Wikipedia:Bots/Requests for approval/SmackBot XXII

The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was

Withdrawn by operator.

SmackBot XXII[edit]

Operator: Rich Farmbrough

Automatic or Manually assisted: Auto

Programming language(s): AWB

Source code available: AWB

Function overview: Delink full dates

Links to relevant discussions (where appropriate): See User:Full-date unlinking bot

Edit period(s): One time

Estimated number of pages affected: Depending on division of labour and other tasks taking priority up to 500,000, more likely 80-100,000

Exclusion compliant (Y/N): Y

Already has a bot flag (Y/N): Y

Function details: Delinks dates pretty much per the spec of User:Full-date unlinking bot

Discussion[edit]

I have been asked to turn SmackBot to this task, as FDUB has run into difficulties with maintaining the throughput planned.

The settings are available here.

They have been tested against FDUB's test data [1].

Regards, Rich Farmbrough, 22:35, 18 November 2009 (UTC).[reply]

When FDUB was approved it was agreed that any given article would only be processed once to prevent human editors finding themselves in an edit war with a bot. This bot should coordinate FDUB so that neither will revisit an article that either of them has edited. Jc3s5h (talk) 01:41, 19 November 2009 (UTC)[reply]

This is a valid point, I have inspected all of FDUB's edits (12,000+ at the time) and there 5 where date linking had been re-introduced to an existing date, by two editors, one was unaware and was simply editing the articles, the other by an editor who wished to retain the auto-formatting as there were mixed formats on the page. Nonetheless it is possible for SmackBot to do this relatively easily, and to create logs provided FDUB can deal with them. Rich Farmbrough, 08:59, 19 November 2009 (UTC).[reply]

Two questions:

Are these edits really so time-sensitive that the community-approved FDUB's editing rate is insufficient? If not, why have two bots doing it?
Has the necessary code for coordination mentioned above been created yet? Given the extreme drama behind this task, I wouldn't want to even start a trial without that being done.

Anomie ⚔ 01:36, 2 December 2009 (UTC)[reply]

When I put the request in FDUB was stalling almost all the time. At the rate it was going it would have taken approximately 5 years to complete the task.
The co-ordination above is relatively simple from my point of view, simply exclude all articles edited by FDUB. But as I have now scanned about 60k off FDUB's articles and only found 2 reverts, I think the fear of a human/bot edit war is misplaced.

Rich Farmbrough, 02:25, 3 December 2009 (UTC).[reply]

You imply that FDUB is no longer stalling all the time. What is the current time estimate? And, frankly, is there really a deadline?

Maybe?

Besides having SmackBot not edit articles FDUB has already edited, there is also the matter of having FDUB not edit articles that SmackBot has edited for this purpose. And given that a large community poll and ArbCom approval was necessary for FDUB to be approved in the first place, and that poll specifically stated that the bot would edit each page only once, I don't think changing that here would be appropriate.

Anomie ⚔ 18:15, 3 December 2009 (UTC)[reply]

FDUB has so far processed 146 thousand articles. With a total estimated scope of 650 thousand, and assuming a processing rate of 9-10 thousand per day, it will take 50-55 days as of today. So, assuming no breakdowns, the job, if left to FDUB alone, will be completed by the end of January at the earliest. As to the second question, the essays say there is no deadline. However, if we treat that as an ethos, everything will be embarrassingly complete - there would be bugs galore in the mediawiki software, there will be not a single GA, let alone FA. As there were initially concerns that there would be resistance, but it is now clear there is none. There might need to be no deadline if we were short of volunteers, but we have in Rich one of the most experienced bot operators in the whole of WP. By all those, the cleanup of seas of blue dates can and ought to be sped up. Ohconfucius ^¡digame! 01:58, 6 December 2009 (UTC)[reply]

I think this is a worthy request. To address the two questions:

The task is so large that even if the current rate and reliability of FDUB were maintained, the task would take an unnecessarily long time to complete.
It should be easy for Smackbot to avoid articles in FDUB contributions list. Given the chronological processing by FDUB, it should be easy to avoid those soon to be processed by FDUB. With 120,000 articles delinked by FDUB, there's been no evidence of link warring, so I believe we can put that behind us. Tony (talk) 14:25, 3 December 2009 (UTC)[reply]

Clarification: Is it intended that this supplement FDUB or replace FDUB's unlinking? --Dweller (talk) 16:55, 3 December 2009 (UTC)[reply]

To supplement, in the sense of doing the same task. So for a given article it would replace it. Rich Farmbrough, 23:04, 3 December 2009 (UTC).[reply]

Comment The idea that date delinking is controversial or drama-inducing is dated. Yes, once upon a time it was a hotly contested issue, but that was before the various RfCs, when the consensus was disputed. Now, nobody bats an eye when they see a user or even a bot delink dates, and this is supported by the bland talk page of the current bot, compared to the talk page of Lightbot's owner, who constantly received complaints about his bot's actions. Dabomb87 (talk) 01:29, 4 December 2009 (UTC)[reply]

Seems a good idea that an experienced bot operator contribute to a major task which is already agreed. Rjwilmsi 22:22, 11 December 2009 (UTC)[reply]

Just a note that you may need to ping some other BAG members who look active, I'd approve it, but I'm recused from all date linking activities. MBisanz ^talk 02:44, 27 December 2009 (UTC)[reply]

Does approval here even matter, since the bot op has decided to run it anyway without approval? Anomie ⚔ 15:07, 27 December 2009 (UTC)[reply]

Well, something needs to be done with this request. Approving it would be the easiest way. MBisanz ^talk 23:19, 27 December 2009 (UTC)[reply]

I think approval would be appropriate, imo. --IP69.226.103.13 (talk) 20:24, 28 December 2009 (UTC)[reply]

Withdrawn by operator. Would have liked approval but it is a dead letter. Rich Farmbrough, 21:41, 29 December 2009 (UTC).[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.