Wikipedia talk:WikiProject Microformats/Archive 2

Microformat Icons

Pardon if this is a silly question, but I see the icons for microformats on your example pages (the little green rectangles with a label) and I am wondering if those are supposed to show on pages somehow? Because your Project page is the only place I am seeing them. When {{coord}} gets used on a page, it just expands out into the actual numbers with no icon button in sight. I am wondering if the icon could function as a "hide/show toggle" clickable link such as what we see in navboxes? Then the visual of the little green labelled icon up in the upper right corner of the article would trigger a recognition response in the user and if we wanted to expand that information, one click would toggle the box to "show" and unpack its contents. BeeTea 14:36, 11 July 2007 (UTC)

It could be used like that, if someone were to write the necessary code; but there was a lot of hostility to using the icon at all, even on non- article pages. Andy Mabbett | Talk to Andy Mabbett 09:30, 27 July 2007 (UTC)

Infobox Theatre

I've recently implemented {{Infobox Theatre}}, and was more or less able to code the necessary things into hCard. I was having some problems, however, getting the url to code properly. I'm trying to code it such that all that would need to be entered is the website, that the theatre's name will code into the brackets, and it would display like this: Majestic Theatre. Any help at all would be greatly appreciated. —  MusicMaker5376 02:40, 27 July 2007 (UTC)

The web address seems to be happening already. I've also tweaked the microformat mark-up; see the template's talk page. Thank you for using microformats. Andy Mabbett | Talk to Andy Mabbett 09:28, 27 July 2007 (UTC)

Thanks for your quick attention! However, when I click "Export Contact" in my Firefox, only the name of the theatre and the city show up in Outlook: no street address, no url. Also, I removed the "nickname" tags from the "othernames" field -- it's a field for historical names, not other names in common usage. Thanks again! —  MusicMaker5376 21:29, 27 July 2007 (UTC)

There is currently no way to add an "adr" class around all the sub-fields of the address, in a wiki-table; and tehre is no way, in wiki code, to add class="url" to an anchor element, unless you show the raw URL, wrapped in a span. "nickname" is still the appropriate class to use, for former names; please restore it. Andy Mabbett | Talk to Andy Mabbett 21:38, 27 July 2007 (UTC)

The way I had it before, the street address and the city both appeared in Outlook. And I don't understand why anyone would want to download a name that hasn't been used for 90 years into their address book, but I'll restore it.... —  MusicMaker5376 22:01, 27 July 2007 (UTC)

They don't have to import nicknames - they can omit them at the point of saving. The way you had it before, three *different* addresses were saved, one with just a street address, one with just a city, and one with just a country. Andy Mabbett | Talk to Andy Mabbett 22:10, 27 July 2007 (UTC)

Note how the address is marked up on {{Infobox Education in Canada}}. Andy Mabbett | Talk to Andy Mabbett 11:54, 31 July 2007 (UTC)

Oh, okay. I see what you're saying. If I included the entire address in one field, it would probably work. —  MusicMaker5376 22:06, 31 July 2007 (UTC)

Commons:Geocoding

Maybe there should be crosslinks with Commons:Geocoding. (SEWilco 15:59, 27 July 2007 (UTC))

Time label

Is there a microformat for Time? I'm labeling my images with their location, but haven't noticed a way to label their coordinates in the time dimension. That could be automagically picked up from camera info, although one has to allow manual override (a bot could supply a default time template which could be manually altered). (SEWilco 16:32, 27 July 2007 (UTC))

Not yet, though it has been mentioned previously. The necessary methodology is in place, because there are date/ time components in several microformats already. There are other use-cases, also, such as allowing user-agents to display dates in the user's preferred style, as Wikipedia does, but on any site. there are also issues, such as the problems currently encountered with use of the "abbr" element's title to hold (in this case) the machine readable ISO date format. I'll re-raise the idea on the microformats wiki and mailing list in an hour or two - I'm just about to step out to eat. Andy Mabbett | Talk to Andy Mabbett 17:04, 27 July 2007 (UTC)

On reflection is there any reason why you couldn't use hCalendar? As a minimum, it takes a (start) date-time, and a "summary", which could be the image file name or the location or coordinates; and you can wrap the coordinates in its "location" property. Can you give a URL for one of your pictures, and I'll mark it up as an example. Andy Mabbett | Talk to Andy Mabbett 19:33, 27 July 2007 (UTC)

So far all the images have been in Commons so they're using Commons templates, but at Robert_E._Lee#Monuments the footnote after "attended this dedication" has a cooordinate, while the image of the statue being unveiled has a Commons Location template. I'd like to tag the image with the date of the event (not that anything uses the time yet, although I see Google Earth has added time sequence abilities). (SEWilco 03:05, 28 July 2007 (UTC))

(outdent) It's not as straightforward as I'd imagined, because of the use of templates. What would be needed (on Commons, at least) would be a version of the location template, which would output something like:

<div class="vevent">
  <span class="summary">Unveiling of the Equestrian Statue of Robert E. Lee.</span>
  <span class="dtstart">1890-05-29</span>
  <span class="location geo">37.553861;-77.460194</span>
</div>

Giving:

 Unveiling of the Equestrian Statue of Robert E. Lee.
 1890-05-29
 37.553861;-77.460194

I've raised that on commons, with additional examples and commentary. Andy Mabbett | Talk to Andy Mabbett 08:39, 28 July 2007 (UTC)

hCalendar - first attempts

I've started to apply hCalendar markup to a couple of templates; see Category:Templates generating hCalendars and Apollo 11. The issues I've found so far are to do with:

• end dates are "non-inclusive" which complicates the required mark-up
• no HH:MM:SS component in {{Start date}} or {{End date}} as yet

Because of these, it's probably best if we stick to events with a single, whole-day date, at present.

Any helpful suggestions will be welcome! Andy Mabbett | Talk to Andy Mabbett 18:33, 31 July 2007 (UTC)

Infobox templates

FYI: Category:Infobox templates. Many of these have (and others could have), hCard microformats; some could have hCalendar. Andy Mabbett | Talk to Andy Mabbett 11:34, 1 August 2007 (UTC)

Wikinews

Hi, just an idea: does Wikinews use geotags? I have just seen a website which places news items on a map. It would be interesting if Wikinews would support something like that. --Tgr 13:28, 1 August 2007 (UTC)

I'm not aware of it yet. Check if they have Template:coord or Template:Location defined. I've been geotagging my items in Commons, but don't think Wikinews and search engines are using that info yet. (SEWilco 04:46, 6 August 2007 (UTC))

I've raised the issue of microformats on WikiNews on their "water cooler" (like Village Pump) page. Andy Mabbett | Talk to Andy Mabbett 14:09, 6 August 2007 (UTC)

Logo Deletion Discussion

The Microformat logo images on Wikimedia Commons have been nominated for deletion due to possible copyright violation.

See : http://commons.wikimedia.org/wiki/Commons:Deletion_requests/Microformat_logos#Microformat_logos

--Ozhiker 22:54, 7 August 2007 (UTC)

hCard use on Wikipedia

hCard use on Wikipedia was requested for examination on Template talk:Coord#Untangling. Does the microformat project have any design plans on microformat implementation on Wikipedia, such as which types of templates generate microformat elements, which need the contents wrapped to construct the element, and which need a complete external microformat wrapper through a template or otherwise? If none exist, could the participants of the project please create some? Such a list would be useful to untangle the dependencies of these seemingly ad hoc implementations. Thanks. --Para 16:24, 8 August 2007 (UTC)

GreaseMap

I haven't tried it, but I stumbled across a GreaseMonkey tool GreaseMap. It adds a map which shows coordinates on the currently displayed page. (SEWilco 04:35, 13 August 2007 (UTC))

Coord RfC

Related to the previous request for microformat design plans, there is now a request for comment at Template_talk:Coord#Request for Comment to annotate coordinates on Wikipedia for use as a title in map services and microformat readers. As it's related to this wikiproject, opinions would be appreciated. --Para 10:51, 19 August 2007 (UTC)

The proposal is redundant and broken. Andy Mabbett | Talk to Andy Mabbett 13:52, 19 August 2007 (UTC)

Biographical metadata

The hcard-person functionality should have been implemented in Persondata, not in every infobox that could possibly be included in a biography. What a mess. We don't need redundant metadata guys. Kaldari 22:49, 27 August 2007 (UTC)

Please see Wikipedia talk:Metadata standardization for broader discussion of this problem. Kaldari 23:26, 27 August 2007 (UTC)

You appear to be under misapprehensions about the purpose of microformats, the differences between them and 'personadata' (which clearly could not include microformats' functionality), and the fact that microformats do not produce "redundant metadata". Andy Mabbett | Talk to Andy Mabbett 12:11, 20 August 2008 (UTC)

Template Vgrelease

Shouldn't Vgrelease be implemented instead of Infobox CVG ? [1] « FMF » 16:31, 20 October 2007 (UTC)

Template:Infobox

In case anyone's still monitoring this WikiProject, I figured I should mention that I've just enhanced the {{Infobox}} template to make it easy to add vcard classes to infoboxes generated using it. I think I've covered all the bases, but if anyone has any suggestions I'm keen to hear them. Bryan Derksen (talk) 20:30, 13 March 2008 (UTC)

See User_talk:Omegatron#hCarding_and_vCarding_._._._.3F for a discussion about another infobox — Omegatron 00:32, 4 April 2008 (UTC)

Thank you. Please see a question about class nesting in relation to those changes. Andy Mabbett | Talk to Andy Mabbett 11:09, 24 August 2008 (UTC)

Broken templates

The changes in this edit broke the hCard microformat. The page is protected, so I've requested that somebody reverse them (but not revert; so as to preserve subsequent edits). Other templates may have been similarly broken. Andy Mabbett | Talk to Andy Mabbett 12:05, 20 August 2008 (UTC)

Also [2]. Andy Mabbett | Talk to Andy Mabbett 13:00, 20 August 2008 (UTC)

Both now fixed. Andy Mabbett (User:Pigsonthewing); Talk to Andy Mabbett; Andy Mabbett's contributions 22:51, 6 September 2008 (UTC)

Removal of microformatted data

These edits:

• List of Gaudi Buildings
• Crossings of the River Severn
• Stourport Ring
• List of impact craters on Earth

(and several others made at the same time) removed data (such as label and note properties) from microformats. Note that in some cases, HTML comments warning of the harmfulness of doing so were also removed. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:54, 8 September 2008 (UTC)

New template for in-line display of vernacular & scientific names

I have created {{Biota}} for the display of in-line scientific names, with the same draft "species" microformat used by taxoboxes.

So far, it works only for vernacular, binominal, trinominal and genus names (example on template documentation), but I'll add a few other ranks soon ; and more complex processing for Foo bar ssp. boo and other such formats, later.

It allows wiki-linking, and the optional emboldening of vernacular names, and italicises scientific names automatically.

Comments welcome. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 12:34, 10 September 2008 (UTC)

TfD nomination of Template:Biota

Template:Biota has been nominated for deletion. You are invited to comment on the discussion at the template's entry on the Templates for Deletion page. Thank you. Hesperian 03:38, 12 September 2008 (UTC)

Withdrawn for now, as it has been pointed out to me that Andy is currently not in a position to respond. Hesperian 03:55, 12 September 2008 (UTC)

Andy is back; TfD is live. Hesperian 00:33, 16 September 2008 (UTC)

Parser for Internet Explorer

There is a new parser add-on, Oomph, for Internet Explorer. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:50, 23 October 2008 (UTC)

Look at a template

Hi, I'm not so into this area. Could someone see if the {{DATEtoMOS}} conforms to the microformat syntax? Nsaa (talk) 11:40, 12 November 2008 (UTC)

Currently, it outputs no microformat classes. These might be dtstart (for hCalendar's start date), bday (for hCard) or updated (for hAtom; not currently used on Wikipedia); or dtend (for hCalendar's optional end date). The first three can exist at the same time (and do on {{Start date}}) without harm; but the two hCalendar dates are mutually exclusive. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 12:38, 12 November 2008 (UTC)

Microformat dates- template start date is the wrong approach

I appreciate the good intent of the template writers of {{start date}}- Pigsonthewing-Andy etc., but it needs to be pointed out that the problem of global specification and conversion of dates is not well appreciated.

The first most obvious technical problem we have is that dtend is exclusive, meaning if an event ended on July 4, 1776, then the correct date to insert is dtend 1776-07-05. At first glance, this appears manageable in wikitext, but for those unfamiliar with implementing date arithmetic, simple schemes quickly fall apart upon examination of the details. Note that if the author instead declared July, 1776, the correct date would be 1776-08, which seems simple enough, but consider July 4, 1776 at 23:59, then 1776-07-05T00:00, but if July 4, 1776 at 23:58 then its one minute earlier on the previous day: 1776-07-04T23:59. Did I mention that if this wants to support microformat dates that it is required to support dates in the format 1997-W01-2 (day two of week 1 of 1997), or 1997-W01-2T235959.9942 (time may be expressed in microseconds, or in decimal form). For further details see ISO 8601 since microformat date-times adhere to this spec, and authoritative docs are ISO 8601[3], rfc3339[4]), and for microformat use[5].

Secondly, historians will appreciate the problem of locally specified dates and times. Canonically proper dates and times should use UTC. Naturally, the human readable date may be specified as a local time. For example, consider the dates of Pacific theater battles that occurred on the other side of the date line. Most of these are known for their news reported date in the US and Europe, which is usually one day prior to their local date in the far east.

So for authoring use, we need a human readable date, and the canonically correct microformat date adhering to ISO8601.

After reviewing the details it is not difficult to appreciate that any canonically proper treatment will require a high level language to do the conversion with correctness, and this will be out of the reach of template wikitext for the forseeable future.

Proposal: Authors declare dates in human readable fashion. An attempt is made in templates to manipulate the date, but it will be too precise for approximate dates and may be off for some end dates. A bot periodically takes the human readable date and inserts the canonically proper dtstart and dtend values. To illustrate, consider specifying the date and time of the first shot fired in WWII pacific between the US and the Japanese.

{{start-date| 7 December 1941 8AM HST}} 

{{start-date}} currently does a fairly good job correcting for Hawaii time to UTC and emits for this:

1941-12-07T18:00Z

However, the author was vague on the minutes, so this is overprecise. The bot upgrades the value to:

{{start-date| 7 December 1941 8AM HST|ISO8601=1941-12-07T18Z}}

(Z indicates UTC). This sort of human readable template works just fine because it relies on the strength of the #time: parser function. By contrast, although {{start date}} is not a bad effort, it is seriously challenged in a number of ways. First, the equivalent syntax for dates is difficult to read, and prone to error:

{{Start date|1941|12|7|18|Z|df=yes}}

might seem to be the correct syntax for the above example, but the user must do their own conversion to UTC, and they can't format the way they like to. They also may run into bugs. The above example displays as:

18:0Z, 7 December 1941 (1941-12-07T18:0Z)

Not exactly a pretty sight. The alternative allows the user to format as they wish- with the time preceding, with the date in european order or not, and so on. In cases where authors which to link to month day or year pages, they may specify a second parameter. EG.

{{start-date| December 7, 1941|[[December 7]], [[1941]]}} 

displays as:

December 7, 1941 (1941-12-07) The template may require some refinements, but it is currently demonstrating valid vevent microformat at Siege of Tsingtao, and at Battle of Changde (For verification of correct vevent values, add Firefox's free Operator toolbar, and after setting options to display in data mode, check the debug data). As for the bot runs, running pywikipedia is trivial and there is good python code for doing sophisticated date arirthmetic and utc conversion, so bot operations would not be difficult. I will publish whatever code I have for future bot operators. (Perhaps I can talk some big bot operator to insert it as a periodically run script.) Finally, as part of this proposal, the use of the old {{start date}} and end date would be deprecated.

Note that after the bot runs, users have the opportunity to hand tweak the value parameter to correct for any conversions the Bot may have mistakenly made. A flag for locking the date might be indicated in an embedded comment or a template parameter |blockbot=yes so that future bot passes would not make any further conversions. -J JMesserly (talk) 16:30, 30 January 2009 (UTC)

Your post makes a number of uncited and bogus claims, for example "if the author instead declared July, 1776, the correct date would be 1776-08". There is no requirement "to support dates in the format 1997-W01-2". You don't appear to demonstrate any flaws in the current operation of {{Start date}}, nor, if they do exist, to say why they are not resolvable in that template. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:27, 3 February 2009 (UTC)

Nowhere in the documentation for {{Start date}} is {{Start date|1941|12|7|18|Z|df=yes}} given as a valid input format: please do not use "straw man" arguments. Also, your latter start-date example, With CSS disabled, displays as 1941-12-07T00:00Z7 December 1941 - that's not the "graceful degradation" which we should perform. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:57, 4 February 2009 (UTC)

I support the proposal of J JMesserly and favor the {{start-date}}: Before all, Wikitext must remain human readable. (BTW: There's in fact currently no chance - even for programmers - to enter a date like "7 December 1941 8AM HST" using {{Start date}}: I vainly tried {{Start date|1941|12|7|18|||Z}}, {{Start date|1941|12|7|18||Z}}, {{Start date|1941|12|7|18|Z}}). --Geonick (talk) 00:05, 5 February 2009 (UTC)

On 7 December 1941 Coordinated Universal Time did not exist. It did not begin to exist in any way, shape, or form until the 1950s and did not exist in the form described in ISO 8601 (with leap seconds) until 1 January 1972. ISO 8601 does indeed say that "Z" is the designator for UTC, so it follows that "Z" should not be used for any date in the ISO 8601 format before 1 January 1972. --Gerry Ashton (talk) 18:08, 9 February 2009 (UTC)

(undent) The human readability issues are currently being discussed at Manual of style talk page. I propose we continue to discuss any of the microformats aspects of this proposal here. -J JMesserly (talk) 22:49, 11 February 2009 (UTC)

Para gives great advice to longtime users of Operator

User:Para helped me identify a problem that may have led some to mistakenly assume that some Commons and Wikipedia articles expose invalid microformat data. One possible source of problems is that as of the time of this writing, Operator does not properly uninstall itself. I have used Operator since its early days, but follow the practice as with all beta software to uninstall before reinstalling new revisions. Para suggested I take a look at my prefs.js file. Sure enough, after uninstalling Operator, there were many lines mentioning Operator extension in that file. Deleting them and reinstalling them allowed Operator to correctly identify the microformat information as valid data. Hope this helps to some other long time users of Operator. More recent users likely will likely not run into this particular problem, but it might be good to check for this if apparently good markup is not being recognized by Operator. -J JMesserly (talk) 16:29, 4 February 2009 (UTC)

←I forwarded the URL of the above comment to my good friend Mike Kaply, the author of Operator, who kindly replied:

Most Firefox extensions do not remove their preferences because if they did, then when you upgraded to a new version, you would lose your old prefs. With Operator, I've actually been migrating the preferences over the years so I'm surprised this happened. Without knowing his exact scenario, I can't really debug, but the leaving the prefs behind is working as designed.

Naturally, I shall be happy to forward any further details. As a user and alpha-tester of Operator, I can also confirm that I have seen no such problems. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:27, 5 February 2009 (UTC)

That most certainly will not be necessary. Just my opinion, but cruft from previous alpha versions of software is an unnecessary detail to trouble a developer about. -J JMesserly (talk) 16:40, 6 February 2009 (UTC)

Who said anything about cruft from alpha versions? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:29, 6 February 2009 (UTC)

hCalendar not for biographies

The hCalendar microformat should not be used to represent a person's life-span (particularly, not for biographical infoboxes) because - semantically - a life is not an event. It is a series of many events, starting with a birth-event, and ending with a death-event. The hCard microformat already caters for the semantic markup of a life, as it as a birth-date property; the addition of a death-date property in its parent vCard specification is in hand. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:51, 7 February 2009 (UTC)

It is unhelpful to state controversial positions as if they were fact. Perhaps you are unaware of the controversy [6]? It actually is very old, and there are many that believe hcalendars are perfect for biographies. So please cite this assertion that placing a vevent on biography infoboxes is incorrect. If you cannot, please show what harm is caused if Wikipedia contributors don't do things the way you believe is correct. The benefits are clear. Allowing a vevent on a biography infobox allows the exposure of a death date on a person. If we do it Andy's way we cannot. -J JMesserly (talk) 15:23, 7 February 2009 (UTC)

I am aware of and was involved in that discussion; there is no controversy. Please avoid creating unnecessary drama and avoid making unsubstantiated claims that some anonymous many support your views. As I said above, the issue is already known and steps are in hand to include death-date in vCard, and thus in hCard. I gave multiple justifications - none refuted - in that discussion; one example is that using vEvent does not allow us to mark-up cases where death-date, but not birth-date, is known; death-date on vCard will allow that. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:12, 7 February 2009 (UTC)

Do you deny that many feel that it is perfectly acceptable to encode birth and death dates as dtend and dtstart in a vevent? If so, you would be factually incorrect. There is a controversy in the microformats communitee. Regardless, you have not demonstrated any perceptible harm to wikipedia users. Therefore there is no basis for reverts on this theoretical issue. On the other hand, I have demonstrated benefit. Wikipedia correctly exposed the birth and death dates of Augustus to microformats aware applications. It is factually demonstrable that your revert prevents this functionality. -J JMesserly (talk) 01:10, 8 February 2009 (UTC)

I have demonstrated harm; you have not demonstrated controversy. The metadata in [{Augustus]] was bogus. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:54, 8 February 2009 (UTC)

If not for me, then for the benefit of other readers, please provide a link to where you show concretely what practical harm a wikipedia user could percieve. -J JMesserly (talk) 21:56, 8 February 2009 (UTC)

(undent)Mr. Mabbett, the well meaning founder of this group has once again misrepresented the truth, this time regarding the death date controversy. He writes, "there is no controversy". Mabbett wrote on the project page regarding this issue that "the addition of a death-date property in its parent vCard specification is in hand"[7]. It is simply not so. Tantek Çelik, editor and author of both the hcard and hcalendar specifications states that death date is not necessary [8]. It is simply not credible to state that there is no controversy regarding use of hcalendars for people's birth and death dates or that the specification of an hcard death date property is and hand. The truth is, that Mr. Mabbett as lobbied vigorously for death date in the microformats community and has so far been unsuccessful( see thread responses from Mabbett here [9]). His POV may or may not be correct- that is not the issue. For now, the guidance on the project page should be factual and NPOV. For now, we know that people are using vevent and dtend for expressing death dates, and that there is no known prohibition on using it for this purpose. The Microformats page has be corrected to reflect NPOV on the hcalendar debate. If there is any assertion that one side or the other in the microformats community has been established as being correct, then please provide citations. NOT citations to their arguments, because we are well familiar with them. We need citations of some standards authority saying categorically that it is proper to do it one way and not the other. -J JMesserly (talk) 06:42, 9 February 2009 (UTC)

I have misrepresented nothing, and you have still not demonstrated controversy. Celik is but one voice in the microformat community; and ha been proven to be wrong on a number of significant issues. As to my supposedly being unsuccessful, date-of-death will be added to the next version of the vCard specification; with no dissenting opinion that I have seen. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:56, 9 February 2009 (UTC)

If that is true, then you can provide a cite for your information. -J JMesserly (talk) 20:52, 9 February 2009 (UTC)

AS You have been asked, and failed, to do so for your claim of controversy. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:05, 9 February 2009 (UTC)

I have documented that there are varying points of view on the subject of how to encode death date. Your response is that you dismiss that evidence. Fine. Afford us the same courtesy. Where is your evidence that using vevent is prohibited for death dates? You have repeatedly made the claim. Surely you can make one citation. -J JMesserly (talk) 02:23, 10 February 2009 (UTC)

The case of Augustus should remind us that some dates may be well known in one calendar, for example, the date of Augustus' birth in the Roman calendar, but the conversion to the Proleptic Julian calendar or the Proleptic Gregorian calendar may have an uncertainty of a few days. Casting the date from the contemporaneous calendar to a well-defined microformat may introduce false precision. --Gerry Ashton (talk) 19:41, 9 February 2009 (UTC)

Is there something wrong with expressing this uncertainty by leaving off the day? To illustrate, I have created a new edit to Augustus [10]. Scan to "bday", you will see: {{bday|0063-09|BC=y|September 23, 63 BC (approximate)}} This emits the date as "0063-09". Does that answer this concern? (By the way, I will remove that "approximate" bit unless you have a citation handy. I don't want to get into an imbroglio with roman history scholars.) -J JMesserly (talk) 20:52, 9 February 2009 (UTC)

I see nothing wrong with giving a approximate date in the proleptic Gregorian or Julian calendar, but giving just the month does not work if the date in question is close to a month boundary. Also, the microformat seems to be indicating AD 63 when the correct year is 63 BC.

As for a citation, Holford-Strevens and Blackburn wrote "Republican dates will seldom correspond to the same nominal day in the retrojected Julian calendar" (The Oxford Companion to the Year, Oxford University Press, 2003, p. 670).

Agreed. It would be nice if we had a precision indication as in science, eg plus or minus 5 microns. But I am unaware of any way to do this in ISO8601. That doesn't mean it's not possible, so until someone comes up with something better, we are stuck with backing off to month granularity, when the citations say the exact day is fuzzy. Regarding the AD/BCE, you will notice in the case 2 description that this issue is described. Actually it is emitted properly as a negative date in the emitted source (see how to display source data in the case2 instructions. Unfortunately, Operator does not display BC in the user interface- this is a known bug. The author can be forgiven perhaps because not many people schedule appointments for dates in the BCE period, and such appointment calendar applications are by far the current dominant use of microformat events data at this stage. I forsee a time when apps like google earth will allow us to navigate in the 4th dimension as well, so I agree that we should get as precise as the source material allows with time/date information. I look forward to the time when WP will be the pre-eminent source of accurate semantic information on historical events. Thank you for participating in that. -J JMesserly (talk) 22:31, 9 February 2009 (UTC)

Use of vevents for specifying death dates

Since no evidence from recognized authority has been provided that using vevents for death dates is prohibited, I shall be adding vevents to all Person infoboxes. I would be happy to reverse this if such information is provided and is accurate.-J JMesserly (talk) 02:23, 10 February 2009 (UTC)

Please provide evidence from a "recognized authority" that using vevents (sic) for death dates is supported. As requested on your talk page, please desist from adding vevents (sic) to all Person infoboxes, until you have demonstrated consensus for this change, per WP:BRD. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 12:46, 10 February 2009 (UTC)

You are the one making assertions that it is incorrect. Please provide citations supporting your position, and please document what harm encoding death dates in this way harms wikipedia users in any perceptible way. Otherwise please desist from your reversions. -J JMesserly (talk) 17:13, 10 February 2009 (UTC)

You are the one making assertions that it is correct. Please provide citations supporting your position. You are the one wanting to change the status quo - . Please demonstrate consensus for that change. I have already demonstrated the harm done; and you have cited discussions in which I did so. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:24, 10 February 2009 (UTC)

I am encoding death dates in a matter that is used by other sites. Assuming good faith assumes innocence until proof of guilt. Where is your proof that this practice is "bogus" as you claim? -J JMesserly (talk) 19:05, 10 February 2009 (UTC)

Then you will assume that the method of marking up biographies used on Wikipedia for the past year and more is correct. What other sites are using your method? (Previously-answered questions not answered again). Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:02, 10 February 2009 (UTC)

Here] is an example: "<adactio> "Date of birth and date of death can (and should) be marked up as hCalendar." Here is an example of code emitting dtend as date of death. I have demonstrated practice whereas you have declined to support your allegation that this encoding is bogus. Other people seem to feel you are mistaken, so my point is that if you cannot show any harm from encoding this way, then really this is an academic assertion on a POV you have advocated in the microformats community but for which there is no official position. If this is not the case, then please provide some citations showing why this is "bogus" as you claim, and show how this encoding presents harm to wikipedia. -J JMesserly (talk) 20:26, 10 February 2009 (UTC)

You said "I am encoding death dates in a matter [sic] that is used by other sites". I asked you "What other sites are using your method?" and, earlier, for "evidence from a "recognized authority" that using vevents (sic) for death dates is supported". You have given none, and I have already shown the harm, in discussions you have cited. Other people feel that The Earth is flat, and that Coca Cola tastes good. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:02, 10 February 2009 (UTC)

(undent) User Pigsonthewing has refused to show how encoding death dates as dtends constitutes harm to wikipedia, claiming that he already has. He claims his POV is not controversial- I have demonstrated that it is very much a matter of debate. He claims the practice he claims is "bogus" has no examples in the real world, I have shown examples of this assumption of coding in the real world. I have exhausted the avenue of discussion with Pigsonthewing as he apparently feels it is unnecessary to further discuss an issue he has already proven to me. If that is the case, then I apologize in advance for cluttering this thread. I think what is required at this point is a third party who can make a binding judgment on the matter. The general question of blocking encodings based on unfelt harms to wikipedia users will be dealt with in mediation or arbcom, and I will post the results here. -J JMesserly (talk) 21:25, 10 February 2009 (UTC)

Now you're just lying. I have made no such refusal. The coding example you gave provided scrapes Wikipedia infoboxes and has nothing to do with microformats. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:37, 10 February 2009 (UTC)

                //as ical
                out= new PrintWriter(new FileWriter(new File(this.tmpFolder,"history.ical")));
                out.println("BEGIN:VCALENDAR");
                out.println("CALSCALE:GREGORIAN");
                out.println("METHOD:PUBLISH");
                out.println("X-WR-CALNAME;VALUE=TEXT:History Of Sciences"); 
                out.println("VERSION:2.0");
                for(Person o:this.persons)
                        {
                        for(int side=0;side<2;++side)
                        {
                        Date date=(side==0?o.startDate:o.endDate);
                        if(date==null || date.getMonth()==null || date.getDayOfMonth()==null) continue;
                        String title=(side==0?"Birth":"Death")+" of "+C.escape(o.getName()+" <a href=\"http://www.google.com\">TEST URL</a>");
                        Place place=(side==0?o.startPlace:o.endPlace);
                        out.println("BEGIN:VEVENT");
                        out.println("SUMMARY:"+title);
                        out.println("DTSTART;VALUE=DATE:1900"+(date.getMonth()<10?"0":"")+date.getMonth()+(date.getDayOfMonth()<10?"0":"")+date.getDayOfMonth());
                        out.println("DTEND;VALUE=DATE:1900"+(date.getMonth()<10?"0":"")+date.getMonth()+(date.getDayOfMonth()<10?"0":"")+date.getDayOfMonth());
                        out.println("RRULE:FREQ=YEARLY;WKST=SU");
                        out.println("UID:"+o.getID()+(side==0?"b":"d")+"@freebase.com");
                        out.println("DESCRIPTION:"+title);
                        out.println("LOCATION:"+(place==null?null:C.escape(side==0?o.getBirthPlace():o.getDeathPlace())));
                        if(o.iconFile!=null) out.println("X-GOOGLE-CALENDAR-CONTENT-ICON:"+this.baseURL+o.iconFile.getName());
                        
                        out.println("END:VEVENT");
                        }
                        }
                out.println("END:VCALENDAR");
                out.flush();
                out.close();

-J JMesserly (talk) 22:05, 10 February 2009 (UTC)

As I said - nothing to do with microformats. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:35, 10 February 2009 (UTC)

Denying what your eyes see does not make your denial so. This program is println-ing to file information about Vevents, dtsart, dtend and so how are we to believe your assertion that it has "nothing to do with microformats"? You will have to provide a little more information, because on the face of it, the facts contradict what you assert. This code does indeed have to do with microformats, and it specifically is outputing birth and death dates using these microformat symbols. -J JMesserly (talk) 00:01, 11 February 2009 (UTC)

As I said - nothing to do with microformats. You do know what an iCalendar is, right? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 09:18, 11 February 2009 (UTC)

Nothing is stopping you from spelling out your argument. Go ahead and see if other readers accept your point. -J JMesserly (talk) 00:27, 12 February 2009 (UTC)

You're asking me to prove a negative. Nothing is stopping you from demonstrating that the code processes the hCalendar microformat - except that it does not. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 13:41, 12 February 2009 (UTC)

(undent) The code shows that the author of believed that DTEND and VEVENT may be used to express death dates. (See line 14). Your response is that this code has nothing to do with microformats. I believe those tokens do have a lot to do with microformats, so your response is very difficult to comprehend without further information. You decline to provide it. Readers can make up their own mind about our respective arguments. -J JMesserly (talk) 22:29, 13 February 2009 (UTC)

Are you able to provide any evidence whatsoever that the author was using microformats; or that he was doing so on a publicly-available website? Perhaps you could post the URL? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:19, 19 February 2009 (UTC)

If anyone is baffled by Pigsonthewing's response, here it is. He has been confronted with a program that is using dtends to express death dates. His response is first that it has nothing to do with microformats. Pick up any book on microformats and you will see that iCalendar has a lot to do with microformats. His point attempts to sidestep the fact the code is using vevents and dtends in precisely the way I describe- to describe a life event. Instead we are told it doesn't matter because it is not being used in html on a web page (although we already saw examples from other sources describing precisely this usage. See <adactio> above), or in greater length here. So let's not lose sight of the focus of our inquiry. Is it bogus to use dtends for death dates? Well apparently not for iCalendars. Is this an isolated idea of JMesserly's that has no merit on the web? No. So what's the problem with using this approach on wikipedia? Pigsonthewing doesn't like it and we are told we should do it his way. Fine. He is entitled to do it his way. Others may do it other ways. Let a thousand flowers bloom and let's see how this microformats thing develops. Why dictate to other contributors how they may or may not use microformats if it is perfectly legal, and it presents no perceptible harm? What is the big deal here? We don't know, because Pigsonthewing refuses to enumerate any perceptible harm this causes to wikipedia users. It is simply a POV on an esoteric point of microformat formatting. That's it. -J JMesserly (talk) 04:55, 22 February 2009 (UTC)

Microformat community comments

Due to Mr. Mabbett's frequent errors with his pronouncements regarding microformats, and because his refusal to respond to repeated requests for him to substantiate his assertion that use of vevents with biographies is "bogus"[11], I have taken this question to the Microformats.org community. Use of vevents for biographies is not bogus, improper or illegal, according to one recognized expert at microformats.org [12]. In fact, it is currently the only way to encode death dates at this time. -J JMesserly (talk) 21:48, 19 February 2009 (UTC)

Your allegation of "frequent errors" is false, as was your claim that you are "encoding death dates in a matter [sic] that is used by other sites". What I referred to as "bogus" in the edit you cite was not the method, but the data emitted (you were in error by roughly 33 years). The advice you were given on the microformats mailing list is that "it does seem a little odd… A better way would be to mark up his birth and death as separate events" (which matches the advice you have been given by another "recognized expert" on microformats); so the "only way" to which you refer is not the method you are currently espousing. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:17, 19 February 2009 (UTC)

Corrections:

1. The Microformats discussion responder stated categorically that the use was permitted. He had a suggestion of encoding death as a distinct event and I agree that is a good idea. He never said it was mutually exclusive of encoding an event that encompassed both birth and death events, and had no response to why this might be so for people, but not for other events such as battle of Gettysburg[13]. Even if he has some theoretical response for this inconsistency, that is a style question, not a matter of what is permissible or not permissible. Adactio (cited above) is also a recognized expert who speaks at conferences who holds an opposing view[14]. WP should take no sides in the dispute (see #Proposed live and let live guidance to microformats editors). Both styles should be permitted until good cause is shown to deprecate another.
2. "Frequent errors" by Pigsonthewing is a fair assessment, and I have good cause to be skeptical of any of your claims about microformats. Let's confine ourselves just to the narrow subject of dates since you brought it up. It has been your assertion that {{end-date}} emits bogus metadata[15]. If the user states the completion date is 2009 with only the year specified, you state that the dtend value should be 2009[16]. This is off by a year from the correct 2010 in the cited example. Anyone interested can see confirmation in question one of the inquiry to the microformats.org site[17]. It is elementary microformats behavior about adding one unit for dtend and you were dead wrong. Let's quickly survey three additional serious errors: In Augustus, you reverted as bogus dates: [18]. You did not notice that the Augustus birthdate was correctly encoded as negative. In Lamian war [19], you made the identical error. This information was there in the source tab for you to check such things in Operator, but somehow you missed it, twice, and proclaimed in both cases "bogus metadata". Well, the accusation was nonsense. On Manual of style, Num discussion, you educated the audience on the use of the {{start date}} function. This required me to point out multiple errors in your example of JFK's death {{Start date|1963|11|22|19|00||-07:00}} [20]- that Dallas never has a time correction of -7:00 at any time of year, and secondly that you were using the UTC time (19:00) as a base, which is an elementary error concerning the ISO8601 dates that microformats are based on.

1. If there is any doubt that Pigsonthewing was challenging the legality of using vevents in the way described, we only have to consider the prior threads of comments on this subject[21]. Pigsonthewing claimed he was only obligated to claim the practice was illegal, and it is the contributor's obligation to prove that it is. This is specious. In the real world, we have a presumption of innocence, and citizens are not obligated to prove what they are doing is legal. It is up to accusers to prove the contrary. Regardless, now that Pigsonthewing has been shown to have been in error, 'again on a point of fact regarding microformats, the goal posts have shifted, and we are told that what his claims of bogus metadata did not really have to do with the method, but particular encodings. That's simply not factually accurate[22].

Andy, you have a POV on microformats. That's it. You did some testing on Operator. Great. Thanks. A lot of other people did too. You were first to propose a wikiproject. Great. It means you were first, does it mean anything more than others regarding microformats or that you are any kind of authority on the subject? Not necessarily. My point in all of this is that it is difficult for you to base your arguments on your authority regarding microformats, because it is unconvincing given this track record of frequent errors, and definitive pronouncements that turn out to be flat wrong.

Now, to the matter at hand. You have not shown that the encoding is illegal, and you have not shown that it constitutes any perceptible harm for wikipedia users. Since it does have benefits for wikipedia users, we are left with the question why it is appropriate to continue to block any usage of the usage you do not favor? Because you don't like it? Does consensus mean only things that Pigsonthewing is willing to permit? I don't have any problem with Pigsonthewing encoding biographies lifetime event with an hcard rather than a vevent. That's fine. But how about some reciprocity? Why should contributors be blocked from encoding them in a perfectly legal way? Because you don't like that other way? Where is the harm? This is a repeated question that you have never answered, except with an "I already answered that", or "my responses above refer". You frequently cite WP:BRD. You have Bold and Revert down cold. There is no doubt. I am asking you to Discuss, seek common ground and be open to compromise for the common good. -J JMesserly (talk) 04:14, 22 February 2009 (UTC)

Compound microformats: vcards nested within vevents

In his edit on "Battle of Bouvines"[23], Pigs on the wing mistakenly believes the location Bouvines should not be encoded as in France because it wasn't in france at the time of the battle. The metadata concept he is expressing (in case you are interested) is that data on location within a vevent is a property of the vevent (the battle). Were it not for the fact that the location is itself enclosed in a parent vcard, then he would be correct. However, it isn't and the properties of vcards encapsulated within a vevent are not inherited by the vevent. For further information, please see[24]:

# [16:02:21] <drewinthehead> "so if i have an hCalendar instance with an embedded hCard, which in turn has a tel with a value, does that value become the value of hCalendar's location?"
# [16:02:52] <drewinthehead> "it's a descendant"
# [16:03:31] <drewinthehead> "or does it only apply for a child?"
# [16:03:32] <mkaply> "Not anymore. I just put code in Operator so that microformats will never inherit from another."

Microformats do not inherit from each other, so it is completely irrelevant that during the Battle of Bouvines, Bouvines was not in France. It is proper to indicate that Bouvines currently is in France, because the country name does not have to conform to the date property of the vevent that the vcard is within. The vcard is not constrained by the date 1214. -J JMesserly (talk) 23:56, 10 February 2009 (UTC)

You once again selectively quote; out-of-context (in this case a discussion on inheritance specific to "value excepting"), to support your mistaken view of how microformats work. Please read the hCalendar specification, which is canonical, and states "A named LOCATION (potentially with an address and/or geo) in iCalendar MAY be represented by a nested hCard in hCalendar". Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 09:17, 11 February 2009 (UTC)

Ok. Let's get concrete. Show me where it says that locations may not use present day place names. If you claim a revert based on an assertion that it is "bogus metadata", then you either need to support that assertion with facts, or desist from reverting. The location for Battle of Bouvines [25] is accurate for present day location. If you believe it is not proper, then make your citation supporting your case that the place name must be accurate for the time period. -J JMesserly (talk) 18:26, 11 February 2009 (UTC)

Please - as you asked me to do - read the article. At the time of the battle, an incident which you wish to enclose in an hCalendar microformat as a single event - Bouvines was not in France. Why would you then wish to assert in metadata that it was? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:57, 11 February 2009 (UTC)

Andy, the first sentence of this thread shows that I understand your point. You have not shown why using present day locations is prohibited for specifying locations for events. If you are correct, then I would be happy to observe this, but if you can't, then you have no basis for reverting the article. Why is that not a fair proposition?-J JMesserly (talk) 18:48, 12 February 2009 (UTC)

Feel free to put France as the location in the infobox, in plain view. If there is consensus to do that, then emit it as the lication in metadata. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 00:04, 13 February 2009 (UTC)

Other editors may feel it is redundant given that the text repeats this, but I appreciate your being flexible. -J JMesserly (talk) 01:25, 13 February 2009 (UTC)

Page description regarding vevents for biographies

Regarding this section, I made an edit that made the statements: "This issue continues to be discussed in the microformats community, and methods of recording birth and death dates may change in the future. There is no consensus for prohibiting either method on Wikipedia; many articles use the former approach but cannot encode death date at this time (see talk)."[26]. I gave a citation above with adactio discussing this in the microformats.org site last spring, so statement 1 is accurate. Given the imbroglio over this issue, I believe that is the no concensus statement is also accurate. Yet these statements have been struck[27] and replaced with the following statements replaces them: "This issue has not been discussed for some time by the microformats community, and methods of recording birth and death dates may change in the future. Consensus for the latter view has not been established on Wikipedia; the status quo is the former approach (see talk)." I have striven for NPOV on this issue, and I fail to see how the current text reflects facts. -J JMesserly (talk) 00:29, 11 February 2009 (UTC)

A discussion "last spring" is "not for some time". The current consensus, and status quo, on hundreds of infoboxes, is to use hCard for biographies. You have demonstrated no consensus - and indeed no support - for your one-person campaign to use hCalendar. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 09:22, 11 February 2009 (UTC)

Andy, it is a point of view that is debated in the microformats community, and it is not Wikipedia's place to take sides. I don't really care which way they go on this issue. I simply think we should be emitting death dates, and this is a valid way of doing it. You have asserted it is not, but have declined to make any citation that it is deprecated in any way by any standards body or other authority. The vevent usage is not an invention of my making, and although you have attempted to minimize the importance of the contributions of others that oppose your POV in that community, I have pointed you to its existence and the controversy outside of wikipedia in which you have been an active participant. So let's work this out. In the meantime the text of the project page should reflect these realities. As for concensus, let's both have a little perspective on this. The "microformats community" today is just two guys who seem to have a difficult time working together. Who wants to join a project like that? The "consensus" that created those hundreds of infoboxes is a contribution that a single pioneer deserves credit for. You did it, and my hats off to you. Look at the talk page. During the year of your absence no one showed any interest, and the project was marked as inactive. You are doing good work here, and I admire your energy and tenacity in advancing your arguments. But really, we are just talking at each other here, and I propose we hit the reset button. The fireworks of this interaction will be buried forever in the archives never to be read again. So what do you say? Let's work together and build a less contentious microformats community. -J JMesserly (talk) 18:17, 11 February 2009 (UTC)

I have not said that we should not be emitting death dates. However, if you wish to do that, then the property way is to create an hCalendar microformat with a summary like "Joe Bloggs Died" and a DTSTART property giving the date on which that happened. Joe Bloggs' death is an event; Joe Bloggs is not. You have demonstrated no controversy. I sought and obtained clear consensus before I started to introduce microformats. You have done nothing like that for the changes you have tried to introduce. Your accusations about my behaviour are again fallacious. Nonetheless, I am - and have always been - willing to work with you; and have offered to do so previously, However, "working with you" does not equate to acceding to your apparent demand to be allowed to do damage to Wikipedia unhindered; nor to leaving your ad hominem attacks and false accusations unremarked. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:15, 11 February 2009 (UTC)

{undent) Perhaps you misunderstand. I am not asking you to use vevents in the way that is being done outside of wikipedia. I am asking you to desist preventing me and others from doing so until you support your assertions that this is deprecated in the microformat community and that it presents some perceptible harm to users. That is all. To date, you have done neither and so we can either work this out here, agree to mediation or take it to arbcom. Your choice. -J JMesserly (talk) 19:25, 11 February 2009 (UTC)

I have already responded to this particular threat, elsewhere. You have been asked to provide evidence of sites outside Wikipedia using hCalendar in the way you wish to; and have failed to do so. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:52, 11 February 2009 (UTC)

It's no threat. I outlined our options. As I said, your choice. I think it is needless to repeat my request to you to support your allegations that this is bogus or constitutes harm. I am aware of the options as are you. Why can't we just settle this between ourselves and work together? What is unfair about my request that I and others be allowed to use microformats in ways that are used outside of wikipedia until another party shows that it is prohibited and constitutes harm to users? -J JMesserly (talk) 20:05, 11 February 2009 (UTC)

You are again repeating yourself; see answers above. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:35, 11 February 2009 (UTC)

No evidence being provided, I shall proceed with the addition of vevents to biography articles. -J JMesserly (talk) 18:30, 13 February 2009 (UTC)

User talk:J JMesserly#Death dates refers. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:49, 13 February 2009 (UTC)

(undent) Could you kindly insert whatever citation you have that says that vevents are prohibited for this use? Could you please provide on the Wikiproject Microformats page evidence of any harm that end users can perceive? You have not provided this information anywhere to my knowlege, including the link you just provided, so I shall proceed with the edit. Please do not revert until you are prepared to back up your allegations with proof. -J JMesserly (talk) 22:21, 13 February 2009 (UTC)

User talk:J JMesserly#Death dates refers. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:57, 14 February 2009 (UTC)

I am unaware of anything in the link provided that is responsive to this repeated request for authoritative information supporting your assertion that this encoding is not permitted/deprecated in the microformat community and secondly that it presents some perceptible harm to users. If you could kindly provide that support we could make some progress on finding common ground and resolving this dispute. -J JMesserly (talk) 18:26, 16 February 2009 (UTC)

Proposed live and let live guidance to microformats editors

One of the foundations I think we should observe is a principle that arbcom established in 2005. As paraphrased in Manual of Style:

In June 2005, the Arbitration Committee stated that when either of two styles is acceptable, it is inappropriate for an editor to change an article from one to the other without substantial reason. For example, with respect to British date formats as opposed to American it would be acceptable to change from American format to British if the article concerned a British subject. Edit warring over optional styles (such as 14 February and February 14) is unacceptable.

Similarly, it should be inappropriate for a microformats enthusiast to change an article from one to another form of microformats encoding if both are acceptable. If both can be used, then we should allow both. We should give guidance for editors to observe a live and let live in this emergent stage of microformats. How does that sound? -J JMesserly (talk) 18:17, 11 February 2009 (UTC)

It sounds irrelevant. The wording you cite refers to two styles, of equivalent meaning, and with a large body of support for each. Neither applies here. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:15, 11 February 2009 (UTC)

I think you may misunderstand. I was refering to general guidance for all microformats disputes, not any specific case of one application of microformats versus another. Surely that is relevant. -J JMesserly (talk) 19:23, 11 February 2009 (UTC)

I didn't "misunderstand"; you started a sub-section under "Page description regarding vevents for biographies" " and have since changed the headings to remove it from there, making my comments appear out-of-context. Wikipedia already has dispute resolution procedures in place; starting with WP:BRD, which you have repeatedly shown a disinclination to follow. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:49, 11 February 2009 (UTC)

I can understand your interpretation and have corrected the error in the inadvertent section nesting. Now, to the point at hand. What about the guidance proposal regarding competing valid styles of microformats encoding? Are you proposing that every single dispute be taken to arbcom? Why should we give guidance any different than the principle established by arbcom above? -J JMesserly (talk) 19:58, 11 February 2009 (UTC)

You are so keen to have a "live and let live" approach that you, er, have an outstanding proposal to deprecate the widely-used {{Start date}} and {{End date}} templates. Where have I ever proposed taking any dispute to arbcom? My previous comment still applies. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:06, 11 February 2009 (UTC)

I do have that proposal but it has nothing to do with differing microformat styles that do the same thing. The issue between the {{start date}} and {{start-date}} does not have to do with the microformats they emit- the style of their microformat use is identical- just a dtstart. The issue is their readability as discussed at Manual of style. If the disagreement were solely based on styles of expressing dtstart, then the proposed agnostic stance would apply. So how does the proposal for this general class of dispute sound to you? -J JMesserly (talk) 22:25, 11 February 2009 (UTC)

You are now replacing {{Start date}} with {{Start-date}} (over 250 such edits in the last few hours) citing this talk page as justification, when there is no consensus for you to do so and in contravention your above proposal. Please desist. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:59, 14 February 2009 (UTC)

I am confused about what you are suggesting is improper. There is nothing wrong with the use of {{start-date}}. I don't cite this page as justification, but for background information on the nature of edits that are being made. Regarding this particular issue on use of {{start-date}} versus {{start date}}, to date, in the Manual of style discussions on the subject (also see former section), multiple individuals find {{start date}} as unnecessarily complex. Actually, to date there has been only one person who feels that the other template is preferable. You. Of course it is your prerogative to revert any of these edits based on some argument, but it is unnecessary for contributors to seek consensus for an editorial preferences they have, which appears to be what you are suggesting I require in this case. -J JMesserly (talk) 18:19, 16 February 2009 (UTC)

Previous date template usage has been loose

I am seeing lots of odd usage of our dtstart emitting templates. I'm not sure if these were bot runs or what, but there is a bit of chaos going on with usage. Examples:

• Toastmasters_International was using start date twice. Once for inception, and once for incorporation.
• Multiple songs have multiple start dates eg: Two Tribes Ziggy startdust had 6 dtstarts [28].
• Previous Bot runs probably went amok. accessdate in a cite was set to start-date in M25 Motorway[29] Also see ziggy example above.
  • muliple start dates in the body of text- looks like BOT went bezerk, converting all dates it could find.
    • Kid A
    • Closer (Joy Division album)
    • Duran Duran
• Large numbers of dtsarts:
  • (22) Aladdin Sane
  • (16) Power Pad
  • (12) 38th Canadian Parliament
  • (10) Band of Brothers
  • (10) Ghost in the Machine (album)

If we want to correct multiple dtstart emitting pages, I can probably handle it, depending on what we decide to do about these cases. -J JMesserly (talk) 02:52, 14 February 2009 (UTC)

If you understood the hCalendar specification correctly, as you claim to do, you would know that this matters little; if inside a microformat-emitting template, the microformat will include the first such date encountered, an ignore the rest. Outside of such a template, there is no effect. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:50, 14 February 2009 (UTC)

Perhaps you misunderstood my meaning. I was just listing my findings from a survey of usage, not presenting an argument that the usage breaks anything. Do you believe that there are not some issues to resolve here? At the time of this writing these articles just have a vevent class on the infobox and no other vevents associated with these dstarts. Sure, it matters little from the parser's perspective because it knows what to do. Do the users understand that when they move a dstart emitting template prior to the first in the list that they are changing the the start of the event? Aside from the human intelligibility issue, the issue is, what utility or semantics do the 21 other dtstarts have in the Aladdin Sane article if it is unclear which vevent they correspond to? What name (summary property) is associated with the 2nd, 3rd, 22nd property? Again, the parser has an answer, but is it the one the contributor intended? How do we make this clear for editors? It seems to me that if we are going to be doing collections of vevents in an hcalendar, we are going to provide some support for it, because otherwise lay users are going to perpetuate these patterns of usage which arguably are GIGO. I think we can defer it because I doubt there are many in the world interested in using their online calendar applications for storing this sort of historical information on games and music esoterica. But when microformat aware apps do show up for information of more mainstream interest, they are going to want this data in a little bit more usable form. -J JMesserly (talk) 16:34, 14 February 2009 (UTC)

It was you who chose to use the word "chaos" - more unnecessary drama, perhaps? There are no issues to resolve; there are no no "semantics [for] the 21 other dtstarts" and they have no "summary", because, as I said above, they are not included in any microformat. There is no "garbage out". Any "microformat aware apps" which acted on such dates would be broken. As to what was intended by their use, perhaps you could ask the editor concerned? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:13, 14 February 2009 (UTC)

However the situation is characterized, multiple occurances of this usage pattern indicates that it is a systemic error that users are making. The issue will be academic until users start wondering why the other dates aren't showing up in their applications. -J JMesserly (talk) 17:30, 15 February 2009 (UTC)

Argument that vevent cannot be used if "the subject is not a single event"

There are some who believe in fictional atomic events. This causes them to argue: Template X should not have a vevent because subject is not a single event.

• The first question is, is any event not made up of multiple events?

• Since all events are made up of constituent events, it would follow from the atomic events folks that no pages on wikipedia should use vevents.

This is not the case. Vevents are currently on ((tl|Infobox Military Conflict}}. Presumably, this is ok, because battles are a "single event". But is this so. Let's do a case study on Battle of Gettysburg. Is it a single event?

Consider the battle of Gettysburg. Maybe to you and I, Gettysburg was a famous battle sometime during the civil war and most would put the event somewhere between 1861 to 1865. If we lived in Pennsylvania somewhere near the battle site we might know the dates from our schooling- that the battle was from July 1 to July 3, 1863. Analogously to your "Better" suggestion, should the beginning and end dates be described as separate events? To a historian, and to Wikipedia and commons, even this granularity is rather crude- on the Gettysburg battle we have a large volume of microfomatable information on its constituent events. There are a half dozen articles on the subject. Pickett's charge for example is of historic importance because it foreshadowed the slaughter of World War I due to the dominance of artillery and rapid fire rifles over massed infantry techniques of the Napoleonic era. Some of the events are extremely fine grained, for example the Little_Round_Top article. We even have large numbers of historic photos of particular portions of the event such as a fierce assault at the location known as "Slaughter Pen", that was part of the "Round Top" sub battle.

Is the battle of Gettysburg one event, or a collection of thousands? Can a Byzantine politician's lifetime be regarded as one event, or a collection of thousands? I see no semantic distinction between the two, other than human vanity that our life events are somehow different in kind than other types of events, all of which are also aggregations of constituent events. If a spacecraft's lifetime cannot be considered an event, then why should the span of time describing the aggregate battles that compose the conflict known as the battle of Gettysburg be described as a single event? There is no consistency here.

An hcalendar is a collection of vevents, and its existence is implied on a web page even if it is not explicitly declared. So use of multiple vevents on a page is perfectly legal. Simply because one vevent, eg life of a spacecraft spans others declared on the page is no error. There is nothing improper in the encoding, and there is no basis for removing this functionality from wikipedia pages [30] [31]. -J JMesserly (talk) 22:49, 19 February 2009 (UTC)

Argument that hcards cannot be used with vevents

Nutshell position of J JMesserly: Specifications allow hcards to be nested in vevents. The utility for wikipedia users is that they can see a thousand foot view of the location of an event, whether it was a battle, an aircraft crash site, or a more typical event like a convention that is stored in a calendar application. Adding these tags presents no harm to wikipedia and delivers benefit to its users. On 12:15, 21 February 2009[32], PigsontheWing reverted Battle_of_Haelen with the note "(rv remove bogus hCard microformat (a battle is not a person, organsiation nor venue)". This is an obscure objection that does not affect the formatting or display of the wikipedia article. Pigsonthewing is free to support his allegations of "bogus metadata". In fact, it is true that hcard microformats are permitted to be emitted from events. The venue for this event was indicated in the hcard. Because he has deleted this, it is no longer possible to see the battle site using Yahoo and Google mapping applications from this article. Pigsonthewing has removed functionality with no other apparent benefit other than furthering a theoretical debate regarding microformats.

• Other aspects of this dispute. There is a request that would prevent the use of hcards on {{Infobox Military Conflict}}. The request from Pigsonthewing is to disallow the default behavior of assigning the Place parameter to the microformat class location. The case for allowing the hcard use, and an explanation of why it is necessary to permit the optional disabling of the default behavior was made at the talk page for that template. For further information on this detail of the dispute, please see Infobox Military Conflict talk page.
• Permutation: There is an unsubstatiated allegation that hcard locations should be reflective of the place names at the period of the event, and that such use is therefore "bogus"[33]. The compromise position of adding "present day" shifted the objection to hcard usage[34]. When viewed in non CSS browsers (how many users are we talking about here?) Hcard encoding provides verbosity that is presumably harmful. For discussion see Battle of Bouvine talk page.

1. Is this use bogus from the perspective of the standard? No. The specifications for hcalendar are specific on this point. hCards are allowed within vevents:

   A named LOCATION (potentially with an address and/or geo) in iCalendar MAY be represented by a nested hCard in hCalendar. Similarly, an address LOCATION MAY be represented by an adr, and a geo (latitude and longitude) LOCATION may be represented by a geo. (source)

2. Does the functionality deliver benefit? Yes. It is easy to see the "1000 foot" views of such events from wikipedia by installing the current best of class microformats addons called Operator. This functionality is promised as native capability in all major browsers. Step by step instructions for seeing the map functionality here It's pretty cool.

On the other hand, we are told that a battle site like Battle of Fort Hindman is not a "venue" (arguable), and so an hcard cannot be applied. This is an interpretation/ style POV. This hCard usage is not prohibited by spec, and does deliver benefit to wikipedia users. Should users have interoperability with other net applications be disabled due to such a quibble? I think not. Unless Pigsonthewing can show some harm, his argument has no merit. -J JMesserly (talk) 17:19, 21 February 2009 (UTC)

You are once again presenting straw-man arguments; and once again the quote you cite, from the hCalendar specification is out-of-context and dies not support your arguement in the way you appear to think that it does. hCards are allowed with hCalendars, as the event's location. The specs do not suggest that an event should be marked us as an hCard, and there is no benefit (let alone "interoperability with other net applications") in so doing. You may wish to argue about whether "Battle of Fort Hindman" (note - the battle, not the site of the battle) is or is not a venue; do you wish to do the same for "World War II" or "the Irish Civil War"? You really should desist from your wide-ranging mass changes, until you better understand what you are doing, and how such things should properly be done. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 16:48, 7 March 2009 (UTC)

You are the one reverting edits that use hcards within vevents, claiming they are "Bogus"[35]. There is no support for your position that it is improper to do so. If you wish to make your case, please do so.-J JMesserly (talk) 17:18, 7 March 2009 (UTC)

You are the one applying edits that use hCards as events, claiming they are not "bogus". There is no support for your position that it is proper to do so. You have tried to make a case, and have failed. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:35, 7 March 2009 (UTC)

No, I am simply doing it in a way that you do not approveof. You have not made your case that it is "bogus", and have refused as usual to make any citation to any authority other than your own, which is dubious. -J JMesserly (talk) 17:38, 7 March 2009 (UTC)

No-one can prove a negative; kindly provide proof that your method is valid and semantically meaningful. And please stop making ad-hominem attacks. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:42, 7 March 2009 (UTC)

(undent) Wrong analogy. You are making the argument that you can pronounce someone guilty, and it is their job to prove themselves innocent. It is perfectly possible to point to traffic code that states when a left hand turn is illegal. Ok. You claim my edit is bogus and improper. Ok- show us an authority that states this. You are the one basing your argument on your own personal authority. I simply stated the weakness of that position. -J JMesserly (talk) 17:45, 7 March 2009 (UTC)

I was not making an analogy, and I'm not making accusations of illegal behaviour. You're the one citing specs and claiming, without foundation, that they support your actions. The specs don't explicitly prohibit what you're doing, in the same way that they don't explicitly prohibit you giving every event a description of a value "cheese". They describe a range of permitted behaviours. Your behaviour falls outside that range; and my objections are based on those specs, not "personal authority". Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:28, 7 March 2009 (UTC)

You assert the behavior falls outside the range but decline to make any citations why, claiming that you cannot prove a negative. I'm sorry, but this sort of sophistry is not going to convince even a grade school student of logic. I fail to understand why you think it can convince anyone here. -J JMesserly (talk) 08:33, 3 April 2009 (UTC)

Tech note: Hcards with urls

Emitting Urls creates some verbosity in non CSS browsers. Here is the explanation of the issue. If we do something like <span class="url">[{{{url|{{fullurl:{{FULLPAGENAME}}}}}}} wikipedia article name]</span>

Then the CSS will display the wikipedia article name (usually the fn) as a link, and no additional text need appear from the CSS version. However, the wikimedia engine quite properly adds a rel nofollow to the link due to abuse by spammers trying to boost their search ranks based on inbound links. A microformat parser seeing the nofollow might assume that somehow an address was emitted that the owner doesn't want published, so they decline to retrieve it.

For this reason, {{placename}} is emitting the url only with the linkback prefix: Example emitted text visible on non css browser:

-linkback:http://en.wiki.x.io/wiki/Falklands_War

Perhaps I have erred? Any suggestions for alternate solutions? -J JMesserly (talk) 18:01, 22 February 2009 (UTC)

New templates?

Could new templates {{Sky}}, {{Moon}} and {{Coor Mars}} be considered part of this project? Telescopi (talk) 18:18, 28 March 2009 (UTC)

Because they emit geo? It seems like the geocoding project is the correct home. -J JMesserly (talk) 23:26, 2 April 2009 (UTC)

Claim that +1 year dtend is in contravention of hcalendar spec

This edit [36] makes two changes:

1. unsubstantiated claim that the prior passage is in contravention of the hcalendar spec. The prior passages discussing dtend year date being +1 the user specified year. The correctness of this calculation has never been disproven by Mr. Mabbett, and the correctness was verified at the microformats.org site[[37]. Actually, Pigsonthewing is simply repeating his error. See discussion above in section #Microformat_community_comments. -J JMesserly (talk) 08:16, 3 April 2009 (UTC)
2. the edit removes the statement that {{end date}} does not perform these calculations. It doesn't, so the passage should remain.

-J JMesserly (talk) 08:16, 3 April 2009 (UTC)

The project page states "Dates are adjusted +1 unit of time, where unit of time is dependent on the precision. EG: {{end-date|December 31, 1976}} would generate 1977-01-01Z". If this is so, it is an error, because the template does not indicate the location where the date is observed (it could be anywhere on earth, and so the date actually covers a period of 48 hours) but the microformat output falsely indicates the date is in Universal Time. --Jc3s5h (talk) 22:29, 3 April 2009 (UTC)

Sharp eyes. Thank you very much. That was a typo- no Z is emitted for that example. A quick way of independently verifying what the template is actually doing is to copy the text displayed by the template then paste it into another application like a word processor. The emitted dtend is enclosed in parenthesis. You will see for example, that the Z will not be emitted by that example if you try in in a sandbox. Other examples may be found on {{end-date}}, with two given that specify timezone- one via location and one via standard Timezone abbreviations. A list of recognized abbreviations I think is linked to from the documentation.

It is heartening to receive an inquiry such as yours. I was beginning to think that only Andy and I had any interest in this stuff. If there is anything else that you'd like to chat about regarding microformats, please don't hesitate to contact me. Cheers. -J JMesserly (talk) 01:01, 4 April 2009 (UTC)

Odd behavior of copying in infoboxes

In all infoboxes, it is possible to copy microformatted text and paste the copied text into a wikipedia edit box or a browser address bar to verify the microformatted value in parentheses. Technically, it should not do that and indeed it appears to be a bug, because it doesn't do this when the text not in an infobox. For example:

Date of battle September 7, 2009 (2009-09-07)

Something to be done is to investigate whether we can remove the copy of hidden text from the infoboxes. It would be bad if a dtend that was day adjusted forward were misunderstood.

-J JMesserly (talk) 20:33, 16 April 2009 (UTC)

RDFa

I'm not hugely experience in using microformats, so (while I'm sure there'll be plenty of more experienced individuals that can ease my concerns) after my modest forays into experimenting with their use, alarms bells started going off in my head re: scalability and, more significantly, forward compatibility. RDFa, while a little more complex, seems a more robust alternative to me at first glance, and I didn't have too much difficulty in trying it out in a few XHTML+RDFa pages. Just wondering if there's any wikiprojects concerned with this topic, or whether any members of this one believe it could be encompassed within this project? Is there any scope for its use on Wikipedia? _ɹəəpıɔnı 20:36, 26 April 2009 (UTC)

For me, I don't care which metadata format is used. KML and microformats have tools (eg lots of virtual earth applications, and with microformats, browsers like Firefox have the support native and there are add ons like Operator and IE has similar announced plans as well as their own microformats add-on. These metadata formats deliver some practical benefit to wikipedia users, so WP has some templates that support those. If RDF delivers some practical benefit to users, I am sure the community is open to supporting it as well. At some point they will likely be rationalized, but I see no point in picking winners or losers prematurely. Let them compete for Wikipedia's attention. May the best metadata format win.

Technical detail: How do you propose to encode the following (from a code sample above) on Wikipedia?:

<a href="http://www.neubauten.org/" rel="foaf:interest"...

html <a href is not allowed in wikitext, so the E-R (entity relationship) (aka node-edge) graph is not going to represent much knowledge if you can't represent a relationship between two entities (function of the rel= in the example above). -J JMesserly (talk) 21:13, 26 April 2009 (UTC)

I can see the obvious arguments against RDFa, particularly on Wikipedia where editors don't have direct control of page source code - this I mentioned as it being "a little more complex" to implement and things like your example above would certainly be impossible as far as I can think without MediaWiki level changes. The change of DOCTYPE is another major issue. So yes, RDFa would be far more difficult a challenge to implement throughout the encyclopedia, possibly undoable. But certainly worth discussing nonetheless.

As for "May the best metadata format win", I fear this is rarely the case. Think of the mess of markup this recent "semantic web" drive is attempting to sanitise. In my experience, it is usually the quick and dirty solutions that are quickly poularised, while the well-thought out, standardised ones languish in minority circles until the masses suddenly realise they've made a mess of the web and it needs cleaning. Microformats seems kinda like the "layout tables" of our time to me. Extremely popular (as tables were in the 90s), but unsustainable for the distant future. The similarities are uncanny - tables (like microformats) fit into the existing html technology at the time, while CSS required major changes and adoption of a new technology (as RDFa requires DOCTYPE changes and the adoption of XML namespacing). _ɹəəpıɔnı 02:30, 27 April 2009 (UTC)

Patience. HTML was an abomination to most people familiar with the promise of SGML- to separate rendering information from declaration of what the items in a document were. XML and CSS has brought us closer to that ideal. 18 years ago I would have thought we would have been much further than where we are today, but perhaps Hero of Alexandria is more entitled to be shocked at how sluggish the world can be at recognizing the power of an new idea (in his case steam power). As for the matter at hand, if there are some practical benefits that can be delivered to Wikipedia users by encoding data in RDF, let's hear your proposal. I'd be happy to help try and figure out how it could be folded into the tagging we have if you have a good case. -J JMesserly (talk) 01:14, 28 April 2009 (UTC)

My motivation was to evoke discussion as, after searching a fair bit, I didn't find anything related to this anywhere. I will go and have a closer look into MediaWiki though, see if I come up with anything.

As for the practical benefit, I would've thought it would have all of the same benefits of microformats, without what may (or may not be) future problems. I don't think it's a matter of impatience, rather foresight and sustainability. I could of course be compeletely wrong though, but microformats seem a little too quick and simple to me. I'll go away and come up with concrete proposals now, enough of this talk of hypotheticals. _ɹəəpıɔnı 10:19, 28 April 2009 (UTC)

I suggest incrementalism is a way of going about this. This extension would be useful to both mf and rdf. If you want to propose it's evaluation by the WP tech guys, I will support the proposal. -J JMesserly (talk) 15:52, 28 April 2009 (UTC)

(undent) You may also be interested in the Semantic MediaWiki extensions that eventually will be used on WP. These support full RDF output, however it could take years for their adoption by WP. Here is an example of a Berlin article [38]. Take a look at Query demo[39]. This is a sandbox, so you can try out stuff for yourself. The Semantic mediawiki site has the documentation on this extension. -J JMesserly (talk) 18:36, 30 April 2009 (UTC)

Ha ha. Great minds - after reading your post above regarding the attributes extension, I had gone looking for something more... "complete". SMW was one of the things I found, via this, although as far as I can tell sofar, that project doesn't seem to have gone too far (the current SMW does support RDF output, but no inline machine readable semantics, neither microformats nor RDFa - that I'm aware of).

The one concern I would have with the Link Attributes extension would again be "future-proofing". It assumes that rel, rev and class are and will always be the only attributes ever in use, and provides little scope for extending this. It is also not hugely intuitive, as the prefix symbols for each are just that, symbols. I see no reason not to leave the syntax as ((class="this that")) rather than obfuscating it as ((.this .that)); the former would at least leave some scope for the future addition of the likes of ((property="cc:attributionName" rel="cc:attributionURL" xml:lang="en")).

Anyway, I'm currently looking into SWA, in particular its use in CC, on which that project I linked was based. Thanks for the info, and opinion. I'll keep thinking away on proposals. _ɹəəpıɔnı 02:09, 1 May 2009 (UTC)

Small point on the SMW and microformats- it actually does have some trivial output of microformats. Scan for vcard on the semantic wikimedia site, and you will see how modest it is. This is a highly extensible engine and it would not be difficult to add more support. Take care and stop by from time to time. -J JMesserly (talk) 02:58, 16 May 2009 (UTC)