User talk:Lightmouse/wishlist/Archive 1

Wishlist
Articles like this one: Chronology_of_the_Irish_War_of_Independence. A separate button to bold the dates that occur straight after the asterisk would be great. I'd apply it after delinking. How to delineate the right-hand side, though? Most articles introduce their facts by just day and month, or month and day, rather than the year as well. And there's the issue of the punctuation after the date, which should, IMO, be either a colon and a space, or space – en dash – space. There's MOS support for the latter, and no proscription for the former. The colon would be best included in the bolding, but not the en dash. This would be very good PR for the cause; not only are we removing DA, but we're replacing it with superior formatting. Tony  (talk)  15:19, 1 November 2008 (UTC)


 * Should be possible. Please provide a 'before' and 'after' on this page with nowiki tags. Lightmouse (talk) 00:31, 4 November 2008 (UTC)

Common terms

 * Current geographical list in code:
 * (African?|Alaskan?|American?|Antarctican?|Arctic\sOcean|Asian?|Atlantic\sOcean|Australian?|Austrian?|Beijing|Belgian|Belgium|Berlin|British|British\sIsles|Canada|Canadian|Central\sAmerican?|Chicago|China|Chinese|Korean?|Paris|People's\sRepublic\sof\sChina|Phillipines|Czech\sRepublic|Danish|Denmark|England|English|Europe|European|France|French|Finland|Finnish|Germany?|Great\sBritain|Greece|Greek|Hawaiin?|Hungary|Hungarian|Hong\sKong|Indian?|Indian\sOcean|Ireland|Irish|Italy|Italian|Japan|Japanese|London|Los\sAngeles|Mexican|Mexico|Mexico\sCity|Middle\sEast|Middle\sEastern|Moscow|the\sNetherlands|Netherlands|New\sDelhi|New\sYork|New\sYork\sCity|New\sZealand|New\sZealander|North\sAmerican?|Norway|Norwegian|Pacific|Pacific\sOcean|Poland|Polish|Portugal|Portuguese|Queensland|Rome|Russian?|Scandinavian?|Scotland|Scottish|Shanghai|South\sAfrican?|Arctic\sOcean|South\sAmerican?|South\sKorean?|Spain|Spanish|Sweden|Swedish|Switzerland|Sydney|Tokyo|Toronto|United\sKingdom|U\.K\.|UK|United\sStates|United\sStates\sof\sAmerica|US|U\.S\.|USA|U\.S\.A\.|Vancouver|Wales|Welsh|Western\sAustralian?|West\sGermany?)


 * Key to interpretation of code above:
 * Vertical line '|' is a separator
 * Question mark '?' makes preceding character optional
 * backslash makes following character special e.g. '\s' is a space character, '\.' is a period


 * Names of countries and demonyms


 * Names of languages


 * Professions


 * Current job list in code:
 * (actor|author|dancer|model|novelist|playwright|poet|politican|record\sproducer|singer|songwriter|writer)
 * actress -> actor

Current code
 * .replace(/\[\[(actor|author|dancer|model|novelist|playwright|poet|politican|record producer|singer|songwriter|writer)\]\]/gi, '$1');
 * .replace(/\[\[(act)ress\]\]/gi, '$1or');
 * .replace(/\[\[(dance)\]\](r)/gi, '$1$2');
 * .replace(/\[\[model \(person\)\|(model)\]\]/gi, '$1');

List elements are used to:
 * delink any element within square brackets e.g. Australian
 * delink piped links e.g. British
 * delink languages e.g. Russian


 * U.S. state --> U.S. state
 * state --> U.S. state

I can do the first one but the second one worries me because it is changing what the reader sees and therefore runs the risk of false positives. For example, if the text says ' blah blah U.S. state blah blah. We need to discuss this. Lightmouse (talk) 09:53, 15 October 2008 (UTC)


 * How about yielding just "state"? Do for piping of "State" and "state"? It's surprisingly common. Tony  (talk)  05:06, 22 October 2008 (UTC)

I am not sure what you mean. If you are proposing a change to what the reader sees, can you make this explicit using the nowiki tags please? Lightmouse (talk) 10:59, 22 October 2008 (UTC)

Technical issues identified by editors other than Lightmouse
''This page is not always on my watchlist. So it may be worth telling me of any changes you make.  Please sign your edits''

date sequence
LM, you asked me to add this example of a false positive: See the 3,4,5? . Tony  (talk)  13:42, 28 August 2008 (UTC)


 * Thanks. Note to self: may need to check for digits adjacent to date. Lightmouse (talk) 18:34, 28 August 2008 (UTC)

No comma between day and three-digit year

 * Croatia—in the infobox. Tony   (talk)  13:55, 5 September 2008 (UTC)


 * Should be possible to fix. Will investigate. Lightmouse (talk) 14:10, 5 September 2008 (UTC)

Month capitalization
As part of the date-formatting functions, could lower-case month names be capitalized? --Elliskev 13:54, 5 September 2008 (UTC)


 * Should be possible. Will investigate. Lightmouse (talk) 14:10, 5 September 2008 (UTC)

hard to anticipate this one!
Finder "Mayor" in this. Almost humorous! Tony  (talk)  11:23, 6 September 2008 (UTC)


 * '27th Mayor' -> 'May 27or'. Needs investigating further. Lightmouse (talk) 10:45, 11 September 2008 (UTC)

Incorrectly formatted wikilinked dates
Lightmouse asked me to add this to the wishlist.

The dates here were incorrectly delinked. The dates were incorrectly formatted previously, having more than the required number of brackets. As a Wiki patroller, I see this incorrect formatting quite often and it would be good if Lightbot recognized this situation and either: Thanks. Truthanado (talk) 01:34, 21 September 2008 (UTC)
 * handled it correctly (may not be so easy to do), or
 * raised a flag so that it gets manually addressed by a human.


 * It isn't easy but I am investigating to see if something can be done. Lightmouse (talk) 10:31, 25 September 2008 (UTC)

comma
Interestingly, August 19 31 showed up delinked without an inserted comma (the year is 31). So rare, probably not worthy of your attention. Tony  (talk)  11:06, 22 September 2008 (UTC)


 * Thanks, I will fix the comma issue soon. Lightmouse (talk) 10:28, 25 September 2008 (UTC)

Missed another date
Article: Clint Eastwood My first edit with this script: here My second edit to get a date that was missed by the script: here Dismas |(talk) 08:57, 23 September 2008 (UTC)


 * That is an issue with dates as abbreviations and punctuations 'Feb. 2'
 * They are not common. As you saw, the script does pay some attention to them but not always in ways you expect. I need to investigate further to make it more rational. Lightmouse (talk) 10:27, 25 September 2008 (UTC)

Here the day and month failed to delink. Tony  (talk)  14:22, 31 October 2008 (UTC)


 * Done. The article contained an error i.e. a double space between 'February' and '12'. I have added code to trap such errors. Lightmouse (talk) 17:10, 31 October 2008 (UTC)

dts again
Please see User_talk:Tony1. Tony  (talk)  02:55, 5 November 2008 (UTC)


 * I can't find dts in that article. Where is it? Lightmouse (talk) 10:48, 5 November 2008 (UTC)

Year in X
Lightmouse, are the bottom two buttons removing the baseball Easter egg? Complaint here. I think it may be preferable not to generate complaints about it at this stage. User_talk:Tony1 Tony   (talk)  09:44, 9 November 2008 (UTC)


 * Updated the code. It should no longer do standalone baseball easter eggs. Easter eggs must not be included in autoformatted dates because it breaks autoformatting. The code will still remove easter eggs from (broken) autoformatted dates and treat them just like any other autoformatted date. Lightmouse (talk) 15:24, 9 November 2008 (UTC)

No comma, no space between date and year
When there is a formatted date (Month Date Year format) next to a formatted year with no space and no comma separating the two, the script runs the date and year together. diff Dabomb87 (talk) 20:21, 9 November 2008 (UTC)


 * I have not seen that error before. I wonder how common it is. I have added some code to address it. Thanks Lightmouse (talk) 01:30, 10 November 2008 (UTC)

Commas when converting MDY to DMY
Could you tweak the script so that when the "Delink dates to dmy+common terms" function (on the monobook script) is used, the commas are automatically removed from between the date and the year? In other words:

This is how I have to do it currently: July 9, 2005→("Delink dates to dmy+common terms")→9 July, 2005→("All dates to dmy" to eliminate the comma)→9 July 2005.

This is how I want it: July 9, 2005→("Delink dates to dmy+common terms")→9 July 2005. Dabomb87 (talk) 02:57, 14 November 2008 (UTC)


 * I think I have fixed it. Please update your cache and try again. Lightmouse (talk) 13:37, 14 November 2008 (UTC)
 * Success! Thank you. Dabomb87 (talk) 04:09, 15 November 2008 (UTC)

January 27th, 2008 (no square brackets) failed to be converted to dmy. Ohconfucius (talk) 08:11, 2 December 2008 (UTC)


 * This reminds me that those who have transcluded the script (there must be many by now) need to be advised to refresh their cache regularly to benefit from the updates/fixes. Is fortnightly a good interval? Tony   (talk)  08:48, 2 December 2008 (UTC)

Ohconfucius, it works for me. As Tony suggests, clear your cache and try again. If it doesn't work, cite the article here and I will investigate. Tony, I don't know the answer to your question, I will ask somewhere. Lightmouse (talk) 08:59, 2 December 2008 (UTC)

Delink common terms when using mdy/dmy+common terms function
Sorry I keep having to ply you with technical problems...but they keep occuring. Now, when I use the "Delink dates to mdy/dmy+common terms" function, only the dates are fixed, the common terms are still linked. See this edit, where United States is clearly still linked (using mdy delink+ct) function and this edit, where I used the dmy+ct function but only 1600s was delinked, not the names of continents (the other thing was a manual edit). Dabomb87 (talk) 01:07, 16 November 2008 (UTC)


 * Code updated. Clear your cache and try again.
 * Your feedback is very welcome. Keep it coming. Lightmouse (talk) 12:50, 16 November 2008 (UTC)


 * Thanks. However, a new problem has arisen. I cannot remove the leading 0s using any of the functions here (MDY) and here (DMY).

Code updated. Clear your cache and try again. The reason for these missing features is that I did a big rewrite and the code is much more efficient and easy to maintain now. Unfortunately, some functions were buried in amongst rubbish and got discarded. More feedback welcome. Lightmouse (talk) 16:18, 16 November 2008 (UTC)
 * Perfect. I'll keep the comments coming as I go. Dabomb87 (talk) 16:47, 16 November 2008 (UTC)

For some reason, whenever I click any of the delink dates functions, the script does nothing. Only the common terms, metric units, and year-in-X functions are working. Dabomb87 (talk) 17:41, 16 November 2008 (UTC)


 * Hold on, I will investigate. Lightmouse (talk) 17:44, 16 November 2008 (UTC)

The problem is too big for me to find out what went wrong. I have reverted it to a previous version of the code from earlier today (or maybe even yesterday). You may have lost the fix for the 'leading zero' or the 'ordinal date' feature that was reported today. Let me know. Clear your cache and try again. If those features are still missing, report them again and I will work on them next time I have a chance. Lightmouse (talk) 18:10, 16 November 2008 (UTC)
 * The script is working again. The ordinals are still removed; the leading 0s are not removed anymore. That's OK, the leading 0's aren't a very common error anyway. Dabomb87 (talk) 18:14, 16 November 2008 (UTC)

That is good to know. Thanks for the feedback. As you say, leading zeros are not common but they are annoying. I may get a chance to add the feature at some stage in the future. The important thing for now is that the main features are working again. Give it a week or so and feel free to remind me. Lightmouse (talk) 18:20, 16 November 2008 (UTC)

ISO conversion glitch
http://en.wikipedia.org/w/index.php?title=Ra%C3%BAl_Gonz%C3%A1lez&diff=248002219&oldid=248001700

Tony  (talk)  16:11, 27 October 2008 (UTC)


 * Done. Thanks. Lightmouse (talk) 17:18, 27 October 2008 (UTC)

LM, this one didn't work, although I hit the bottom button (convert to mdy). Tony  (talk)  11:05, 28 October 2008 (UTC)


 * Done. Good find, thanks. Lightmouse (talk) 11:46, 28 October 2008 (UTC)

Second try at those buttons:. It hasn't worked yet! Tony  (talk)  02:35, 31 October 2008 (UTC)


 * Works for me. I added it to my sandbox. Clear your cache and feel free to test it there. Lightmouse (talk) 11:18, 31 October 2008 (UTC)

Your input requested
User_talk:Tony1. Tony  (talk)  10:11, 10 November 2008 (UTC)


 * Done. See: User_talk:Lightmouse. Regards Lightmouse (talk) 10:12, 10 November 2008 (UTC)

script amalgamation
As the results of the RfC discussions become clear, I believe the time may be ripe to ask for the effective amalgamation of the all dates to dmy and dmy scripts for use within AWB. The former script converts all dates to dmy but leaves linked ISO dates and unlinked ordinal dates alone. The latter does not convert unlinked mdy dates into a format (dmy) consistent with the remainder of an article.

The commonest problem I have come across in articles, apart from linked dates, is inconsistent date formatting -where a mix of date formats dmy, mdy, ordinal, dates containing zeros (as in 04 December). It seems that the latter script works better and removes most of these errors as well as strips linked ISO dates, but it fails to render that consistency by ignoring unlinked. Thus, it strikes me that we need one good script for dmy delinking and conversion, and another one which delinks and converts dates to mdy throughout. If one is choosing to convert linked dates to a given format, it should follow that all other dates in a different format should be switched too. The incremental time for the script to run would not be any issue. I don't much care for a script which merely delinks dates without conversion, as this is something which can surely be done by bot, but won't be needed anyway if the above dmy and mdy scripts are available. Ohconfucius (talk) 06:20, 19 December 2008 (UTC)

PS. An editor has asked if it would be possible for '&amp;nbsp;s' to be added to dates so that so date fragments aren't split by linebreaks? Ohconfucius (talk) 06:20, 19 December 2008 (UTC)


 * Good ideas. As I mentioned before, those AWB scripts are not carefully considered function sets, they are just whatever I was thinking at the time I created them. I probably need to reorganise and delete some. I agree with you that the time is ripe for a top-down review. My thinking is that the monobook script is the best precedent for function sets. It has four basic options:
 * Delink dates to dmy
 * Delink dates to mdy
 * All dates to dmy
 * All dates to mdy
 * I will need to keep other options such as 'delink without changing format' for bot purposes. The leading zero function should be in all of them. They can all be made to unlink ISO dates. Unfortunately, including a function to convert ISO dates is not worthwhile because false positives make supervision too stressful. That might be a good task for a separate script.
 * When you mention nbsp, it makes me cringe. I don't want to start a rant about excessive use of nbsp so I will merely say 'I prefer not to write code to add nbsp'. Ideas are always welcome though.
 * I will get round to creating the scripts soon. Lightmouse (talk) 11:24, 19 December 2008 (UTC)

ISO conversion
I know we've been here before, but I feel there is a problem with ISO formatted dates (linked of otherwise), mostly within citation templates but also occur elsewhere which are relatively small in number but still a pain in the butt to convert manually to dmy or mdy. Could you give me back a script solely for that purpose for use in AWB? I promise to be careful when using it :-) Alternatively, could you integrate some lines of code within the current AWB script which convert the ISO dates (within references only)? Ohconfucius (talk) 02:39, 4 January 2009 (UTC)


 * Of the two types of text (reference material, running text), I happen to think ISO dates are better suited to reference material. For example, outside Wikipedia, I have seen ISO dates on international airline tickets. What do you think? Lightmouse (talk) 12:31, 4 January 2009 (UTC)


 * Actually, I don't disagree with you. In the past, I have always used dmy or mdy for the |date= parameter, and ISO for the |accessdate= parameter. However, the issue is with the section of WP:MOSNUM which states that the format should be uniform within the body of the article, and separately within the references section. What I have found is the mixed date formats in the refs section are frequent, but ISO dates are more prevalent here than elsewhere (except for tables). I am at a bit of a loss thinking which way the dates should go. I would be inclined to keep to what I have been accustomed to do (as described above), but that does not sit well with the MOSNUM requirements. Then we had an editor complaining about this edit which Tony executed. This is the manual edit which the complainant saw fit to do to bring about consistency in the formats, hence my request. Ohconfucius (talk) 13:18, 4 January 2009 (UTC)

Yes, perhaps we can bring Tony in on this debate. I suspect he would recommend eliminating ISO which would be a shame. Lightmouse (talk) 14:03, 4 January 2009 (UTC)
 * I find ISO difficult for a number of reasons.
 * First, on a personal level, I haven't exposed myself to it sufficiently to have automatised my processing of the second and third elements.
 * Second, the difference between North American and other numerical date formats makes this worse, and I'm sure not just for me. Which is the month and which the day in 2000-08-03? When it's around the other way (03-08-2008), I see 3 August, but North Americans see March 8. This causes processing problems.
 * Third, no one says it aloud in the ISO order ("2008 August third"), so it's gawky on that count.
 * If these are problems for me, I can't imagine the problems for readers who are less experienced. I hate it. But I've been around long enough to know that some editors really like it and will object strongly to its removal. In the fuss over ending the grip of date-autoformatting, and trying to sort out another system for determining the raw format for articles, I've not wanted to rock the boat on this issue. I do wish that the citation templates allowed readers to choose their format, which would make harmonising the format throughout an article a doddle. Why is it so inflexible? And of course the square brackets around ISO dates just have to go. Tony   (talk)  15:18, 4 January 2009 (UTC)
 * Thanks Tony. I accept that there can never be any ambiguity if the month is spelled out.
 * I also have issues with templates: for example, some citation templates are written to render ISO dates in dmy (as opposed to leaving dates in ISO). Take a look at Project Chanology, which appears to use cite news exclusively. In the refs section, you will see these dates are a mess because some dates were input in mdy and some in ISO. If all dates within references were in ISO and so displayed, then perhaps I would have less of an issue. Unfortunately, things are not so simple: we have articles cited with and without templates. Oh, what to do??? Hence my suggestion to unify all dates to dmy or mdy within the template to the style within the body of the article. Ohconfucius (talk) 16:25, 4 January 2009 (UTC)

I can't remember how I came to know about the ISO format but I probably first saw the format in the context of airline tickets. I am rather amused that the solution to the holy war between US/non-US date formats is to inconvenience both sides in equal measure. I am frustrated by the lack of progress with citations. I think all my scripts remove square brackets around ISO dates. Lightmouse (talk) 16:39, 4 January 2009 (UTC)
 * Your first point is still a common experience. As far as numerical formats are concerned, the slash formats ('11/12/2008' and '12/11/2008') are by far the most common. I am sufficiently exposed to both slash formats that I always have a base level of uncertainty. However, the ISO format '2008-12-11' looks distinct to me, it has the year first and hyphens. Of the three numerical formats, ISO is more comprehensible to me and I don't have the base uncertainty that I do with numerical slash formats.
 * Your second point... Lets distinguish between writer and reader. I have never seen a written format that looks like ISO but doesn't follow ISO ymd sequence. I am sure that such things are rare. So the issue relates to the reader and is probably most important with non-US readers. Fair point.
 * Your third point about reading dates out loud. A fair point, the sequence is not congruent with any spoken order. Of course, we remove other spoken artifacts too ('of' and 'th') although those are less important than sequence.


 * I find the format completely logical and natural that I never had a thought about it, in much the same way as we have had £sd. It's also probably because that's how Chinese dates appear. So what do you propose for now? Ohconfucius (talk) 14:47, 5 January 2009 (UTC)

I must admit that I had suspended thinking about it on the basis that citation wizards would produce a solution "Real soon now". I think we need an update from them. Lightmouse (talk) 16:45, 5 January 2009 (UTC)
 * Where is that discussion taking place? Soon there will be no dates outside templates, and those arguing for DA will get their way. I cringe at the thought. Ohconfucius (talk) 04:24, 6 January 2009 (UTC)

I can't remember. There are several citation templates, and some project pages. It would be good if you could investigate. The discussion was scattered across several. I am afraid I haven't the energy to go back there myself. Lightmouse (talk) 11:57, 6 January 2009 (UTC)
 * I had a look at two template talk pages before I asked you, but couldn't find it. I'll have a wider search when I have a moment. Ohconfucius (talk) 13:25, 6 January 2009 (UTC)

lack of space between links
While we're on the subject, July 31,1986 gets changed to 31 July,1986 and 99 Month9999 gets changed to 99 Month9999 and 4BC gets changed to 4 BC. Add 'em to the list... --Closedmouth (talk) 14:37, 31 December 2008 (UTC)

Thanks. I will take a look and see what I can do. Lightmouse (talk) 14:47, 31 December 2008 (UTC)

Convert all dates to dmy
Would you be able to fork a script from this that converts all dates to dmy, not just delinks them? As you can see here, just converting the linked dates and the templates can cause a fair bit of inconsistency, and since I'm only editing articles about people and things in countries that use dmy, they should be in that format anyway. --Closedmouth (talk) 09:10, 6 November 2008 (UTC)


 * Below 'Delink dates to mdy', you will see 'All dates to dmy'. Try it and let me know if that is what you meant. Lightmouse (talk) 09:46, 7 November 2008 (UTC)

No, sorry, I meant the dmy AWB script (User:Lightmouse/javascript conversion/dmy), not the browser one. Sorry if I didn't make that clear. --Closedmouth (talk) 09:53, 7 November 2008 (UTC)


 * Ah. Yes it is possible. I will need to summon up the energy for coding. Leave it with me. Lightmouse (talk) 10:28, 7 November 2008 (UTC)

First attempt is at: User:Lightmouse/javascript conversion/all dates to dmy. Test it and let me know if it works. Lightmouse (talk) 10:35, 7 November 2008 (UTC)


 * I'm getting this error. I can't figure out where to put the bracket it's asking for; no matter where I put it, the error becomes "too many brackets" :-/ --Closedmouth (talk) 13:10, 8 November 2008 (UTC)

Oh, haha, of course, I figured it out after I made the bug report. Hooray. Although now it's adding spaces before the months regardless, is there a way to fix that? --Closedmouth (talk) 13:18, 8 November 2008 (UTC)


 * Oh my god, I figured that out as well. I am on a fucking roll tonight! Not bad for a complete regex noob. --Closedmouth (talk) 14:01, 8 November 2008 (UTC)

Alright, now it's doing this. That is waaaaay over my head, I think I'll quit while I'm ahead and leave that one to you. --Closedmouth (talk) 14:08, 8 November 2008 (UTC)


 * Sorry for those errors. I have been busy and the code I gave you was untested. I am still busy but give me a day or so, I will do my own testing and produce a better version. I would like this code to work. Regards. Lightmouse (talk) 14:21, 8 November 2008 (UTC)

No rush. --Closedmouth (talk) 01:12, 9 November 2008 (UTC)


 * Try User:Lightmouse/javascript conversion/all dates to dmy now. Let me know how you get on. Lightmouse (talk) 14:51, 9 November 2008 (UTC)

Perfect, thanks for all your work! --Closedmouth (talk) 07:52, 10 November 2008 (UTC)


 * dates with missing spaces are not 'restored' eg May 301888 becomes 30 May1888
 * also, I notice that the script has not been updated since November, so the version I am using in AWB does not incorporated the bug-fixes and improvements since then. Thanks, Ohconfucius (talk) 04:36, 3 December 2008 (UTC)


 * I should be able to fix the '30 May1888' problem. Leave it with me.
 * As you probably know, I have several AWB scripts and the popular monobook script. I don't have a system for bug fixing, it is all rather chaotic. If somebody reports a bug, I just fix it in the offending code. I don't check for the same bug in the other codes. I don't even consider one piece of code as 'master'. Now that this issue is starting to be serious, it is worth considering a more structured approach. It might be possible to deem the monobook code as master (because more users have tested it) and then translate that into AWB. But the translation task is not trivial. Alternatively, I could deem User:Lightmouse/javascript conversion/delink full years (it doesn't reformat to dmy or mdy) as the master. Lightmouse (talk) 12:29, 3 December 2008 (UTC)


 * thanks for your explanation. I had wrongly assumed that one code was master. The reason I requested amendment to the 'all dates to dmy' script is because I am primarily interested in ensuring consistent date formats within articles, and I use AWB to achieve this instead of the browser script. I'll continue watching the changes in the AWB browser until the time you have made the code changes. Keep up the good work. Ohconfucius (talk) 07:29, 4 December 2008 (UTC)


 * script fails to convert '18 January, 1277' to correct format '18 January 1277' Ohconfucius (talk) 09:36, 4 December 2008 (UTC)


 * Probably a result of a deliberate decision not to remove commas in unlinked text. I don't want false removal of commas. Please cite the article and I will take a look. Lightmouse (talk) 09:52, 4 December 2008 (UTC)


 * Same problem with the comma in Stop the War Coalition, which the 'all dates to dmy' do not completely neutralise. It has now been done manually whilst in AWB. Date delinker (aka Ohconfucius)hard problemsproverbs 09:14, 30 January 2009 (UTC)

I am bit confused by that edit. I tried it myself ono a previous version of the article and was confused too. I am worried that it seems to be changing legitimate article titles. Can you give me specific examples of text changes that you think are wrong and what you think it should do. Thanks. Lightmouse (talk) 10:39, 30 January 2009 (UTC)


 * Script failed to detect and correct unlinked dates in ordinal format such as 20th September 2004 (see 65daysofstatic) Ohconfucius (talk) 08:28, 18 December 2008 (UTC)

Works for me. Clear your cache and try again. Lightmouse (talk) 12:09, 18 December 2008 (UTC)

Monobook script failed to detect and correct unlinked dates in ordinal format such as 25th October 1916, 31st October 1916, 29th August 1918, in Jack Cornwell. Ohconfucius (talk) 04:59, 9 February 2009 (UTC)


 * Done. Clear your cache and try again. Thanks for the feedback. Lightmouse (talk) 09:09, 9 February 2009 (UTC)

dates to dmy scripts
So far, very few serious problems have been encountered. Here is a brief list of what I have come across so far, all relate to the 'all dates to dmy' scripts:
 * Dec. 12, 1996 missed in Leslie Cheung


 * Ah. It is not coded to look for periods that signal an abbreviation. Periods also signal an end of sentence. I worry about the false positives if it were to confuse the two. However, I will see if I can do something. Lightmouse (talk) 12:44, 31 December 2008 (UTC)


 * '31 Oct 08' is falsely converted into '31 8 Oct' in Fiona Sit. Likewise, '30 November 04' became '30 4 November' in Mohamed Rahim.


 * It thinks the date is 'Oct 08' as a month-day sequence. Tricky one but needs fixing. I will get my thinking head on. Lightmouse (talk) 12:44, 31 December 2008 (UTC)
 * Try now with the monobook script (clear your cache first) and with User:Lightmouse/AWB/scripts/all dates to dmy. I have not finished work on the other AWB scripts. Lightmouse (talk) 14:47, 31 December 2008 (UTC)


 * '14th November 1963' in Mir Alam Shah went undetected.


 * That was also in the monobook script. Done. Lightmouse (talk) 12:44, 31 December 2008 (UTC)


 * Does not appear to delink dates within tags in 1996 Garley building fire (is this an ISO problem?)


 * Works for me. Please give exact text string that you think it misses. Lightmouse (talk) 12:44, 31 December 2008 (UTC)
 * It picked up some but missed two linked ISO dates at Stanley Internment Camp. Ohconfucius (talk) 03:15, 2 January 2009 (UTC)
 * I have not tested the AWB script, but the monobook script delinks three ISO dates. I couldn't find any other linked ISO dates. So I will count this as a 'works for me'. If it still doesn't work for you, let me know. Lightmouse (talk) 11:04, 2 January 2009 (UTC)


 * Ignores linked ISO dates altogether - it would be good to have the functionality, as exists in the dmy script for general use.


 * Works for me. Please give exact article and text string that you think it misses. Lightmouse (talk) 12:44, 31 December 2008 (UTC)


 * 'May 07, 1997' within tags becomes '07 May 1997' in Hong Kong Special Administrative Region Establishment Day


 * Works for me. Are you sure you are using User:Lightmouse/AWB/scripts/all dates to dmy ? Lightmouse (talk) 12:44, 31 December 2008 (UTC)


 * Yes, it does work now, sorry. However, 'June 01 2003' becomes '01 June 2003' in George Orwell. Ohconfucius (talk) 09:07, 2 January 2009 (UTC)

Cheers, and Happy New Year! Ohconfucius (talk) 08:12, 31 December 2008 (UTC)


 * I have added code to fix '01 June 2003' in George Orwell. The code exists in the master (the monobook script) so clear your cache and try it. But I have not passed it down to the children (the AWB scripts) yet. It will get there when they are next updated. A feature of the master-children strategy is that the workload has increased but I think it is worth it. Lightmouse (talk) 10:55, 2 January 2009 (UTC)


 * Good feedback thanks. Keep it coming. Lightmouse (talk) 12:44, 31 December 2008 (UTC)


 * probably quite rare case of ordinal dates, but 'the 9th of August 1859' was undetected in James Andrew Broun-Ramsay, 1st Marquess of Dalhousie. Ohconfucius (talk) 09:49, 2 January 2009 (UTC)


 * None of the scripts are designed to address dates containing 'of'. I suspect that a lot of quotes will contain such terminology. I think User talk:Rich Farmbrough targets those, he does a lot of date fixing with scripts, you might wish to be aware of his work. If you want code for it, I can create it. Lightmouse (talk) 11:00, 2 January 2009 (UTC)


 * Strange one, this: in Breaker Morant, '2nd July, 11th August, and 7th September' becomes '2 July, 11th August, and 7 September'. Ohconfucius (talk) 03:06, 3 January 2009 (UTC)


 * Yes, that is strange. I made some minor changes but I can't solve it. It is beyond my regex skills right now. Perhaps something will come to me. Lightmouse (talk) 09:20, 3 January 2009 (UTC)
 * I've only seen this once on my travels, so it's hardly a problem important to fix. Ohconfucius (talk) 02:41, 4 January 2009 (UTC)

Thanks. In that case, I won't make it a priority. Lightmouse (talk) 13:07, 4 January 2009 (UTC)


 * '(May 28 1760 – July 23 1794)' becomes '(28 May 1760 – July 23 1794)' in Alexandre, Vicomte de Beauharnais Ohconfucius (talk) 10:00, 8 January 2009 (UTC)

Done. Clear your cache and try again. Thanks. Lightmouse (talk) 12:23, 8 January 2009 (UTC)


 * 'The London telephone area code' becomes 'The London telephone area code' in London. Is this deliberate? Ohconfucius (talk) 15:50, 11 January 2009 (UTC)

Hmm. That is a result of a recent code simplification. There were so many weird and wonderful variants of redirects (e.g. 1 January, last year ) that the code was getting more and more complicated. I simplified it so that it treats all redirects to date articles the same. The code treats '020' exactly the same as a 3 digit year like '412'. I will have to make a change to that. Leave it with me. Thanks for letting me know. Lightmouse (talk) 23:06, 11 January 2009 (UTC)


 * In the corporate restructuring section of the Asus article, 3 January 2008, appears twice, once as body and once as part of a ref. The AWB script renders this: "3 January 2008, ASUS has been in the process of... " Ohconfucius (talk) 03:28, 14 January 2009 (UTC)


 * Yes. The script has a global setting that makes it ignore text in references. That is to avoid false positives. Unfortunately, that is 'miss' related to that feature. I can't think of a way round it unless I break the script into multiple components so you would have perhaps five different scripts to run. It is a legitimate item for a 'wishlist' such as this. One day, I might have a better solution. Lightmouse (talk) 10:36, 30 January 2009 (UTC)