Wiki143:Bots/Requests for approval/Kingbotk

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

User:Kingbotk

--kingboyk 07:51, 1 September 2006 (UTC) I am using this account with AWB in a semi-automated fashion to tag talk pages with WikiProject templates. This is a case where automation is of very clear benefit - we on the Wikipedia 1.0 team need these talk pages tagged for our article assessment drive, and the work is ridiculously simple for a bot but slow and boring for a human. I always test my regular expressions and settings on some test pages with a totally manual run first and the work is simple. Duration: As long as it takes. --kingboyk 12:13, 6 August 2006 (UTC)

Ok, go ahead and trial it, post some diffs and we'll take a look -- Tawker 00:58, 9 August 2006 (UTC)

What kind of diffs are you after? Or would you care to choose a few edits at random from the bot's history? That might be better: I have nothing to hide and it's all simple stuff :) --kingboyk 15:00, 11 August 2006 (UTC)

It's been a week, and I'm sure people are sick of seeing my bot's edits in recent changes. I've tagged masses of talk pages, received 3 barnstars for the work, had a number of enquiries which have all been dealt with, and no complaints. Let's get this bot bit then please! :) --kingboyk 13:25, 18 August 2006 (UTC)

The bot is running way to fast for a non flagged bot. I've blocked it for 15 min trying to reset it / to see if we can get a bot flag assigned. -- Tawker 21:20, 20 August 2006 (UTC)

Yes, I plead guilty to that. I have such a mass of work to do I've been running a bit fast, sorry. I'll hold off until I hear from you. Cheers. --kingboyk 21:33, 20 August 2006 (UTC)

Bot flag is set -- Tawker 04:22, 22 August 2006 (UTC)

Can we clarify how the "automation" is in any way "semi"? Is there some human decision-making process involved on a per-article basis? I'm highly dubious about the benefit of this: there was much protestation about how Stub-Class articles were in some largely-unspecified way crucially different from stubs, necessitating the two parellel categories be created... and now the one is being mass-populated from the other? If an "assessment" is going to be done on this basis, surely it defeats the purpose of the exercise of making the supposed distinction. This doesn't strike me as "laborious work being done by humans that needs to be automated", as "164,911 edits thrashing the server that would probably never have taken place otherwise, to no clear benefit". There's well over half a million stubs, are they all going to have their talk-pages "bluelinked" with a notice announcing they're now (co-incidentally enough) "automatically assessed stub-class articles"? That's just far beyond the pace of any likely meaningful assessment exercise, it appears to me, so this is simply shuffling "in no way assessed" to "supposedly assessed" with no meaningful addition of information. Incidentally, this (semi-?)bot appears to be often making more than six edits per minute, which is too many for a flagged bot. Alai 05:35, 28 August 2006 (UTC)

My bot is AWB which is why it's described as "semi". I can change that, it's only semantics so let's not distract ourselves from the topic at hand. Like any other AWB botmaster, I run any new regular expression on test cases in my sandbox, then live on a few articles, before switching it to auto.

The primary goal of my bot is to get the living persons warning onto biographical talk pages; a second goal is to tag articles within the scope of WikiProjects with the WikiProject's banner. Approximately 100,000 of my edits have been placing the Blp warning, with around 15-20,000 still to do.

As for the stub-class solution; no, there's no intention to place a banner on every stub. What would be the point in that? However, if the article is getting a tag anyway, and it has a stub template, I'm tagging it stub-class with a very prominent message about what's been done. My thinking is that it's easier to get editors to check if the article really is Stub-Class and just remove the auto=yes parameter if it is, than to get them to assess from scratch. These articles are placed in a special category where they can be picked up by editors for fast-track assessment.

Stub-Class and stubs are not the same, you are correct. The assessment process won't work without the Stub-Class categories. They contain talk pages and are only one-level deep, containing the articles within the scope of one WikiProject or workgroup. If anyone is going to seriuously challenge that distinction again (and, really, in the greater scheme of things what's the price of a few extra categories?) then I may as well stop my work because we need those categories.

That said, most stubbed articles are Stub-Class or Start-Class. A fraction of 1% will be any better quality than that. If they don't get reassessed by a human within a few weeks then it's also fair to assume they're low priority too. Stub or Start Class + Low Priority = unlikely to make it to Wikipedia 1.0. Flagging these articles this way is a clear benefit from my bot - instead of being totally unassessed, we know it's probbaly Stub or Start and nobody has reassessed it yet.

As I see it, there are two seperate issues here: tagging with WikiProject templates, and setting a stub parameter. I only do the latter if the former is taking place. The latter has been done only on a subset of articles as an experiment, but I've received feedback from others that it's a good and helpful idea. If there's any consensus that it isn't helpful - and I doubt that because you're the first person to even mention it a negative light - I can quite easily and relatively happily turn that off. However, you'll still see the article getting tagged just not with that parameter so I'm not sure what that would achieve.

Unless my bot tag is revoked, I don't have any plans to stop the tagging of any article within the scope of WPBiography or other WikiProjects I work with. The work my bot has done has been invaluable. In addition to propogating the Blp warning it's helped the assessment system come alive. In just a few weeks of operations, WPBiography has assessed nearly 10,000 articles, of which only 3500 are stub-class, and - no doubt due to the high-profile templating - has gone from a dormant WikiProject to one of Wikipedia's most vibrant. I've had barnstars, messages of thanks and praise, and requests from other WikiProjects to do work for them (which I've had to decline). I'm totally confident that my work is not only beneficial and useful but has community consensus behind it.

I'm tagging at an appropriate speed as far as I'm concerned. You're probably getting confused with untagged bots. The recommended top speed for tagged bots is 10 seconds per edit. I need to go as fast as I reasonably can because of the numbers involved. --kingboyk 07:51, 1 September 2006 (UTC)

I have to add my voice to Kingboy's in that if the assessing work is already cut down for us, it would be wonderful because I manually assess articles at a rate of 1 per 2minutes and it is a pain in the neck so if the process is somewhat cut down already for assessors like me, it is a great thing.

Secondly, deadlines for such projects as Wikipedia 1.0 and 0.5 (alpha release version) are so close and if the bot can lend a hand in this process I clearly see this in a good light. Lincher 11:46, 1 September 2006 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.