User talk:DOI bot

From Wikipedia, the free encyclopedia

Reporting a bug? Please go to the bugs page. For urgent matters, please leave notes directly on the operator's talk page so the bot can be paused as soon as possible.

Contents

[edit] Eats up en-dashes

In this change the bot changed – HTML entities to plain hyphens. I consider this a bug, since the Manual of Style calls for the use of en-dashes in number ranges. Jyotirmoyb (talk) 10:35, 2 June 2008 (UTC)

It actually converted them to en dashes. Please see the discussion below. Smith609 Talk 11:15, 2 June 2008 (UTC)

Related to the eating up of en dashes... did the bot actually do anything in this change? Kehrbykid (talk) 22:25, 13 June 2008 (UTC)

[edit] Which templates?

Does DOI bot handle {{citation}}, or only the {{cite}} series? —David Eppstein (talk) 15:15, 23 May 2008 (UTC)

It's restricted to {{Cite journal}} at the moment, as these are the most likely to have DOIs - allowing the bot to edit other templates would open up a whole new world in potential of little bugs. Maybe I'll broaden its scope to include these someday... Smith609 Talk 15:51, 23 May 2008 (UTC)
I think {{citation}}s that include a journal= field are just about equally likely to have DOIs. Other cites, and non-journal citations, are iffier, I agree. —David Eppstein (talk) 17:22, 23 May 2008 (UTC)

[edit] Abstract

[1] Is it really reasonable to label something an "abstract" when the linked page has the full document linked in PDF and/or GIF form? WilyD 12:50, 24 May 2008 (UTC)

Thanks for pointing this out. Now fixed.

Is it really fixed? DOI bot has gone through all my marked articles with (abstract). See, for example, mired and D65. Just because a subscription is required it does not mean that the full article is not provided. Are you going to fix this or do we have to go through the backlog manually? --Adoniscik(t, c) 12:27, 28 May 2008 (UTC)

I think the "abstract" summary is accurate. The page the URLs on these pages link to is the abstract, or a summary of the article. If the editor had intended to supply a link to the full text, they would have saved their reader a click and linked directly to the full text, which the bot would mark "subscription required" if necessary. Most academic readers will know how to procure a full text article from an abstract link, or through their own library via DOIs. An abstract link gives the causal reader an impression of what is on the other end of the link; it doesn't say "abstract only" and imply that no other information is available. Perhaps "free abstract" would be a better wording?
I have disabled this feature for the time being until the matter is resolved; feel free to argue back! Smith609 Talk 12:51, 28 May 2008 (UTC)

Not at all. I linked to the abstract page as a matter of policy for two reasons:

  1. There is no way to go back from the PDF to the abstract page. The abstract page, however, does link to the PDF. Frequently it also provides invaluable information not readily available from the PDF, such as a permalink. The permalink generally does not resolve to the PDF.
  2. I connect to subscriber sites through a proxy server affiliated with my educational institution. The redirection is not entirely transparent and hinders my ability to save the PDF to the hard disk (which is my preferred course of action) rather than attempting to open it in the browser. However, if I first attempt to access the abstract page, I go through the redirection process and then can readily right click and save the PDF on the redirected page. If I attempt to right click and save the PDF without first going through the redirected abstract page, I merely save the HTML page where I have to enter my user credentials for my educational institution.

I do appreciate that your bot fixes the en dash in the page numbers! --Adoniscik(t, c) 15:33, 28 May 2008 (UTC)

If you've linked to the abstract page, then surely the bot should mark the link as a link to an abstract page? Smith609 Talk 16:08, 28 May 2008 (UTC)
Only if the abstract is all that is available. That is what I understand when I see the "(abstract)" remark. Maybe a better wording is indeed advisable. I can live with (free abstract), but I think it is just clutter. I would probably say nothing at all. Adoniscik(t, c) 16:22, 28 May 2008 (UTC)
How about "abstract page"...? Smith609 Talk 16:43, 28 May 2008 (UTC)

[edit] Barnstar

The Minor Barnstar
For the great job of improving citations. utcursch | talk 16:11, 25 May 2008 (UTC)

[edit] Please restrict your bot to DOI's

I'm grateful for the handful of DOI's that your bot came up with. Nonetheless, I am going to revert your bot's changes to problem of Apollonius, an article I have been working on for months, for the following reason. I deliberately write & n d a s h ; instead of – to help me proofread the article. Forgive me for saying so, but it doesn't seem wise to make invisible formatting changes that also interfere with a human editor's ability to maintain the article. My suggestion would be to turn off that feature of your bot. Otherwise, well done! :) Willow (talk) 21:58, 27 May 2008 (UTC)

Done: The bot will no longer replace –. Thanks for your feedback, and sorry for the inconvenience! Smith609 Talk 07:19, 28 May 2008 (UTC)

Thanks, Smith! :) I plan on using your bot a lot in the future; thank you a lot for that as well! Willow (talk) 11:20, 28 May 2008 (UTC)

I see that em-dashes are fair game, still. As they should be. Again, thanks for DOI bot; although I've been running around cleaning up after its changes on my watchlist (mostly either replacing or removing deadlinks), I find what it does very valuable. —David Eppstein (talk) 17:15, 28 May 2008 (UTC)

[edit] Broken format with two templates in the id field

this change broke the reference formatting. A line id = {{arxiv | archive = math.DG | id = 0309408}} {{MathSciNet | id = 2178969}} was incorrectly changed to id = 2178969}}, losing information and spuriously adding an extra pair of braces that messed up the url on the following line. —David Eppstein (talk) 19:58, 28 May 2008 (UTC)

Thanks. I've taken the bot offline until I fix it. Smith609 Talk 21:02, 28 May 2008 (UTC)

Probably the same basic problem: id=doi diff broke the template. JackSchmidt (talk) 17:12, 3 June 2008 (UTC)

[edit] approved for this?

I was looking over my watchlist and I noticed this change. I then checked out the two bot approval links on the bot's page. I may have missed it, but it doesn't look like your bot was approved to make changes like this.--Rockfang (talk) 16:09, 31 May 2008 (UTC)

The bot is approved to "correct common mistakes". Many edits such as this one were made during the bot's trial period and did not elicit comment. I hope it's not causing you any inconvenience? If your watchlist is becoming cluttered by bot edits, the "hide bot edits" link may come in handy. Smith609 Talk 06:12, 1 June 2008 (UTC)
Cool. I just wanted to make sure it was approved. I reread the bot approval requests and found it. Thanks for the info.--Rockfang (talk) 06:36, 1 June 2008 (UTC)

[edit] Parsing page numbers

I noticed in this edit that the bot picked out part of the pages parameter (|pages=223), but not the whole (|pages=223–235). Perhaps it is because Blackwell Synergy (gasp) use en dashes for their page range, and not the more common and less correct hyphen-minus? Who knows. Thought I'd let you know. +mt 18:35, 7 June 2008 (UTC)

Unfortunately the database that the bot consults very rarely contains "end page" data. I feel that scraping it from the website itself is an unjustifiable use of time and resources - so unfortunately you'll have to be content with the start only. Thanks for pointing it out, and sorry I can't do anything about it! Smith609 Talk 21:13, 7 June 2008 (UTC)
Quite alright, I'll just keep my eye out for it. Thanks for the info. +mt 17:26, 8 June 2008 (UTC)

[edit] Page numbers

This bot is replacing the dash for pages numbers with the endash. While this appears to be correct style, this is not an approved function for the bot. --EncycloPetey (talk) 19:21, 9 June 2008 (UTC)

Did you see the discussion on similar changes two sections up? —David Eppstein (talk) 20:06, 9 June 2008 (UTC)
Yes, I saw that claim, but that function is not listed on the user page for this bot, nor could I find the discussion approiving it for that function. When "correcting common mistakes" was explained in the bot approval, stylistic mistakes were not discussed. What was discussed were correction such as replacing id= with pmid= when appropriate, or correcting Journal= to journal=. These corrections are invisible, and this is not the same as making stylistic changes in editing. --EncycloPetey (talk) 20:35, 9 June 2008 (UTC)
This is becoming petty.
I've created a very specificially worded request for bot approval for this task. You are invited to contribute to the discussion. Smith609 Talk 07:31, 10 June 2008 (UTC)

[edit] Time Delay

Hello,

I missed reverting some vandalism due to a bot edit to Aluminium diff. Can the bot be configured not to make an edit within x minutes of an IP user, or user with less than x edits? Whilst it wasn't long before someone read and reverted the vandals edits diff, for a less major article than Aluminium and less blatant vandalism, the vandalism could have gone undetected for a while. Something along these lines would be good, as I ignored the bot edit on my watchlist. User A1 (talk) 11:47, 12 June 2008 (UTC)

[edit] issue = 17-5

A recent instance of DOI bot added a volume=17, but left issue=17-5. I think fixing that would fall under "common mistakes".

Suggested rule: If volume=xxx is added, and issue=xxx-yyy, xxx.yyy, or xxx(yyy) already exists, set issue=yyy .

There's a chance of a mistake it's really volume 1 issue 1 (January), and what's written is issue=1(January), but that seems a rare occurence. — Arthur Rubin (talk) 23:32, 13 June 2008 (UTC)

I frequently use issue=nn–nn for journals in which multiple issue numbers appear in a single bound copy of the journal. See e.g. the Hagauer, Imrich, & Klavžar reference in median graph. In that case, there should be a volume as well, but it's not obvious to me that no volume and issue with dash should be "fixed" in this way. I'd prefer in this case that if the bot has a database that indicates the correct volume and issue that it use the data from the database, and that otherwise it leaves it alone. —David Eppstein (talk) 23:54, 13 June 2008 (UTC)
I think what happened in that particular edit is that there isn't an issue number in the database, so the bot left the incorrect issue=17-5 alone. Someone would have to look at the DOI entry to be sure. However, and this would delete your sensible generally sensible convention, fill the issue number from the database, or at least flag the edit for further review if an apparently incorrect field is filled in. — Arthur Rubin (talk) 00:20, 14 June 2008 (UTC)