Wikipedia talk:Date formatting and linking poll/Archive 1
Initial discussion
I don't understand why we need another RfC specially on what is called "full date mark up". Is this what is normally known as date autoformatting? People have resoundingly rejected the idea in an RfC already. Does anyone here disagree? If this is referring to DA, I do not agree with the list of disadvantages; nor do I think they're worded in a way that is easily understandable by most WPians.
I though the issues concerned only where year and month-day links might be used. Can we please keep the RfC to what has not already received overwhelming consensus?
The current wording is very confusing to me, and will be the more so for most editors. Tony (talk) 14:15, 5 March 2009 (UTC)
- Well, what do you suggest regarding wording? It seems clear to me. I've no preference regarding full date linking - I suspect it will get overwhelmingly rejected in this poll so we might as well just take it out. Ryan PostlethwaiteSee the mess I've created or let's have banter 14:37, 5 March 2009 (UTC)
- I'd take it out. It would be profoundly irritating to all those who bothered to contribute to the first one to have to do it again. The results will be contaminated because many people would simply not contribute again. Here it is.
- May I suggest that instead we isolate the issues that might be RfCed. If possible, the RfC(s) should be kept as simple as possible, and should be worded so that everyone can understand them. Why is it necessary to have two phases? I'd have thought something like dabomb's list of exceptions might form the basis of a yes/no list of responses (or even Likert-scale choice of 1–5 to indicate approval–disapproval – I'm unsure until it's more concrete). I haven't researched fully how the opposing parties feel, but I notice Arthur Rubin said "DaBomb's new summary of the consensus is reasonably close to my understanding" at the Workshop page. Here is his summary of "When to link". Could this—preferably tweaked and reduced in size, if possible—form the basis of a series of questions? I think people need to discuss this here first. Tony (talk) 15:33, 5 March 2009 (UTC)
- Date auto formatting should be in this RFC as the matter was not settled by the last two RFCs (in fact a majority of Wikipedians seemed to support some form, as evidenced at WP:MOSNUM/RFC (see Question #2)). —Locke Cole • t • c 22:09, 5 March 2009 (UTC)
When will those polled be asked how often they want chronological items to be linked? In phase 2? Dabomb87 (talk) 22:19, 5 March 2009 (UTC)
- Cole, your mantra that "a majority of Wikipedians seemed to support some form, as evidenced at WP:MOSNUM/RFC (see Question #2))" needs to be nipped in the bud, as I've already done at MOSNUM. There, I clearly pointed out the reasons that your RfC funnelled respondents into a "sometimes" category, giving skewed results compared with your No. 1 RfC and the simple one I put up, which asked whether people wanted to keep the current deprecation or go back to "normally" autoformatting dates. Here is the text I put up, again.
That the following text in MOSNUM:
Autoformatting: Dates should not be linked purely for the purpose of autoformatting (even though in the past this was considered desirable).
be changed to:
Autoformatting: Dates (containing either day, month and year, or day and month) should normally be autoformatted.
The result was overwhelmingly in favour of staying put. I've explained how yours funnelled people into a middle category: it's the old extreme book-end technique in questionnaires.
Caught between two choices. Users were given a stark choice. By entering a "Support", they were somehow ensuring that developer time will be used effectively ("To make sure [MediaWiki developers'] time is being used effectively")—it seemed like the easy, positive thing to do. By contrast, declaring Oppose was framed as turning down the opportunity to have developers deliver a date-autoformatting approach that works—and worse, as interrupting professional work towards this goal, something that many folk would think twice about doing ("If not, the developers should be informed of this so they may focus on other aspects of the software that need improving"). I'd feel a heel myself at spoiling their ongoing project. Support responses. A look through them clearly shows that many supporters were influenced by the blue-sky prospect of a new generation of technology, just as in the first RfC; again, many showed a confusion of the issues and technicalities, as would be expected when non-specialists are faced with a complex feature. Conclusion. Again, writing an RfC is an exercise in trying to avoid bias and contamination, a difficult task indeed; the data are only as good as the NPOV of the stimulus. I submit that the language and choices presented to users rendered the result significantly unreliable, and explains why it generated much higher "Support" numbers than the first RfC above or the more straightforward RfC. Regrettably, this RfC does not deliver useful data.
Arthur Rubin does not seem to agree with you on this point, and I suspect he would not go along with yet another RfC on the same issue. (Does he?) Are you the only one, Cole? Tiring out WPians, asking them back and back again to respond to the same issue, will quickly degrade the results (many people will spurn the RfC if we're not careful—perhaps you want that?). Better, as dabomb hints at, to focus on more fine-grained stuff that has not already been decided. Tony (talk) 04:31, 6 March 2009 (UTC)
- You're quite right Tony, as consensus has not been established to remove all auto formatting, we don't need to waste the communities time by asking them again. We can safely continue with developing and implementing a new auto formatting system until such time as you get community consensus to totally remove the system. Thank you for making this simpler. —Locke Cole • t • c 06:15, 6 March 2009 (UTC)
- Stumbling over yourself, there? After resoundingly declaring that it does not want to go back to the old DA system, the community then declared straight after that context that bots do not need prior consensus at MOSNUM to assist articles to comply with a guideline at MOSNUM. This was not quite so resounding as the first RfC (above), but still a large majority. Here is the text:
==Automated and semi-automated compliance==
The use of an automatic or semi-automatic process to bring article text into compliance with any particular guideline in the Manual of Style (dates and numbers) requires separate and prior consensus at [[WT:Manual of Style (dates and numbers)|the talk page]].
Within one or two, by my counting, the results were:
- 79 opposed the requirement of separate and prior consensus for bots to enable compliance (with the deprecation of DA people had just approved of above).
- 22 supported the requirement.
- 2 were neutral.
So the community doesn't want DA, and they strongly support the use of bots to enable its removal; this interpretation appears plain to me. What else could it mean?Asking people again would risk irritating them to the extreme.
This RfC should concern date fragments—when and how often they should be used, as per dabomb's summary, which summarised what came out of Cole's RfCs, (didn't it?)—not date autorformatting. Please let's not mix up the two. Tony (talk) 08:14, 6 March 2009 (UTC)
- It's not clear if this response was meant for me, but if it was, it was irrelevant to what I said. —Locke Cole • t • c 21:32, 7 March 2009 (UTC)
- Who, me? Yes, it was demonstrating why you were indulging in, at worst, a lie—just above—engineered by a skewed and fatally flawed RfC. Please don't keep claiming your cooked results as a true representation. Tony (talk) 02:18, 8 March 2009 (UTC)
- Response to Dabomb87
- Could I suggest that we don't ask the question "how often they want chronological items to be linked"? I think this does not represent the thought processes of an editor deciding on whether to make a link. I mean, editors don't think "What proportion of other date-links already exist? Therefore I must/mustn't make another one." They should be thinking "Is the link I'm considering relevant/useful/helpful/etc. (or whatever criteria is decided)?" So, the answer to the question about "how often" simply gives an editor no real guidance for any particular edit. That was the worst flaw in the recent complex RfC. Please stay with asking the community "Under what circumstances (if any) should [a particular class of date-link] be made?" --RexxS (talk) 00:01, 9 March 2009 (UTC)
Wording and structural issues
Numbering: I believe the points should be numbered, not bulleted, to enable easy referral.
Technicalities need glossing: Points 2–4 in the "advantages" category for the first one are hard for me to understand, let alone WPians who are less acquainted with the issues:
- Clearly indicates which strings are actual dates (as opposed to, e.g., quotations of dates.)
- What is an "e.g." quotation of dates? I have not the least idea without an example or clearer explanation. Italic-close wrongly positioned, BTW.
- Simplifies automated processing of article text (i.e. gathering metadata).
- This is being put up as an advantage of linking, but no one has ever provided an example of how it could be so. If the point is to be made, can we have a short gloss for "metadata" and an example, if possible in just a few words? I'm foggy on what an advantage would even look like.
- Populates "what links here" pages with possibly relevant data.
- This is for editors of year and day-month pages, is it? I think that should be stated, if so.
Skew accepted but needs to be noted: I note that "advantages" are given first position over "disadvantages" in every case. While either had to come first, consistently, this should be regarded as an advantage in itself in further negotiations over the text and structure. Tony (talk) 04:31, 6 March 2009 (UTC)
.
Use of linking
Can anyone provide me with a list of specific examples of when to link that needs clarifying? I've got a couple so far on the main draft page, but I could do with help getting a full list. Ryan PostlethwaiteSee the mess I've created or let's have banter 19:44, 7 March 2009 (UTC)
- Not trying to toot my own horn, but User:Dabomb87/Summary of the Date Linking RFCs#When to link is pretty detailed in that regard. I will provide additional examples later. Dabomb87 (talk) 22:20, 7 March 2009 (UTC)
Suggested autoformatting question
Shall the wish of some developers to preserve the present autoformatting markup (such as the square brackets in "[[March 9]], [[2008]]") be overruled and the automated removal of any such markup that is present only for autoformatting purposes begin? The stated reason for preserving such markup is the possible future availability of better autoformatting software.
--Gerry Ashton (talk) 20:58, 7 March 2009 (UTC)
- A major statement on the notion of involving a proposed "Son of Autoformatting" is here. Tony (talk) 15:03, 9 March 2009 (UTC)
Points removed from year markup disadvantages
- The search box is readily available as an alternative to low-relevance links, which can also be highlighted and/or piped by "gateway" links in the "See also" section to avoid cluttering the text.
This isn't a disadvantage of year markup, it's a discussion of the workaround to issues caused by overlinking - so I've removed it from the list. It's also poorly-written, colliding separate points in one sentence. -- Earle Martin [t/c] 21:40, 10 March 2009 (UTC)
- Complicates the syntax for editing pages.
This poorly-worded assertion (what "syntax" is made more complicated? Are you trying to say that the text editing area is cluttered by extra brackets? If so, say it) is entirely unproven. Also, as Ckatz comments, so does the syntax for linking, bold and italic formatting, templates, headers, and every other bit of Wiki code. -- Earle Martin [t/c] 03:40, 11 March 2009 (UTC)
- right: every bit of markup complicates editing - but this RfC is not about those other types of markup. the question of whether date-linking/markup is worth the added complication is important to some of us. please discuss the wording instead of deleting the point, okay? thanks Sssoul (talk) 08:09, 14 March 2009 (UTC)
- Indeed, Sssoul has said it well. Tony (talk) 08:36, 14 March 2009 (UTC)
- No, the onus is on you to explain how linking to a year is somehow of such fiendish complexity that it merits being claimed as a "disadvantage". -- Earle Martin [t/c] 09:11, 15 March 2009 (UTC)
- Indeed, Sssoul has said it well. Tony (talk) 08:36, 14 March 2009 (UTC)
Clerk
If that isn't a statement by Ryan (it looks similar, but not identical, to some statements made by Ryan as a proposed neutral statement), it needs to be clearly marked proposed wording. — Arthur Rubin (talk) 01:47, 14 March 2009 (UTC)
rationale please
the proposed wording includes the statement "For instance, specific day-month articles may be linked in articles about holidays that fall on the same day every year, e.g. Christmas Day, April Fools' Day, and Cinco de Mayo." could someone please explain what the perceived value of links to 25 December, 1 April and 5 May is in those articles, or provide a link to where the rationale is given? if the main aim is to visually highlight the date in such articles, that should be specified as one of the perceived advantages of linking. Sssoul (talk) 08:01, 14 March 2009 (UTC)
Two queries
First:
"Theoretically, other forms of date markup could be adopted that would not create links to articles – for example {{24 March}} or {{datetemplate|24 March}} – but would serve to designate the dates as dates."
What on earth does it mean? Designate for the readers, or for the collection of metadata? Why would one want to use a date markup that did not link to date articles? If I am confused, I can't imagine what visitors to this RfC would make of it.
Can someone please fix up this text so the meaning is clear, especially for non-experts? Otherwise, I'm going to remove it in a few days' time as a dysfunctional part of the RfC.
- sorry if it's unclear. my point is that the previous version wrongly stated that "date markup means linking to articles". "linking" and "markup" are not synonyms and it's misleading to phrase this as if they were synonyms. i trust we can make the wording clearer without making it sound like date markup = linking.
- a couple of reasons someone might "want to use a date markup that did not link to date articles" are mentioned prominently in this RfC: to enable autoformatting or to facilitate the gathering of metadata. neither of those functions require linking, and the same examples of theoretically-possible nonlinking markup are mentioned in the section on autoformatting below. Sssoul (talk) 10:59, 14 March 2009 (UTC)
Second:
Less seriously, but still an issue, is this:
"Clearly indicates which strings are actual dates (as opposed to quotations of dates.)"
"Actual" is not a good word, since dates cited in the main text are every bit as real (or not real) as dates within quotes. I can't work out what point is being made here. Dates in the main text turn blue, setting them off from the surrounding text, but dates within quotations are smoothly integrated into the quotation text? If this is it, can the advantage be made more explicit? Otherwise, readers will glaze over and skip on without understanding your point. Tony (talk) 09:01, 14 March 2009 (UTC)
- i don't understand this either. Sssoul (talk) 10:59, 14 March 2009 (UTC)
PS Earle Martin: yes, you're right about removing the hyphen in "date fragments". Tony (talk) 14:35, 14 March 2009 (UTC)
"What links here" pollution
One of the "Advantages" of month-day linking is: Populates "what links here" pages with possibly relevant data. I went to the March 14 page and clicked on What Links Here. I was able to click "next 500" 18 times, over 9000 pages link to March 14. (Filtering to Articles only reduced this to 13 "next 500" clicks.) The links are listed in the order they were created. So if a reader clicks "what links here", that can peruse 9000 random order pages that may or may not have anything enlightening. How is this more useful than pressing "Random article"?
The same goes for year linking. I selected 1931, the year in which Hope Lange was not born. The 1931 What Links Here shows about 6000 random order pages. I can't see where this type of linking finds any relevant data.-- SWTPC6800 (talk) 19:47, 14 March 2009 (UTC)