User talk:Dispenser/Archive 4

It appears that you know some javascript

So I was wondering if you had any idea how to implement my idea or could direct me towards some sort of tutorial to help me. I'm looking to make a script which adds a tab to the when in the user or user talk spaces which when clicked blanks the page and adds {{indef}} to the page. Just looking for a nice clean, quick way to do it. Thanks, --Terrillja _talk 23:41, 15 January 2010 (UTC)

if(wgNamespaceNumber==2 || wgNamespaceNumber==3){ // User || User_talk
	addOnloadHook(addPortletLink('p-cactions', 'javascript:ReplaceWithTemplate("'+wgPageName+'", "{{indef}}", "User is indefinitely blocked")', 'Replace w/ {{indef}}'))
}
function ReplaceWithTemplate(title, text, summary) {
	div = document.body.appendChild(document.createElement('div'))
	div.innerHTML = '<form id="submitme" method="post" action="'+wgScript+'?action=edit&wpIgnoreBlankSummary=yes&title='+
	title+'"><textarea name="wpTextbox1">'+
	text+'</textarea><input name="wpSummary" value="'+summary+'" /></form>'
	document.getElementById('submitme').submit()
}

This is just the basic implementation to get you started, it might result in edit conflict or deletion recreation warnings since some fields aren't given. Those can be gotten through the API, but that's a lot more than just 10 lines. With the API, it is possible for it to working silently in the background. Your best off asking for something like it to be implemented into wp:Twinkle. — Dispenser 01:48, 16 January 2010 (UTC)

I think I'll have to ask on the twinkle page, it removed all my top tabs for twinkle and friendly without replacing them with a new one. Thank you for the code, I'll post there and see if they can integrate it into twinkle without the two fighting each other. It's a lot farther than I got on my own, there is probably one hidden variable missing or something like that. Thank you, --Terrillja _talk 02:28, 16 January 2010 (UTC)

Wikipedia Day NYC

{{Wikipedia:Meetup/NYC/Invite/Wikipedia Day 2010}}

Coordinates log

Hi Dispenser, could you please add the region codes RS-KM and RS-VO (see ISO 3166-2:RS)? Thanks --тнояsтеn ⇔ 11:29, 16 January 2010 (UTC)

Added — Dispenser 05:26, 22 January 2010 (UTC)

geosearch.py

wanted: namespace filtering for Geosearch.py

I'm wondering if you might add namespace filtering to Geosearch.py . At the very least, I'd like to be able to limit searches to articles (mainspace) so I'm not distracted by user sandboxes, talk pages, etc. --Stepheng3 (talk) 00:36, 18 January 2010 (UTC)

I'm actually surprised that you are able to use it at all. Geosearch.py was one of my first SQL tools, before I knew how to optimize the stuff. That page should really be more automated, creating report and posting them on the toolserver. Anyway I've added the namespace stuff, but now the tool will only search in a single namespace at a time. — Dispenser 21:48, 27 January 2010 (UTC)

Geosearch.py is, admittedly, a bit tricky to use. However, it fills the need for near real-time searching, and I'd miss it terribly if it were to disappear. Single namespace searches should be adequate for my purposes. Thank you for your help, --Stepheng3 (talk) 23:51, 27 January 2010 (UTC)

Geosearch.py bug

I notice that Geosearch.py truncates the article title at the first ampersand. For instance, if you search on "37_47_30_N_122_23_45_W" it finds the Pacific Gas & Electric Building article, but gives the title as "Pacific Gas ". Is this something you can fix? --Stepheng3 (talk) 20:01, 27 January 2010 (UTC)

There is a possible fix from ghel.py:

# [[bugzilla:?????]] FIXME FILEBUG
# MediaWiki will store %26 as & breaking things
# We try to correct this as best as we can
# http://en.wikipedia.org/w/api.php?action=query&prop=extlinks&titles=7th_%26_Capitol_and_8th_%26_Capitol
# Update Dec 27, 2009: Seems to have been fixed, but the workaround doesn't hurt much
link = re.sub(r'&(?=[^&=]*(&|$))', '%26', link)

The bug seems to come and go, likely due to a lack of a regression test for it in MediaWiki. As the implementation in geosearch.py goes the link still works. — Dispenser 21:48, 27 January 2010 (UTC)

I agree that the link works. If fixing the title as well is a lot of work, then don't worry about it. Cheers, --Stepheng3 (talk) 23:54, 27 January 2010 (UTC)

lagging

I notice that Geosearch.py is suffering from a 5+ hour lag, as if its data had stopped being updated as of this morning. Usually its lag is no more than a few seconds. Any idea what's going on? --Stepheng3 (talk) 00:07, 28 January 2010 (UTC)

If a user runs a long query and don't set them up correct, the database doesn't freeze the state. Freezing is the default so the results aren't inconsistent. It back to normal catching up at -0.21 sec/sec. — Dispenser 03:21, 29 January 2010 (UTC)

PDFbot operation

Hi there, I looked through the source for PDFbot and the archives related to binary prefixes. I noticed that back in 2007 whenh this topic was last discussed the lack of consensus on WP:MOSNUM was given as a reason for not tidying those KiB/MiB links to something consistent. Now, in 2010 there is consensus on WP:MOSNUM that "The IEC standard prefixes kibi-, mebi-, gibi-, etc. (symbols Ki, Mi, Gi, etc.) are not familiar to most Wikipedia readers, so are generally not to be used...". The PDFLink manual page also shows the KB/MB notation and not the KiB/MiB notation. So I think now would be a good time to allow PDFbot to spot the KiB/MiB and convert to the more familiar KB and MB terminology. It would also have tidy the "what links here" pages for the "mebi/gibi" pages. It would also help keep the article pages consistent because right now there is a bit of a mess with some article pages using the unfamiliar KiB/MiB and most of the other pages using KB/MB. Fnagaton 20:54, 19 January 2010 (UTC)

PS. I forgot to add, if you don't have time to alter the bot's code I can make the changes if you want? Fnagaton 07:09, 22 January 2010 (UTC)

As for KiB, changing this would be a bad idea because bots have trouble understanding the meaning of prose, and because there are some instances where these are allowed, and bots don't know how to make exceptions.

Why does the bot change 6 hex digit color triplets to 3 hex digit color triplits? Where is the syntax for these color triplets documented? --Jc3s5h (talk) 18:11, 22 January 2010 (UTC)

To change the code so that the PDF links which use the XiB prefixes changed to use the WP:MOSNUM guidelines is easy. There isn't any of this problem with "meaning of prose" you write about because PDF links are not part of the main article text. The bot code already spots the type of prefix used for PDF links and preserves it, so it is trivial to change the code to make the bot follow WP:MOSNUM. Fnagaton 13:08, 23 January 2010 (UTC)

Thanks for the explanation. The explanation of the bot's function on its user page seems quite incomplete. If the changes are confined to reporting the size of linked PDF files, I see no problem. --Jc3s5h (talk) 17:01, 23 January 2010 (UTC)

No problem. As for the hex digit colouring, have you got a link to a diff that demonstrates what you mean?Fnagaton 05:06, 24 January 2010 (UTC)

OK, now I feel stupid. I've just noticed PDFbot removing mebibytes and replacing them with MB. For example this diff. I'm sure I remember it not doing this a week ago. Maybe I should have had more wake-up tea that day? :) Fnagaton 13:39, 24 January 2010 (UTC)

It took me a while to remember where I saw PDFbot changing color triplets, but an example may be found here. Apparently because of the large number of changes, my browser is very sluggish when viewing this diff. --Jc3s5h (talk) 19:46, 24 January 2010 (UTC)

It may have something to do with Web colors and "web-safe colors".Fnagaton 04:07, 27 January 2010 (UTC)

Dab solver

Moved to /Dab solver

Your help at User talk:Dschwen

Hello Dispenser. It works! Thanks a lot. --Vejvančický (talk) 09:07, 1 February 2010 (UTC)

Common fixes library

Decapitalizing {{Cite ...}}

Could you please have Reflinks not change {{Cite...}} to {{cite...}} (i.e. upper to lower case)? I don't care that your tool uses lower case in citations it creates itself, but it applies this change as a "fix" to any pre-existing citation that it encounters. As a result, it is effectively editwarring with every user who prefers to upper-case templates like that (or all templates), which is probably tens if not hundreds of thousands of users, and with other tools such as Cite4Wiki. No one will keel over and die as a result, but it's a waste of time, energy, server cycles, and most especially editor patience, since many people do not auto-trust an edit just because it is flagged as "bot" or "minor" and will go examine changes made to watchlisted items. :-) If other bots of yours are making similar case changes, please have them not do that. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 20:46, 1 February 2010 (UTC)

Its part of the commonfixes library which is used across many of my tools. The part take you issue with is consolidates all the redirects to cite web where I had to make the choice of capitalization. So here are the reasons:

1. Capitalizing completely unnecessary and is a key extra thus a potential for screwing up
2. The template documentation consistently uses all lowercase
3. Forcing all lowercase makes coding and reviewing more understandable, compact, and simpler. For example with refs where the casing can be anything, just to start matching them it look like <[Rr][Ee][Ff][^>]*>...

Also, note that my tools aren't bot, they are merely tools to assist humans in editing quicker. — Dispenser 01:12, 2 February 2010 (UTC)

Not trying to be a smartass or anything, but:

1. I don't see what you're getting at. By this rationale, everything on WP should be all-lower-case other than that which must be capitalized in articles in order to be grammatical. I.e., we should refer to the "manual of style", and wp:afd and arbcom, and so on, exclusively.
2. So? Human editors don't.
3. So? Human editors don't do this.

Basically, you've not actually responded to the concern I've raised, which is that your tools are enforcing a personal spelling preference of yours by basically editwarring with thousands and thousands of editors, triggering innumerable watchlists in the process. I have no issue with changing <REF> to <ref>, since upper-casing of element names is bad form in XML. That hasn't anything to do with whether first letter of template names are capitalized, which by real human editors is about half of the time if not more. I mean, I see what you're getting at, but it doesn't match the reality of editor behavior, and the only thing you have to do to get around it is use a [Cc] regexp. It's not a huge deal to me, I just don't see that this lower-case "fixing" actually serves a useful purpose that outweighs its very considerable unhelpful side effects. I.e., the ends are not justified by the means. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 17:19, 2 February 2010 (UTC)

There is far more debate about XML-style tags then there is about this. The documentation uses all lowercase, so that people don't have to memorize archaic casing. To simplify coding for me, I convert all variants of the template into a single standardize form. We know that standardization and consistency helps get work done faster and we know from our usability studies that users learn by copying code off other pages. You have failed to make a case about why I should take time to accommodate this non-standard behavior, only suggesting that it should be allowed since the software allows for it.

Since I log the edits my tool makes we can also look at how this impacts pages. Now I've semi-randomly selected 30 edits to see what real world usage is like, There were 13 edits where there was no citation style prior (bare URLs), 15 where there was no change (all templates using lowercase), and there were two one with a dominate lowercase naming and the other using mainly text ref. So drawing from the data it seems that the lowercase style is far most prevalent. — Dispenser 20:07, 2 February 2010 (UTC)

Question

Hi Dispenser, I wonder if you could help me? I've been using some of your scripts; Reflinks, Checklinks and most recently Dab solver.

I think these tools are excellent, by the way. Thanks for creating them. :-)

The question I would like to ask refers to the ‘Apply common fixes’ feature of the scripts:

One of the changes made (by Dab solver, for example) is the {{reflist}} template is changed to {{reflist|colwidth=30em}}

I agree with this change as I believe it make pages display in a neater fashion. I've also see this edit done frequently by other users.

It's just that I've been asked by another editor where to find the relevant Wikipedia guidelines for this particular change.

I'm not sure myself and I thought you might know the answer. :-)

Kind Regards -- Marek.69 ^talk 20:58, 2 February 2010 (UTC)

There isn't any policy or guideline that specifics how to format the references part. Editors generally fall into one of two camps, <references /> and {{reflist}}. The former claims that their is no reason to specially treat that section and it is just as important as the rest. While the latter want to shorten the "gunk", as it is known outside of Wikipedia, in the article by using a smaller font and columns. They apparently also have trouble with scroll wheels. There's also a debate about font-size since at certain sizes IE renders text larger than others. Now, the only reason why it changes it is to beat the inevitable editor who uses {{reflist|2}} which doesn't scale as well across different screen sizes. — Dispenser 21:59, 2 February 2010 (UTC)

OK, I understand.

Thank you for your help, Dispenser.

Kind Regards -- Marek.69 ^talk 23:05, 2 February 2010 (UTC)

Incorrect capitalisation in piped references

I've noticed that capitalised piped references have been converted to lower case in some edits, and gave been given to understand that this is due to the use of Reflist. As an example, a reference to a [[Coach (carriage)|coach]] gets changed to [[coach (carriage)|coach]] with leading lower-case letter. As the title of the article being referenced is "Coach (carriage)" the capitalised form shows the actual title of the article, and is correct. I was under the impression that this invoked an unnecessary redirect from "coach (carriage)" to "Coach (carriage)", but this may not be the case; even so this unwarranted case conversion is undesirable. There is no issue of an originally incorrect pipe in these cases; Reflist simply lower-cases the leading letter with no other change. If other tools than Reflist do the same the same remark applies. (added later) While other issues have been raised about capitalisation what originally brought me to this page), I think this is a very clearcut case of a tool introducing actual errors.(Pol098 (talk) 11:05, 8 February 2010 (UTC)

The software is case-insensitive for the first letter on the English Wikipedia, so no redirect is being used. Using a pipe trick, it can be written as [[coach (carriage)|]]. The logic is very simple and I haven't found any errors that weren't with the label text. Plus, this change makes easier reading of wikitext.

(for future tools) Also, remember that WP:redirects are cheap, if it make sense to use a redirect that describe the term more accurately then why not use it? It could even save work in the long run, for example, when a character becomes more important and their section is expanded to an article, it'll save the work of re-targeting pages. — Dispenser 15:57, 8 February 2010 (UTC)

I'll merely say that some people (maybe a lot) consider that an article with the title "Coach (carriage)" should be piped with the title exactly as the article title, without changing the case of the leading character. I accept that there are no considerations of efficiency, but an edit that simply changes [[Coach (carriage)|coach]] to [[coach (carriage)|coach]] is unnecessary and arguably wrong. Certainly someone entering a piped reference may choose not to bother with a capital, or to use the pipe trick; but once a link is there with a capital there is no valid reason to change it to differ from the actual article title. You say "I haven't found any errors that weren't with the label text". I would interpret this to mean "the tool only changes labels (links) which are clearly wrong (and in these cases also, as it happens, changes the initial capital to lower case)". Maybe I don't understand what you mean, but there are plenty of cases where Reflink has taken a valid and correct link and changed the initial capital to lower case, with no other change - not originally "clearly wrong". I think we'll have to agree to disagree here, I'll just put on record that in my opinion, for piped references at least, Reflist changes a correct piped link to one that is arguably not correct (though it does work). An example (the coach link mentioned above) is in Marek69's Reflist edit to Hungary. This isn't an isolated instance; there are a great many similar cases. Pol098 (talk) 16:39, 8 February 2010 (UTC)

I also dislike the decapitalization in piped links. I came here because of [1] where PDFbot did it dozens of time. It makes it harder to check diffs for the real changes. And I personally find a piped link more readable when the target is capitalized. I suspect many others do the same. Is the bot approved for this? Is there any guideline or similar supporting the decapitalization? I don't consider it support either way if an example in a guideline or documentation does or doesn't use capitalization without making a recommendation. PrimeHunter (talk) 03:43, 9 February 2010 (UTC)

I know of no guideline which says that in piped links, a lowercase initial letter is preferred to a capital, nor vice versa, except where the pipe trick is employed: but guidelines do exist that state that it doesn't matter. See, for example, Wikipedia:Linking#Piped links; Wikipedia:Naming conventions (capitalization)#Software characteristics; Wikipedia:Piped link#When not to use. However, it's best to avoid change for change's sake: since [[Coach (carriage)|coach]] and [[coach (carriage)|coach]] are visually and functionally identical, there is absolutely no need to change one form to the other. The same goes for template and category names. --Redrose64 (talk) 12:16, 9 February 2010 (UTC)

Reflinks

May I commend you on this tool and also on your userpage, first is extremely useful, second is just cool. I'd just like to note that smackbot goes crazy whenever I use reflinks, changes all the "cite" in the citation template to "Cite" with a capital C (eg.[2]). I'm assuming changing this would be a very simple operation, so if it is it may be worth it to alter it. If it's not that simple (my poor knowledge of computer editing is revealed) then it's probably not worth the bother! Chipmunkdavis (talk) 13:44, 27 September 2010 (UTC)

Capitalization

Further information: [[:Wikipedia:Administrators' noticeboard/Incidents#Please block Rich Farmbrough - thousands of unnecessary capitalization changes|Wikipedia:Administrators' noticeboard/Incidents § Please block Rich Farmbrough - thousands of unnecessary capitalization changes]], [[:Wikipedia:Administrators' noticeboard/IncidentArchive639#Battle of the bots: Reflinks vs. SmackBot|Wikipedia:Administrators' noticeboard/IncidentArchive639 § Battle of the bots: Reflinks vs. SmackBot]], [[:Wikipedia:Bot owners' noticeboard#Filling recent changes with pointless changes|Wikipedia:Bot owners' noticeboard § Filling recent changes with pointless changes]], [[:Wikipedia talk:Citation templates#Template caps: "Cite" or "cite"?|Wikipedia talk:Citation templates § Template caps: "Cite" or "cite"?]], and [[:User talk:Dispenser/Reflinks#Reflinks conflict of interest with other bots|User talk:Dispenser/Reflinks § Reflinks conflict of interest with other bots]]

I strongly recommend that you come to one of the (several) noticeboards where this is discussed. The situation is coming to a head. I tried to point out, a week and a bit ago, that this sort of thing is usually sorted out informally by 'bot owners. But my argument is being somewhat undermined by the fact that it still hasn't been. Uncle G (talk) 15:46, 29 September 2010 (UTC)

FA cleanup

Thanks for everything you do, everywhere ! You are a busy bee ... SandyGeorgia (Talk) 19:48, 3 February 2010 (UTC)

Thanks for fixing my boxes, Dispenser. Will have to get around to reviewing and writing some FAs again soon. Amerique^dialectics 21:45, 3 February 2010 (UTC)

etwiki log

hi, etwiki log has weird content tools:~dispenser/view/File_viewer#log:coord-etwiki.log (?) --WikedKentaur (talk) 16:55, 6 February 2010 (UTC)

Its an error that new code I wrote didn't work out as well as I intended. I've regenerated late night's run. — Dispenser 17:48, 6 February 2010 (UTC)

Thanks. --WikedKentaur (talk) 08:17, 7 February 2010 (UTC)

noreferences.py

Hi,

I would like to ask you to update the noreferences.py file you use for your Checklinks service on the toolserver. Recently, the Hungarian Wikipedia has replaced <references> with a template with Hungarian name -- until you update the template definitions in your version of noreferences.py the online version of the reflinks bot will believe that there is no references section and will try to add it for 99% of all pages on huwiki.

Thank you, --Dami (talk) 19:01, 8 February 2010 (UTC)

I've updated noreferences.py to the latest SVN revision. Thank you for your reminder. — Dispenser 19:22, 8 February 2010 (UTC)

PDFBot broke pdf links at Amber (color)

The bot recently made this edit Amber (color) Remove incorrect {{PDFlink}} which converted two pdf links in the external link sections into references causing a cite error. Might have been a result of the links becoming dead confusing the bot, I've since updated the links. PaleAqua (talk) 02:53, 9 February 2010 (UTC)

Reflinks

NYT API

I recommend that you pull information for NYT articles using their article search API rather than the current screen scraping. A simple json query would return author, publication date, publisher, geo information...etc. Let me know if this is worth devolping...Smallman12q (talk) 19:32, 11 February 2010 (UTC)

There is a discussion at Wikipedia:Village_pump_(technical)#Reference.2FCitation_Auto_Citation you may be interested in.Smallman12q (talk) 22:01, 11 February 2010 (UTC)

I would be interested in a collaboration in the future. However, I need to finish and fix up the other tools of mine before I can embark on this challenge. — Dispenser 19:14, 26 February 2010 (UTC)

Small bug

I believe there might be a small bug in Webreflinks when dealing with bot generated titles. See this, for example. It doesn't strip the first two characters of the comment. Nymf _talk/contr. 14:35, 26 February 2010 (UTC)

I have never been able to figure out where this odd stripping happens, but since tool is long over due for a rewrite I've just put in a hack. — Dispenser 19:14, 26 February 2010 (UTC)

WP:ACTOR and non-standard Filmographies (redux)

Hi. I love your tools; I use reflinks daily. Thanks.

I just read the long thread at:

• MediaWiki talk:Common.css/Archive 9#Header color in wikitables

I've stumbled into this issue during the last few days, having worked on maybe 20 filmographies attempting to clean-up poor markup and styling.

I think the time has come to sort the issue of non-standard styling in the name of a wikiproject. The whole initiative originating there reeks of article ownership, wikiprojects as "governing" bodies, and impeding site accessibility in the name of meretricious appearance.

I think the appropriate outcome is filmographies either implemented as bog-standard wikitables, or a suite of templates/css that allow central control of the styling (i.e. to site-wide conventions).

The current point of discussion is at:

• Wikipedia talk:WikiProject Actors and Filmmakers#Filmography

I hope to see you there.

Cheers, Jack Merridew 07:20, 7 March 2010 (UTC)

• The thread above has rolled well-along and I've a request at WP:AN#Stuck RfC @ WT:ACTOR#Filmography asking that it be reviewed. Cheers, Jack Merridew 03:40, 31 March 2010 (UTC)

type:settlement in GeoHack

Would you be willing to add type:settlement to GeoHack, as a synonym for type:city? --Stepheng3 (talk) 16:42, 8 March 2010 (UTC)

NYC Wikipedia Meetup Sunday, March 21

{{Wikipedia:Meetup/NYC/Invite/March 2010}}
This has been an automated delivery by BrownBot (talk) 15:15, 13 March 2010 (UTC)

Usability of GeoTemplate

Comments on the usability of {{GeoTemplate}} (the page listing mapping services found by clicking on coordinates in articles) are invited, at Template talk:GeoTemplate#Usability redux. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:50, 17 March 2010 (UTC)

Alt text

Hi Dispenser, am I right in thinking you operate the altviewer that produces this? SlimVirgin ^{TALK contribs} 03:27, 22 March 2010 (UTC)

Yes. — Dispenser 03:48, 22 March 2010 (UTC)

Would you mind updating it to reflect the current state of affairs? WP:ALT is no longer a guideline or in the FA criteria because we found we'd been doing it all wrong, and that in fact not all images should have additional alt text. They do need an alt attribute, but it can be an empty one, or "alt=refer to caption", and when we do have alt text it arguably shouldn't summarize an image's appearance, but should be a substitute for its function. You can see some discussion here at the MoS, which links to further discussions if you want to follow them. See the lead of WP:ALT for the current advice.

Your current text is: "Alternative text (alt text) is meant for readers who cannot see an image, such as blind readers and readers who use a text or mobile browser. It should summarize an image's appearance, and should not duplicate its caption. Every image should have alt text, except for purely decorative images, which should instead have "|link="."

Could you change that to something like: "Alternative text (alt text) is meant for readers who cannot see an image, such as blind readers and those who use a text or mobile browser. For more information about how and when to add it, see WP:ALT." SlimVirgin ^{TALK contribs} 13:47, 22 March 2010 (UTC)

I have updated the software to better identify different types of default alt text. It will classify into the following types:

• Default alt text Alt text was automatically produced by MediaWiki. Thumbnails and videos are typically blank (there are exceptions); for inline images it is the file name.
• Caption alt Special case for inline images, if a caption is specified it will be used on both the image's alt attribute and the link's title attribute. (NOTE: The general behavior of screen readers is to read the alt attribute, if none exists it will fallback to any title attributes.)
• User defined User a |alt= was present in the markup. Note it can be defined to to an empty string (|alt=) or a blank space (|alt= ), which is different from not defining it at all (i.e. no |alt=, aka NULL value). A special case been made for flags (titlealt) where the alt text matches the target article title to include it to this type.
• Unlinked Appear if dimensions are larger than 100px on a side, was not generated MediaWiki, and was not match by any of patterns.

The problem with the argument is that with our image policy nearly all images are for illustration purposes. And if not we should question if the image should really be included. I've tweaked the description while still (trying to) retaining instructions on what users should do with the image. — Dispenser 04:15, 14 April 2010 (UTC)

Thanks for providing the image delinking script. It provided useful information. Graham has helpfully explained what screen readers read out. It appears that if the HTML alt text is blank (or a space), and the image is embedded in a link (most are), then the reader software tries to find something to describe the link. One reader program tries to use the link title attribute if available (which is defined for the enlarge icon on thumbnails) and the filename otherwise, but JAWS always uses the filename. Currently, with most alt text blank, screen reader users suffer the filename twice for thumbnail images. IMO we should reopen bug 21042 as bug but 371 was incorrect. The comment on 21042 makes two errors: that only blind users need alt text and they won't care about the link, and that setting alt="" for a linked image is an improvement.

I keep learning new things about alt text and accessibility. If you are able to review the WP:ALT "guideline" and help the discussion with your knowledge on the subject, that would be great. Colin°^Talk 18:46, 19 April 2010 (UTC)

geocoordinate tools

Last night's coord-enwiki.log failed with a Python error, and today geosearch.py finds less than 1000 coordinates. I've heard a rumor of toolserver breakage. Do you know what's going on? --Stepheng3 (talk) 20:11, 26 March 2010 (UTC)

See WP:VPT#Stable Toolserver link breakage, they changed the address and I haven't changed the tools. I've updated the hack used in ghel for the short URL in so its no longer matches incorrect URLs. I had accidentally ignored the error this morning with the storm of SQL server missing (hyacinth). (Sorry for the messy message, traveling now) — Dispenser 04:01, 27 March 2010 (UTC)

I see how the name change fouled things up. Coord-enwiki.log succeeded last night, and I'm grateful for that. (If you've noticed the report getting shorter over the past few months, that's had a lot to do with me.)

I don't understand about ghel and hyacinth, but I'm glad to know that you're on top of things despite travel. I hope for resolution of the geosearch.py breakage in the near future.

You really should consider changing your username to User:Indispensable. What would I do without you? --Stepheng3 (talk) 21:41, 27 March 2010 (UTC)

I see that geosearch.py is still not working properly. In fact, the number of coordinates visible to it has fallen below the 500 mark. Is there a workaround? --Stepheng3 (talk) 05:57, 2 April 2010 (UTC)

I see that it's fixed. I assume you had something to do with it, so thank you. --Stepheng3 (talk) 23:04, 17 April 2010 (UTC)

User:Dispenser/Checklinks/json.js dead?

User:Dispenser/Checklinks/json.js is dead I guess? Since tools:~dispenser/cgi-bin/jsonchecklinks.py no longer exists. Gary King (talk) 05:07, 6 April 2010 (UTC)

It didn't make sense to continue maintaining it. Only people six users ever installed the script and months between uses. I might bring it back for something like integration with the peer reviewer script. — Dispenser 22:51, 12 April 2010 (UTC)

locateCoord.py doesn't work anymore for many languages?

Hi Dispenser, I just want let you know that for some reason, recently locateCoord.py doesn't work anymore for some lang database. For example, english does work, italian doesn't (see this page for example) I've seen you edited something recently. By the way, thanks for the excellent tool! I always use it. Keep up the good work! :) —Preceding unsigned comment added by 93.48.47.222 (talk) 22:01, 16 April 2010 (UTC)

I fixed the problem shortly after you reported it. IIRC, an error was thrown in some new code went when the database we being rebuilt that night. — Dispenser 01:04, 28 May 2010 (UTC)

Adminship

It looks like you've been doing some good work, know your way around the 'pedia, and have been here for awhile. And it's always good to have technically-savvy admins on board. Would you care to be nominated for adminship? Thanks, Tisane (talk) 12:49, 8 May 2010 (UTC)

While I have been thinking and reading up on this over the past several weeks, I have to respectfully decline. I have been trying to free up more time for other projects, writing is difficult for me and I do not think I can respond in a timely manor to questions posed. I have no significant content contributions to articles and much of my work has been with templates, so my function as admin would likely be limited to improving templates and completing editprotected notices. While this is not a bad thing, the regulars at RfA like seen familiar names and often drill on their area of expertise of deletion, counter vandalism, dispute Resolution, and content creation. Perhaps, at some later point in time we can re-engage in this. — Dispenser 20:02, 27 May 2010 (UTC)

NYC Wikipedia Meetup Saturday, May 22

{{Wikipedia:Meetup/NYC/Invite/May 2010}}
This has been an automated delivery by BrownBot (talk) 21:00, 12 May 2010 (UTC)

GeoHack, map for Slovenia

Hi there.

I left a message to Magnus Manske about an addition of a map service for Slovenia, but I see he hasn't logged in for two weeks already, so I decided to write you.

It's a free 3D visualization based on NASA's World Wind in Java, but it has aerial imagery from the Slovene government. This is the entry:

|-
| Gaea+
| colspan=3 | [http://gaeaplus.si/webstart?version=gaeaplus&command=FlyLookAt;{latdegdec};{londegdec};400;WGS84;0.0;45.0; 3D Satellite]

You can give it a try here

Thanks, Mariano(t/c) 13:41, 26 May 2010 (UTC)

I do not have Java install so I cannot view to see if it is any good. However, I recommend leaving a note on Template talk:GeoTemplate; or if you're bold enough, to add the link to the page. The GeoHack team is not actually responsible for the content or links which appear when using GeoHack, that is up to each individual language. — Dispenser 19:05, 27 May 2010 (UTC)

Thanks for the feed back, will first leave a message on the talk page and if everyone is OK with it I will also add it. Mariano(t/c) 08:26, 28 May 2010 (UTC)

Geodata–check tag

Thanks for your fix–up. I boldly made that change, but I like yours better. However, when I click on a set of coords and the GeoHack page comes up I'm not seeing the notes that you added to {{GeoTemplate}}. What am I doing wrong, or when do they show up? — TRANSPORTERMAN (TALK) 16:37, 28 May 2010 (UTC)

Because the cache expiration for GeoHack seems to be longer some reason now a days. Might be related to the switch Vector and the hack I put in since they removed the <!-- start content --> markers. — Dispenser 18:01, 28 May 2010 (UTC)

Thank you

		What a Brilliant Idea Barnstar
		for Dab solver DuncanHill (talk) 11:19, 5 June 2010 (UTC)

Geohack/osm

Hello, I make some changes on GeoTemplate/osm but couldn't see an effect a half hour later. Also reload and purging of template doesn't help. In the source code I can't detect the problem (caching or so). Could you help? --Kolossos (talk) 10:49, 8 June 2010 (UTC)

I heavily suspect it's the &useskin=monobook hack I put in after the usability team changed skins. Vector is missing the <!-- start content --> and <!-- end content --> markers and is breaking tools because of it. When asked, they mention to use the API but admit, besides the poor examples, that it is not cached as well as the squid servers. A bug will need to be filed to add the tags to vector. — Dispenser 13:43, 8 June 2010 (UTC)

Metawiki seems still running on monobook. I try a wget to http://meta.wikimedia.org/wiki/Template:GeoTemplate/osm from wolfsbane and get an actual file with start-/end-content-tag back. The interesting thing is that I don't get a mistake, I get an old but correct answer on geohack. Do you have an other idea? --Kolossos (talk) 17:27, 8 June 2010 (UTC)

It is related to the skin issue since the squid caches don't properly empty when a non-standard URL is used. I'll talk to the usability team and other MediaWiki developers next week about adding the tags back in. — Dispenser 20:30, 12 June 2010 (UTC)

The problem seems still to exist. Perhaps is the way over the API much cleaner. So we also need to transfer only the pure text of the page and reduce so the transfer volume. --Kolossos (talk) 20:52, 2 August 2010 (UTC)

It sounds like a better solution until your realize squid caches are far more reliable and less taxing than the API. I'll have another go about nagging the usability team to include the tags. — Dispenser 21:33, 2 August 2010 (UTC)

Filmography table begin

Hi,
could you comment at Template talk:Filmography table begin, please?
Thanks, Amalthea 11:21, 11 June 2010 (UTC)

Barnstar

Hi, Dispenser. I know you prefer your work to tell people who you are, but some things need to said as well. --Philcha (talk) 06:27, 21 June 2010 (UTC)

		The Invisible Barnstar
		Dispenser's actions are worth 1,000 words

THANK YOU SO MUCH FOR YOUR HELP

RE: [3]

		What a Brilliant Idea Barnstar
		This Barnstar is awarded to Dispenser, thank you for taking the time to help me. You are truly an asset to the project. Keep up the wonderful work.  :) Adamtheclown (talk) 02:04, 25 June 2010 (UTC)

Readability

Hey, I noticed you shut down the Readability Tool. Could you resurrect the non-readability parts of that tool, such as # of links, # of interwikis, # of references, maybe even size, and place them in a new tool? Thanks, Purplebackpack89 15:59, 29 June 2010 (UTC)

No, one reason why I took the tool down was because I didn’t want to maintain the wikiparser behind it. You could ask User:Dr pda if he would implement these statics into User:Dr pda/prosesize.js, which is also better supported and documented. Also, why would you need to know the number of links? — Dispenser 22:14, 29 June 2010 (UTC)

If a tool was not implemented correctly, that's one thing. But please don't assume that others are incapable of using a tool with intelligence and discretion. We are, and some of us know more about reading research than you might imagine. Macdonald-ross (talk) 17:11, 29 June 2010 (UTC)

The method used by the algorithms is not a very good proxy of a person's understanding of a word. And the purpose of readability and our target audience in Wikipedia is unclear at best. Most users are unfamiliar with the subject of readability and do not understand the shortcomings. Seeing an automated tool gives the impression that readability is an exact science.

For example, an algorithm was found to accurately determine featured status mostly by counting the number of references. This creates a problem if it were easily accessible, it would become a pseudo-requirement and people would start gaming the metrics.

Nevertheless there were reason specific to the implementation. The syllable counter was very simplistic and it can be compounded such that an word often-repeated can throw off the score significantly (almost chaotic in nature). People knowledgeable on the subject of readability are not aware of the exact triggering conditions and may trust a particular score over another. This leaves the people who do understand the subject and limitation better served copy and pasting a paragraph, section, or the full article into a program. — Dispenser 22:14, 29 June 2010 (UTC)

I agree with MacDonald...if you have lost confidence in the readability tool, you should pass the code of the tool, in whole or part, to someone else, rather than deprive the community of the tool. I'm not sure that the other tool you mention will work for me or other users for a plethora or reasons; chief among them being the js suffix often means trouble for Internet Explorer users Purplebackpack89 16:39, 30 June 2010 (UTC)

It doesn't make sense to pass on a broken program (and accompanying framework) when there are better open source implementations such as GNU's "Style and Diction". The prosesize script has an IE-only feature, so I doubt that there will be trouble when install correctly. — Dispenser 20:23, 30 June 2010 (UTC)

glupt.log

Very nice. Thank you! --Stepheng3 (talk) 01:37, 4 July 2010 (UTC)

Checklinks Reflinks issue

I ran theses tools on the article internet and both tried to replace on internal wiki link to BBC#interent with an external url. This doesn't seem to be the intended behavior. They others work fine. Phatom87 ^{(talk • contribs)} 20:18, 5 July 2010 (UTC)

It is actually a redirect which points to the section BBC#Internet not an external url.

Commonfixes selects the best matching redirect, even if it is nonsensical. It is easier to correct a few redirects (which needs to be done anyway) when spinning off a section then it is crawling through articles to fix. I would suggest creating BBC Internet, but it is already spun off as BBC Online. — Dispenser 01:45, 29 July 2010 (UTC)

GeoHack

I don't know if your a go to person for this kind of a problem or not, but I've noticed your account associated with GeoHack. It seems there is an anomaly in the way GeoHack URLs are formed. See these examples in my user space. The first uses unsigned decimal degrees and produces a URL containing "37.2971555_N_118.6587268_W". The second uses signed decimal degrees and produces a URL containing "37.2971555_N_-118.6587268_E" which is a bit unusual. There is an ongoing discussion at User talk:Justin.Johnsen. I don't know if this is related to the that discussion or not. Best.  –droll [chat] 05:12, 19 July 2010 (UTC)

Double negatives are expected and valid syntax and are correctly handled in GeoHack, ghel, and many other programs. These programs also correctly handle -10°-10"S and the semicolon separator syntax, lat;lon_params. URLs are transcribed as directly as possible to reduce template complexity and enable geosearch.py to work well. Reusers should use one of the robust libraries or data sets we provide. — Dispenser 21:59, 28 July 2010 (UTC)

Idea for new tool

Hi, Dispenser. For finding and fixing WP:D#Incomplete disambiguation pages, could you create an automated list of disambiguation pages that have (foo) in the title, where (foo) is anything besides (disambiguation)? Thanks (for considering it, not necessarily implementing it). -- JHunterJ (talk) 04:24, 23 July 2010 (UTC)

I'm not sure exactly what your trying to achieve. Are you looking for foo_(bar) when foo_(disambiguation) doesn't exist or is a redirect to foo? Or when foo_(bar) has a high link overlap with foo_(disambiguation)? — Dispenser 16:01, 28 July 2010 (UTC) PS: I just ran a query excluding redirects there are 2,000 dab pages with something besides (disambiguation). — Dispenser 16:01, 28 July 2010 (UTC)

Readability tools

I'm told you hosted a couple of (ease of reading) Readability tools, for example, the flesch kincaid analyzer. Would like to have these tools back if you are inclined, so we can plug Simple English articles in them as a QC check. Is this possible, and with your help? Thank you for all you do. Best, NonvocalScream (talk) 10:49, 29 July 2010 (UTC)

No, see #Readability. — Dispenser 12:43, 29 July 2010 (UTC)

Optimize the OSM-map on geohack

Hello, many people really like the OSM-map in the german Geohack. But we should try to reduce the load on OSM-Servers from our side. In the moment we use standard embed.html from OSM, the default buffer parameter is so by default 2. The effect of this parameter is that you see only 6 tiles on geohack-screen but download 42 tiles because you have on each side a buffer 2 columns. This produce a lot of traffic and in the most case the users simply click to a link without use this tiles. So I copy the embed.html to toolserver and set buffer to 0[4]. So could you please copy this file to Geohack-SVN and integrate the new file in german Geohack? Thanks. --Kolossos (talk) 21:07, 2 August 2010 (UTC)

de:MediaWiki:GeoHack.js is the file that you want to change. Some other ideas while you're tinkering around: 1) a button to expand the iframe, 2) load a KML file with nearby articles, and 3) alternative tiles (like satellite). — Dispenser 22:02, 6 August 2010 (UTC)

Thanks. It's now activated. A script function increase the buffer form 0 to 2 with the first paning/zooming. I also minimize the OpenLayers.js with openlayerer so that it now only transfer 87 kB instead of 171kb. I use php to transfere OpenLayers.js in compressed version if possible. --Kolossos (talk) 10:44, 12 August 2010 (UTC)

Wiki-Conference NYC (2nd annual)

Our 2nd annual Wiki-Conference NYC has been confirmed for the weekend of August 28-29 at New York University.

There's still plenty of time to join a panel, or to propose a lightning talk or an open space session. Register for the Wiki-Conference here. And sign up here for on-wiki notification. All are invited!
This has been an automated delivery by BrownBot (talk) 15:15, 3 August 2010 (UTC)

Reflinks tool opposition

Not by me, though; I must show in any usage logs, a lot. I thought you might care to comment about this thread:

• User talk:Gimmetoo#Asinine actions

Sincerely, Jack Merridew 19:34, 12 August 2010 (UTC)

care to comment on this?

Template talk:GeoTemplate#Ireland. --Stepheng3 (talk) 18:45, 17 August 2010 (UTC)

Geohack is not up-to-date

Some edits like http://de.wikipedia.org/w/index.php?title=Vorlage:GeoTemplate&diff=next&oldid=77717978 doesn't have an effect to german Geohack also after a week. That's too long. Where is the Geohack get it old data after such a long time? --Kolossos (talk) 08:36, 18 August 2010 (UTC)

Dispatch

This is relevant to you. ResMar 15:17, 22 August 2010 (UTC)

GeoHack broken

I just discovered that http://toolserver.org/~geohack/geohack.php appears to be somehow transcluding the main page of enwiki. Please look into this. --70.36.159.31 22:24, 23 August 2010 (UTC)

Never mind. It's working again. --Stepheng3 (talk) 22:28, 23 August 2010 (UTC)

Thank you

Thanks for fixing my user page. Anna Frodesiak (talk) 00:59, 25 August 2010 (UTC)

Entry descriptions on disambiguation pages

Please be advised of this guidance from MOS:DAB#Individual entries: "The description associated with a link should be kept to a minimum, just sufficient to allow the reader to find the correct link. In many cases, the title of the article alone will be sufficient and no additional description is necessary." I reverted your recent edit to Boo because it seemed to consist, at least primarily, of adding lengthy descriptions to every entry when they weren't necessary. For example, if there's only one TV series on the page, it's doubtful that a user looking for that TV series will need to know that it's "a children's television series shown in the United Kingdom on the CBeebies channel, and occasionally on BBC One or BBC Two" in order to figure out that Boo! (TV series) is the right article. Propaniac (talk) 18:03, 25 August 2010 (UTC)

I had just finish implementing an automated description system. Call me lazy, as I only spent time cleaning up other MoS issues on Boo. I suspect you are too, as you didn't edit/remove the descriptions.

I have to disagree with the MoS on the descriptions. There isn't enough information to pick the correct entry if the only knowledge was that an actor had a role in Boo. This is compounded further if that the correct entry was a red link, which are treated like a bastard children around here. These problems have become more apparent with the tools I have been developing. So I've reverted your edit, but copy edited the descriptions to be shorter. — Dispenser 19:14, 25 August 2010 (UTC)

Are you saying that your strategy is to automatically add the first sentence of the linked article to every single entry? Please, please don't do that. Short descriptions are one thing, but what you're talking about is going to create so much excessive text and clutter. Also, while I can mostly live with the current Boo page, please don't add descriptions that simply repeat the article title, i.e. ""Boo" (CSI: NY episode), an episode of CSI:NY". I'm trying to be polite, but it is my belief, which is supported by the MOS (which is an official guideline), that a cleaner and more condensed disambiguation page is easier to navigate, and also that doing things the wrong way only encourages others to think that's the right way. Propaniac (talk) 19:48, 25 August 2010 (UTC)

I have since added "lazy" buttons to remove untouched modified text. — Dispenser 21:47, 14 October 2010 (UTC)

Signpost

I think you should take over the tools article. I think that most of the stuff under possible additions is way out of scope and the IE library thing sounds like a coding session not a Dispatch. But I don't care really. I'm tired to death of this idiotic argument. Maybe you'll have better luck running a wall head on then I did. Regards, a tired ResMar 19:49, 31 August 2010 (UTC)

Thanks for fixing the Reflinks entry (I added it in the last minute before publication, to make the article more consistent and complete - although I had already suggested to cover it on August 20 [5], at the very beginning of the whole saga... ;)

I hope you will contribute to the next edition too, see also Newsroom. Regards, HaeB (talk) 02:49, 7 September 2010 (UTC)

DOIbot

I've recently noticed an issue with the bot, namely that when going over coauthor= fields in references it takes all of the enclosed names and encapsulates them into last1=, first1=, last2=, first2=, etc, which is exceptionally clunky once you get to 5+ such params. Sounds like it needs a code alteration :) ResMar 19:02, 9 September 2010 (UTC)

Dispatch p. 2

You indicated you wanted to do some work on it; can you do this before publication...? ResMar 21:33, 16 September 2010 (UTC)

CatScan 2

Thanks for your post on the technical pump; a potential indeed for rich queries. I just tried the tool out with a few combinations, and I don't know if you're aware but there can be a blank white screen returned if the query takes too long. (Perhaps if the Db query times out?) In case it's helpful to you, the query I experienced this with is below.

Categories

2000s singers

Living people

2010s singers

Negative categories

Girl groups

The first three or four times, it went away for about 2 mins before clearing the page to a blank white page. Tried it a bit later, and it worked, so I'm guessing I hit a busy time on the server. You may wish to consider using a longer timeout somewhere, or handling a timeout by displaying a timeout message.

Thanks for the tool; hope this info is useful to you. PL290 (talk) 18:02, 21 September 2010 (UTC)

I'm an not the author of that tool, that's User:Magnus Manske. But developing similar tools I know that it is typically a result of 1) user's browser times out before completion and 2) that SQL tables were not cached in memory. — Dispenser 18:49, 21 September 2010 (UTC)

FA tools

Would you mind having a look at this edit on the FA tools template, designed to prevent unrelated previous FACs appearing? (SandyGeorgia requested me to ask you to check.) Ucucha 23:55, 24 September 2010 (UTC)

Query for bot help

Wikipedia:Bot_requests#Query_on_AFD_statistics - any chance you could help out with this? :) Thank you for your time, -- Cirt (talk) 16:25, 28 September 2010 (UTC)

I have no experience in anything deletion related. To my knowledge, there have been several changes to the format and process over the years. There is undoubtedly bad data from things are speedy kept. So my suggestion is to contact someone who is familiar with the deletion process and put together a criteria list. A name that jumps out at me is MZMcBride, who has extensive experience and maintains Wikipedia:Database reports. I would also advise finding a statistician to perform T-tests and other statistical measurements. — Dispenser 20:48, 28 September 2010 (UTC)

Wikipedia NYC Meetup Sat Oct 16

{{Wikipedia:Meetup/NYC/Invite/October 2010}}
This has been an automated delivery by BrownBot (talk) 16:04, 9 October 2010 (UTC)

API edit conflicts

Hi, did your comment at Wikipedia:Village pump (technical)/Archive 126#Edit conflict problem mean that people editing through the API are not getting checked for edit conflicts? Regular editors still get e/cs from what I can tell. EdJohnston (talk) 15:56, 14 October 2010 (UTC)

It's an issue with the HTML interface, my tools given an empty wpEdittime. For anonymous users, it triggers an edit conflict; for logged in users it skips the edit conflict detection. I speculated that edit conflict detect is likely broken in several other ways. — Dispenser 20:06, 14 October 2010 (UTC)

Wikipedia:Deletion review/Log/2010 October 23#MediaWiki:History short

Hi Dispenser. Because you participated in Wikipedia:Miscellany for deletion/MediaWiki:History short, you may be interested in Wikipedia:Deletion review/Log/2010 October 23#MediaWiki:History short. Cunard (talk) 06:48, 24 October 2010 (UTC)

Wikipedia geocoding

(Not sure if you get my mail.) Hello Dispenser, you know that I use your extraction for my Wikipedia-World project and I like this way of usage very much. I found two problems:

• Could you make your script so that it is sure that the gc_name parameter is correkt UTF-8?

There was an article in the japanese Wikipedia with a name-parameter longer than 255 bytes. So MySQL cut such things after 255 bytes in the middle of an utf-8-character and take it without problems. But PostgreSQL want to be sure that it is correct UTF-8 and breaks the import complete. I hope you understand that it makes no fun to find and repair a mistake in line 21.345 of an japanese dump-file.

• You have 4 digits in your extract, that's in the most cases perhaps enough but not enough for the crazy german guys ;-)

In the german Wikipedia we have some extreme things like geocoding single figures on the roof of a church [1] So, please give me 6 digits.

Thanks

Greetings Kolossos

[1]

• http://de.wikipedia.org/wiki/Liste_der_Mattielli-Statuen_der_Katholischen_Hofkirche_in_Dresden
• http://toolserver.org/~dispenser/cgi-bin/locateCoord.py?dbname=coord_dewiki&lon=13.7371&lat=51.053173&range_km=.05
• http://maps.google.com/maps?q=http%3A%2F%2Ftoolserver.org%2F~para%2Fcgi-bin%2Fkmlexport%3Fproject%3Dde%26article%3DListe_der_Mattielli-Statuen_der_Katholischen_Hofkirche_in_Dresden

ALERT: Punctuation after the references is the dominate format!

Just noticed this message in reflinks, and I think I know what you mean; I think you mean "dominant" instead of "dominate".

Also, you might want to explain it a little; I just explained it to someone like this;

I suspect what they mean is that, in that article, people have put the refs *before* commas and periods, not after them and therefore, when editing it, they want you to keep to that same style

chzz is 194<ref> Book </ref>.    instead of the usual    chzz is 194.<ref> Book </ref>

 Chzz  ►  14:53, 25 October 2010 (UTC)

hmm... WP:REFPUNCT no longer states anything about punctuation after the reference. I suppose its been deprecated, I'll comment out the message. — Dispenser 19:29, 25 October 2010 (UTC)

Quotes with Redirect checker

Been finding the redirect-only "what links here" tool useful. But I noticed it seems to truncate links that have double quotes in them. Example Gramophone record. Vadmium (talk) 02:02, 30 October 2010 (UTC)

This should be properly fixed now. Thanks you for alerting me to this security venerability. — Dispenser 02:46, 30 October 2010 (UTC)

gc_primary

Hi, i use a script from alexz/Mr Zman (tools:~alexz/coord/) but i've some problems with. I just want to know what's the column "gc_primary", how is this useful? Thanks. Myst (talk) 22:38, 2 November 2010 (UTC)

gc_primary indicates if the coordinate is for the entire article or just a part. Consider an article about a park. The primary coordinate covers the entire park, while the non-primary coordinates are statues in the park.

If you want more help, you should state the problem you're trying to solve. Are you trying make a version Z-man's tool for frwiki or are you trying to create a new tool? — Dispenser 03:31, 3 November 2010 (UTC)

Geodatabase dumps

Hi Dispenser,

I'm trying to use the database dumps with coordinates you provide to make links between Wikipedia articles and points in OSM. It would be useful for me to know when the coordinate from a certain article contained only degrees, or only degrees and minutes (to determine if I should then update the data with the one from OSM). I tried simply converting the decimal values back to dms, but I believe conversion errors are too significant to allow me to make the wanted decision correctly. Is there any way you could provide that information from your parser code, at least for the Romanian wikipedia?

Also, I am reluctant to installing the geo extension for MySQL and I would love to have [6] export it's data in csv format. Also, it would be nice if one could query all the databases at once. Thanks.--Strainu (talk) 12:45, 9 November 2010 (UTC)

List from database

Hi there! You appear to be able to generate lists of various kinds using the Toolserver. Please forgive me if I am making a ridiculous or impossible request, or if you simply do not have the time to do this, but I wonder if you might be able to get me a text file with links to every article in Category:Disambiguation pages that I can use in AWB? There are currently 129,508 pages in the category so I cannot load them all in AWB. Could you please let me know if this is something you can do? Thanks for your consideration of this matter! --Andrew Kelly (talk) 04:48, 24 November 2010 (UTC)

SELECT CONCAT("* [[",page_title,"]]") as Article
FROM page
JOIN categorylinks ON page_id=cl_from
WHERE page_namespace=0 /* article only */
AND cl_to="Disambiguation_pages";

That would be your query, but you'll need to ask at WT:DBR to run it for you. I think you're sticking your head down the wrong hole. Lists like that get dated very quickly and it doesn't have any useful context like when it was last touched, project banners, or intersecting categories. If you're looking for stuff like that, I recommend using CatScan 2. — Dispenser 05:19, 24 November 2010 (UTC)

Cat Scan 2 did the trick nicely. Thanks for the link. If I had known about it I would not have bothered you. Cheers! --Andrew Kelly (talk) 05:45, 24 November 2010 (UTC)

Reflinks blurb

Hi Dispenser, I noticed your request for some writing to help noobs understand reflinks so knocked this up. You may want to add some links to it, but I wasn't sure on the syntax you'd need to use so have left it. (You might want to check it as I've written a few things based on why I think they happen, but it might not be true!) Thanks for the tool, it's so useful!

Hi and welcome to reflinks! This tool will help you to turn bare urls into templated references, hopefully leaving you more time to write and reducing link rot caused by bare urls. The tool visits each webpage that is in a bare reference and collects the page title and some other information automatically, the tool has to be checked over manually to make sure that the references are filled in correctly by the tool. You may have to remove some extra information from the template and add extra details if the tool is not able to find it for you. Some links may be marked as dead links incorrectly as the site blocks the tool, you may wish to manually check these dead links before adding the dead link tag to the article. The tool also does a few other minor maintenance tasks related to references.

SmartSE (talk) 20:27, 24 November 2010 (UTC)

That's prefect. It may even deter the click and forget crowd that my simple non-verbose warning fail to do. — Dispenser 06:04, 25 November 2010 (UTC)

Dabfix fix

Could you please adjust Dabfix so it stops recommending {{tocright}}. The correct template name is {{TOC right}}. Thanks. --ShelfSkewed Talk 21:33, 27 November 2010 (UTC)

That name is the bastard child of Wikipedia (all lowercase, "right") and MediaWiki (ALL_UPPERCASE, "TOC") naming convention. And there is no "correct template name". The naming convention is actually consistent within the tool: all lowercase. — Dispenser 01:31, 28 November 2010 (UTC)

Hmmm. Okay--I guess it has to be lived with. But it is a little annoying having editors "fix" dab pages I watch replacing the direct link to the template with that redirect. There are worse things. Cheers!--ShelfSkewed Talk 04:41, 28 November 2010 (UTC)

(TPS butting in) Except TOC (or even ToC, but that's a bit awkward) is upper case because it's an abbreviation. And systematically replacing a direct link with a redirect is just not very sensible. Can you please change it? Rd232 ^talk 11:41, 30 November 2010 (UTC)

Not all initialisms need to be capitalized, laser, radar, scuba, and sonar are some of the more well known. Our citation templates tend to stick to all lowercase parameter name (including doi, isbn, issn...) which precludes newbie casing mistakes and less is prone to mistyping. Users should not be burdened with memorizing archaic casing. And in our usability studies, new users learned by copying existing wikitext from another pages and modifying it. — Dispenser 05:42, 1 December 2010 (UTC)

Those are acronyms (pronounced as words), and were originally written uppercase but aren't now. TOC is pronounced as an abbreviation not a word, so there's nothing archaic about writing it uppercase. With the existence of redirects, it doesn't matter if users get it wrong (though with the typical copypaste from the correct template transclusion, I don't see what they'll get wrong). If you want to suggest renaming the template, then do that, this is not the way. Rd232 ^talk 11:36, 1 December 2010 (UTC)

Chiming in, it really is irritating to have pages "fixed" to use a template redirect instead of using the correct template name. older ≠ wiser 13:02, 22 May 2011 (UTC)

NYC Meetup: Saturday, December 4

{{Wikipedia:Meetup/NYC/Invite/December 2010}}
This has been an automated delivery by BrownBot (talk) 22:16, 3 December 2010 (UTC)

nl GeoHack

Hello Dispenser, We have translated the template GeoTemplate and we wish the links and view of the nl-version of the GeoHack page on the toolserver is updated with the newest version. We have tried to be carefull as possible but if the page could be tested would be nice as well. How can we arrange that the toolserver page is updated? Greetings - Romaine/Romaine (talk) 02:05, 6 December 2010 (UTC)

It is unfortunate the WMF squid caching servers aren't better integrated as you've done a good job of designing a template. The old cache version should expire after (IIRC) 10% of elapsed time since the last modified date. This sadly means in about a week or two. Writing our own custom caching code is on the to do list and I hope to have it done before the year is out. — Dispenser 06:07, 6 December 2010 (UTC)

Thank you for the reply, I hope it will go faster but we will have patience. Thanks in advance for working on it. Greetings - Romaine (talk) 13:28, 7 December 2010 (UTC)

Dabfix

Hi, I just noticed Dabfix in an edit summary, and followed a link to WT:MOSDAB#New_MoS_bot_like_tool. The changes at Akira look very useful; I'd like to encourage use of the tool by WP:WikiProject Anthroponymy on human name pages. However, I had some difficulty understanding how to install it and use it. I managed to use it in the end but may not have made the best of it.

The "dabfix" link at the above talk page goes to http://toolserver.org/~dispenser/cgi-bin/dabfix.py which says at the top, "Preparing diff. This may take a few minutes. Please review before saving" although nothing is happening until a page is entered -- perhaps this could helpfully be re-worded. I tried clicking on the bookmarklet link before running the tool, but IE7 said "Error on page". I've never used IRC. I didn't know where to find out what dab solver was either.

May I recommend that you prepare step-by-step instructions for editors who are not familiar with such tools?

Also, would you be prepared to tailor it further for use on anthroponymy pages? There is no rule against more than one blue link per line on these pages, and links to obscure sports, bands etc are useful. I therefore suggest that the tool should not remove blue links after the first on a line, if the page has "name" in its title, or is otherwise detected as an anthroponymy page rather than a dab page. - Fayenatic (talk) 14:28, 6 December 2010 (UTC)

Dabfix, again

Firstly can I congratulate you on dabfix. I find it very, very useful especially for finding missing pages. One thing it does though is replace {{TOC right}} with {{tocright}} which is a redirect to the template. Is this deliberate or an oversight? I don't mind but occasionally other editors will revert that bit. Anyway, thanks again. Tassedethe (talk) 05:17, 10 December 2010 (UTC)

Dabfix uses lowercase template names for consistency, see #Dabfix fix above. — Dispenser 13:14, 10 December 2010 (UTC)

Right, didn't see that bit. Thanks for the info. Tassedethe (talk) 14:58, 10 December 2010 (UTC)

The Cleanup Barnstar

		The Cleanup Barnstar
		For all your efforts with Dabfix Tassedethe (talk) 06:15, 14 December 2010 (UTC)

Merge proposal needed

I noticed you'vce added Merge tags to the Air-cushioned landing craft and Landing Craft Air Cushion. You need to complete the process by statring a discussion on one of the talk pages, and giving your rationale for the mnove, and linking to the discussion in both merge headers. If not, I'll probably remove the tags as incomplete. Please note that the articles aren't about the same thing: Landing Craft Air Cushion is about a specific type of hovercraft used by the USN and JMSDF, while Air-cushioned landing craft is about the use of hovercraft as landing craft in general. As such, I'll be opposing the merge if you complete it. (I'll be happy to help you complete the process if you need it.) Thanks. - BilCat (talk) 02:31, 24 December 2010 (UTC)

I'm sorry, I added the tagged after attempting to cleanup LCAC (disambiguation) and being perplexed at what the two pages were trying to cover. As Air-cushioned landing craft claims to be a synonym of LCAC and I thought it was a sub-topic of LCAC. — Dispenser 22:48, 24 December 2010 (UTC)

The original LCAC, Landing Craft Air Cushion, is a specific type, but the term LCAC is also used for any Air-cushioned landing craft. It's no wonder you were confused! No worries. - BilCat (talk) 01:29, 25 December 2010 (UTC)

unlinking common publisher names in cites

Hi. I think one of your scripts unlinks common publisher names (like "CNN") as part of what it does. I think some other scripts or bots out there do this too. I don't object to this, but I would like to know if there a list of common publisher names that you are working off for this, that you can point me to? Then I can know not to bother linking them in the first place, when writing cites. Thanks. Wasted Time R (talk) 16:28, 29 December 2010 (UTC)