Wikipedia:Bots/Requests for approval/LemmeyBOT 2c

From Wikipedia, the free encyclopedia

[edit] User:LemmeyBOT

tasks • contribs • count • sul • logs • page moves • block user • block log • flag log • flag bot

Operator: Lemmey talk

Automatic or Manually Assisted: Automatic

Programming Language(s): Python

Function Summary: Destroy Bastards (Referneces that is...)

Edit period(s) (e.g. Continuous, daily, one time run):Daily / on command

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

Function Details:

SHORT VERSION: This task specifically replaces Bastard References with {{subst:Fact-now}} tags. A Bastard is defined as any named reference like <ref name = Lemmey/> in an article that has never, in its entire history, had a parent tag like <ref name = Lemmey>www.example.com</ref>

LONG VERSION: This task explanation uses the analogy that named references (i.e. <ref name = Larry>) in an article act as a family. The explanation uses crude but clear language to effectively describe the task.

  • Parent Reference <ref name = Larry> reference source or cite tag </ref>
  • Child Reference <ref name = Larry/>
  • Siblings all mentions of <ref name = Larry/> are siblings to each other
  • Orphans any child reference that does not have a parent reference
  • Bastards any child reference that does not have a parent reference in any previous verson of the article
  • Inheritance the copying of content of a dead parent to an orphan reference, the Orphan becomes a Parent to its former siblings

Guidelines of BOT behavior

  • Orphans are bad. They produce produce red line errors in the reflist and can not be seen by readers. They can not be linked to from the reflist. They appear to give a statement a source but it can not be verified by readers. Orphans must be fixed.
  • Bastards are really bad. Though they may not be intenetional they are giving false sense of authority to a statement within a paragraph. They should be replaced with Fact tags that can be filed in (or the entire statement removed) by editors. Bastards must be destroyed.

LemmeyBOT already does a fine job of restoring Orphaned References but many Bastards still exist. A reader will not realize a statement is infact unsourced unless they look at the reflist(which in itself is difficult because the inline links don't work on orphaned or bastard references)


[edit] Test

I've run the BOT on the article Fsix Corporation. A short history article I've nominated for deletion as not notable, it provides a clean real world test for this task. This article had two child references that were orphans. LemmeyBOT searched through the article history and determined the references were actually Bastards. LemmeyBOT replaced the refs with fact tags.

The diffs can be seen here [1][2]


[edit] Discussion

A couple questions:

1. What will the bot do if it finds an orphaned reference?
2. Where will it get the list of pages to run on?

--Mr.Z-man 20:25, 22 May 2008 (UTC)

LemmeyBOT already replaces orphans with their parents information taken from an older version of the article. This is already an approved task. LemmeyBOT processes articles found in the Category:Pages with incorrect ref formatting. --Lemmey talk 23:21, 22 May 2008 (UTC)

I don't think this can be automated. In your example article, the author did try to give references (diff 16:06, 24 September 2007). Although it's not in wikisyntax, you can clearly tell from that diff exactly what the author intended to go with REF1 and REF2. A human reading the wikitext could figure it out and fix it. Replacing the REF1 and REF2 with {{fact}} makes it much more difficult to fix; it's not an improvement. Gimmetrow 03:53, 23 May 2008 (UTC)

Perhaps you could figure out United States housing market correction then. --Lemmey talk 03:57, 23 May 2008 (UTC)
Yes, I could. And that's why I don't think you should just replace them with fact tags. Gimmetrow 12:06, 23 May 2008 (UTC)
Actually, looking at the revision before the bot's edit to United States housing market correction, it "fixed" the wrong thing. The ref with name "ffiec.gov-hmcrpr" was already defined a few lines above where the bot added it. Mr.Z-man 22:47, 23 May 2008 (UTC)
That was actually due to a fault in the earlier version of the bot where it didn't recognize the difference between "ffiec.gov-hmcrpr" and "ffiec.gov-hmcrpr. The fault has since been fixed. --Lemmey talk 22:53, 23 May 2008 (UTC)
I do tend to agree with Gimmetrow though, simply removing them makes it harder for humans to fix them as it removes the context of the reference name. I think it would be better to simply comment them out and/or leave a different inline tag than {{fact}}, something like [broken footnote]. Mr.Z-man 23:06, 23 May 2008 (UTC)
I agree, commenting them out (and adding a comment "can't find reference foo_1 - check for typos") is a much much better idea. There are always typos, homographs, issues with people not quoting names, and so on. I'd also rather see a different tag, appearing as [missing reference] or something. It's much less insulting. Pseudomonas(talk) 13:57, 25 May 2008 (UTC)

The references the bot would replace are listed here: User:Lemmey/L

From this version of United States housing market correction The refs for #18 and 19 are:[1][2] for #38 is:[3] for #49 and 50 are:[4][5] and for #52, 53 and 55 are:[6][7][8] Can your bot do that? Gimmetrow 23:08, 23 May 2008 (UTC)
  1. ^ "Alan Greenspan Interview with Jim Lehrer", The NewsHour with Jim Lehrer, 18 September 2007. 
  2. ^ "Greenspan alert on US house prices", Financial Times, 17 September 2007. 
  3. ^ Shiller, Robert (2005). Irrational Exuberance (2d ed.). Princeton University Press. ISBN 0-691-12335-7. 
  4. ^ This article classified several U.S. real-estate regions as "Dead Zones", "Danger Zones", and "Safe Havens."
    Fortune magazine Housing Bubble "Dead Zones"
    "Dead Zones" "Danger Zones" "Safe Havens"
    Boston Chicago Cleveland
    Las Vegas Los Angeles Columbus
    Miami New York Dallas
    Washington D.C. / Northern Virginia San Francisco / Oakland Houston
    Phoenix Seattle Kansas City
    Sacramento Omaha
    San Diego Pittsburgh
    "Welcome to the Dead Zone", Fortune, 4 May 2006. "Welcome to the dead zone: The great housing bubble has finally started to deflate, and the fall will be harder in some markets than others." 
  5. ^ Tully, Shawn. "Getting real about the real estate bubble: Fortune's Shawn Tully dispels four myths about the future of home prices", Fortune, 25 August 2005. 
  6. ^ "Adjustable-rate loans come home to roost: Some squeezed as interest rises, home values sag", The Boston Globe, 11 January 2006. 
  7. ^ "Mass. home foreclosures rise quickly", Boston Herald, 29 August 2006. 
  8. ^ Shiller, Robert. "The Bubble's New Home", Barron's, 20 June 2005. "The home-price bubble feels like the stock-market mania in the fall of 1999, just before the stock bubble burst in early 2000, with all the hype, herd investing and absolute confidence in the inevitability of continuing price appreciation. My blood ran slightly cold at a cocktail party the other night when a recent Yale Medical School graduate told me that she was buying a condo to live in Boston during her year-long internship, so that she could flip it for a profit next year. Tulipmania reigns."  Plot of inflation-adjusted home price appreciation in several U.S. cities, 1990–2005:
    Plot of inflation-adjusted home price appreciation in several U.S. cities, 1990–2005.
    Plot of inflation-adjusted home price appreciation in several U.S. cities, 1990–2005.

Apparently Yes

--Lemmey talk 23:25, 23 May 2008 (UTC)

You realize those refs are not in the page history, right? But they can be found. Gimmetrow 23:28, 23 May 2008 (UTC)

Bot is now blocked for doing an unrelated task. There are reasons why edits like this shouldn't be done. Some editors intentionally write articles without using named refs, and a bot shouldn't be imposing this. Gimmetrow 23:58, 23 May 2008 (UTC)

I can see it why may be helpful in some cases as having lots of refs sprinkled about paragraphs can be annoying, but using names like "jrvhew" and "trgnbs" which aren't at all related to the content of the reference isn't very helpful. Mr.Z-man 03:24, 24 May 2008 (UTC)

Bot unblocked. If you want to restore the rest of the missing refs from United States housing market correction, you might try pasting the article from United States housing bubble and removing it, then running your ref-restoring script. I suspect that's where almost all the named refs came from. Gimmetrow 21:44, 24 May 2008 (UTC)

Cluebot and msisiabot won't like it, but I can deal with being blacklisted for a week. --Lemmey talk 21:48, 24 May 2008 (UTC)
Can you do it in software? Load up the text of the other article and treat that as a history version of the current article. Doesn't need to be actually saved. (Though I wonder, with the overlap between the articles, if something shouldn't be cut out of one or the other.) Gimmetrow 21:56, 24 May 2008 (UTC)
Please note that Lemmey and his bot has been blocked per this AN thread for sockpuppetry. LegoKontribsTalkM 04:43, 1 June 2008 (UTC)