing documentation.
-- Andrew Garrett Contract Developer, Wikimedia Foundation agarrett@wikimedia.org http://werdn.us
So you want to use flagged revs to correct some other content ? You can look on this blog , the same idea is described on writing a probabilistic grammar checker based on flagged revisions (the ones flagged for some incorrect grammatical content) http://www.orchycore.com/index.php?/archives/96-Proposal-for-a-probabilistic...
Chad wrote:
All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
The idea crossed my mind after the past several days noticing quite a bit of information on Mediawiki.org is either poorly worded, outdated or just plain wrong. Now, MW.org doesn't suffer from most of the issues that are seen on other projects: we're small, we don't really have anything to edit war over, and we don't seem to get (as much :) spam and vandalism. I was curious as to whether we could use FlaggedRevs as a quality control over our documentation.
A lot of the docs have been written by people other than developers, and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Hopefully we can improve the overall quality of the docs on MW.org. I'm certainly open to other ideas too.
-Chad
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Sun, Jul 5, 2009 at 1:44 AM, randomcoder1randomcoder1@gmail.com wrote:
So you want to use flagged revs to correct some other content ? You can look on this blog , the same idea is described on writing a probabilistic grammar checker based on flagged revisions (the ones flagged for some incorrect grammatical content) http://www.orchycore.com/index.php?/archives/96-Proposal-for-a-probabilistic...
Chad wrote:
All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
The idea crossed my mind after the past several days noticing quite a bit of information on Mediawiki.org is either poorly worded, outdated or just plain wrong. Now, MW.org doesn't suffer from most of the issues that are seen on other projects: we're small, we don't really have anything to edit war over, and we don't seem to get (as much :) spam and vandalism. I was curious as to whether we could use FlaggedRevs as a quality control over our documentation.
A lot of the docs have been written by people other than developers, and a lot of the docs have never been read by a developer. That being said, using FlaggedRevs we might be able to deliver more solid docs on MW.org by flagging docs at like two levels. One could be like a basic "has been looked over for glaring errors and basic readability" and a second could be "has been thoroughly reviewed and is considered the doc on the given subject."
Hopefully we can improve the overall quality of the docs on MW.org. I'm certainly open to other ideas too.
-Chad
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
I'm not really sure how this relates to what we're discussing. The idea in that post is to use Wikipedia's edit history as a basis to build grammar checking rules around.
I'm not suggesting that this system will fix issues in our documentation, but rather that it will assure the reader as to the level of review has gone into the docs they are currently reading.
-Chad
On Thu, 02 Jul 2009 21:07:01 -0400, Chad wrote:
All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
One issue on Mediawiki.org is that the help namespace is in the public domain, intended to be mirrored for local use. But if someone copies it at the wrong time, they could end up pulling in all sorts of nonsense onto their local wiki; someone fixing it after the fact won't help them as much.
So there really should be some way to copy the reviewed version of that space, even if that's not the version that's shown by default, just to be assured that you'll get good content, since in the help namespace that's much more important that having things perfectly up to date.
On Mon, Jul 6, 2009 at 12:11 PM, Steve Sanbegssanbeg@ask.com wrote:
On Thu, 02 Jul 2009 21:07:01 -0400, Chad wrote:
All,
I had this idea cross my mind earlier today, but I got so tied up in meetings that I couldn't sit down and write out a proper e-mail until this evening. I was curious as to whether we think FlaggedRevs might be of use to Mediawiki.org, and if so, how exactly would we use it?
One issue on Mediawiki.org is that the help namespace is in the public domain, intended to be mirrored for local use. But if someone copies it at the wrong time, they could end up pulling in all sorts of nonsense onto their local wiki; someone fixing it after the fact won't help them as much.
So there really should be some way to copy the reviewed version of that space, even if that's not the version that's shown by default, just to be assured that you'll get good content, since in the help namespace that's much more important that having things perfectly up to date.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Exactly, which is my main point I'm trying to make here. We've got a situation right now where we've got documentation that we say is _the_ source for info about MediaWiki, and yet we cannot vouch for it.
I don't think our major issue is vandalism or people purposefully inserting false information, but more so users who don't really know exactly what they're changing, they just know what worked for them. I've got several docs I've looked at in the past several days that suffer from this: clueless people trying to be helpful.
Not that we should discourage helpful people, but helpful people don't always know what they're talking about; at least not well enough to allow them to set the documented standard :)
-Chad
Exactly, which is my main point I'm trying to make here. We've got a situation right now where we've got documentation that we say is _the_ source for info about MediaWiki, and yet we cannot vouch for it.
I don't think our major issue is vandalism or people purposefully inserting false information, but more so users who don't really know exactly what they're changing, they just know what worked for them. I've got several docs I've looked at in the past several days that suffer from this: clueless people trying to be helpful.
Not that we should discourage helpful people, but helpful people don't always know what they're talking about; at least not well enough to allow them to set the documented standard :)
Yes, but the help namespace is a much simpler, less controversial place to start than other parts of the wiki.
This issue was brought up there years ago, long before there was any technical means to address it.
The idea is these pages could be mirrored to a local site, but the problem is that mirrored set could easily contain vandalized, wrong, or nonsense pages that would need to be sorted out, which pretty much rules out any kind of automated mirroring, which was part of the point of setting it up that way. Flagged revisions could simply fix that.
In most pages, unless the flagged revision was shown by default, it wouldn't be too useful. Here, as long as you can access the flagged revision easily, i.e. via the API, it would solve a real problem without discouraging editors by preventing their work from being shown on the site.
And, really, help pages don't need to change that often, so a brief lag wouldn't be a problem. If fact, it could even be useful, i.e. to document new features as they are implemented, but flag which version of the page applies to the latest quarterly release.
Although it's good if the developer can document every detail, it may not be practical for them to spend a lot of time polishing the document, if they can at least get their point across quickly. From what I've seen, misspellings and bad grammar tend to cause a bunch of IP edits, but these tend to sort themselves out. The problem is when the documentation is unclear, people may clarify things that they don't understand.
On Mon, Jul 6, 2009 at 3:34 PM, Steve Sanbegssanbeg@ask.com wrote:
Exactly, which is my main point I'm trying to make here. We've got a situation right now where we've got documentation that we say is _the_ source for info about MediaWiki, and yet we cannot vouch for it.
I don't think our major issue is vandalism or people purposefully inserting false information, but more so users who don't really know exactly what they're changing, they just know what worked for them. I've got several docs I've looked at in the past several days that suffer from this: clueless people trying to be helpful.
Not that we should discourage helpful people, but helpful people don't always know what they're talking about; at least not well enough to allow them to set the documented standard :)
Yes, but the help namespace is a much simpler, less controversial place to start than other parts of the wiki.
This issue was brought up there years ago, long before there was any technical means to address it.
The idea is these pages could be mirrored to a local site, but the problem is that mirrored set could easily contain vandalized, wrong, or nonsense pages that would need to be sorted out, which pretty much rules out any kind of automated mirroring, which was part of the point of setting it up that way. Flagged revisions could simply fix that.
In most pages, unless the flagged revision was shown by default, it wouldn't be too useful. Here, as long as you can access the flagged revision easily, i.e. via the API, it would solve a real problem without discouraging editors by preventing their work from being shown on the site.
And, really, help pages don't need to change that often, so a brief lag wouldn't be a problem. If fact, it could even be useful, i.e. to document new features as they are implemented, but flag which version of the page applies to the latest quarterly release.
Although it's good if the developer can document every detail, it may not be practical for them to spend a lot of time polishing the document, if they can at least get their point across quickly. From what I've seen, misspellings and bad grammar tend to cause a bunch of IP edits, but these tend to sort themselves out. The problem is when the documentation is unclear, people may clarify things that they don't understand.
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
As a followup to all of this discussion: is there really any reason we can't go ahead and give this a try?
I think Help and/or Manual are really the only namespaces that need it, as they are the documentation namespaces. Most other stuff tends to be just general notes, etc which don't really need review.
-Chad
On 7/13/09 11:34 AM, Chad wrote:
As a followup to all of this discussion: is there really any reason we can't go ahead and give this a try?
I think Help and/or Manual are really the only namespaces that need it, as they are the documentation namespaces. Most other stuff tends to be just general notes, etc which don't really need review.
For those not in on the fun in IRC right now, we're trying this out right now. :)
The group permission config may be under flux for a bit...
-- brion
wikitech-l@lists.wikimedia.org