User talk:Kendrick7/Summary of the Date Linking RFCs

Latest comment: 15 years ago by Arthur Rubin in topic Comments

Raw numbers

edit

It's not that I don't trust Dabomb87, but you may wish to double check the raw counts. Further, if we really want to get down to what the community really thinks, it may be worth our while to read each !votes comment and consider setting aside comments which are misinformed (those suggesting a technical solution is impossible, for example). Another thing to consider is "what is consensus"? For example, someone running for adminship needs 75-80% support to pass. The first question in the 2nd RFC, re: deprecating dates linked purely for auto formatting, is barely over 80%, with 14% opposed (and that's without hand pruning !votes that are based on incorrect assumptions). Just some things to consider that I think might shed more light on these results than simply counting votes. —Locke Coletc 03:19, 13 February 2009 (UTC)Reply

I'm way too lazy for all that. If you want to double check his results, feel free to have at it. I have no objection to tweaking about the edges here. -- Kendrick7talk 03:38, 13 February 2009 (UTC)Reply
Please feel free to do it yourself Locke, I fully understand your desire for accuracy. Dabomb87 (talk) 01:11, 14 February 2009 (UTC)Reply
I noted that my margin of error on the larger counts is ±3, I think that would count it. Dabomb87 (talk) 01:13, 14 February 2009 (UTC)Reply

Sorted results of RFC2Q1

edit

I've broken down the !votes at RFC2Q1 at User:Locke Cole/Date Linking RFC Analysis, here's a table with the results.

!Votes Percentage Category
1 0.33% Editor supported deprecation but wrote nonsense
16 5.25% Editor supported deprecation but seemed to support fixing autoformatting
158 51.80% Editor supported deprecation but seemed unaware autoformatting could be fixed
6 1.97% Editor supported deprecation but only via a vote
1 0.33% Editor supported deprecation but provided a confusing reason
2 0.66% Editor supported deprecation but did not sign
1 0.33% Anonymous editor supported
62 20.33% Editor supported deprecation
51 16.72% Oppose deprecation
7 2.30% Neutral about deprecation

Comments? —Locke Coletc 06:14, 14 February 2009 (UTC)Reply

Edit your results in sirrah. 'Tis your playground. -- Kendrick7talk 06:19, 14 February 2009 (UTC)Reply
For ease of understanding, possibly reduce the number of categories by merging (nonsense and confusing?; chuck the stragglers into one category?), and present a more logical order. Does the "Editor supported dep." category mean that they wrote just that without comment? Possibly put that cat. first? Tony (talk) 06:58, 14 February 2009 (UTC)Reply

Proposal on year linking

edit

This is copied directly from where I proposed it on the Greg L RFC.

Any year may be linked to, once (in accordance with our guidelines preventing overlinking), in an article. The determination of whether any particular year should be linked is an issue of editors' judgment, as with most other links in the encyclopedia, and should be debated on article talk pages if necessary. Wikipedia readers should be treated as adults who are capable of deciding for themselves how they wish to browse the articles of this project - purposefully, casually, or randomly, as they see fit.

Hex (❝?!❞) 13:24, 13 February 2009 (UTC)Reply


Well, it's only a guideline. -- Kendrick7talk 19:05, 13 February 2009 (UTC)Reply
Here's a rephrase with more guidelininess and less blunt editorializing, which isn't really necessary outside the context of heavy MOSNUM discussion.
Years may be linked to in articles, but usually only once, in accordance with our guidelines preventing overlinking. The determination of whether a particular year is worth linking to from an article is an issue of editorial judgment, as with all other article links. If necessary, the utility of a year link should be debated on an article's talk page.
Hex (❝?!❞) 19:37, 13 February 2009 (UTC)Reply


Clarification/more info?

edit

Proposal 2: "While some are opposed to the autoformatting mechanism itself, some are only against the current autoformatting method of wikilinking full dates. Supporters of autoformatting believe that it is necessary to prevent date format wars, to suit user preferences and for consistency. Opposers believe that autoformatting is unnecessary because the differences are trivial, the current method of autoformatting is harmful, and that it is too complex for little gain."

Please consider specifying the numbers (where possible), where you say things like "supporters of DA believe ...". And they didn't all specify all three reasons, did they? That's what your statement appears to say. More importantly, when characterising the vast majority who "opposed", can you give a sense of what "some" means in terms of the numbers or proportion. There are two "some" statements at the top. Would be clearer if the second and third sentences were swapped in order. Tony (talk) 04:14, 14 February 2009 (UTC)Reply

Since this is an exact copy of DaBomb's findings, you might want to mention this to him also. -- Kendrick7talk 04:18, 14 February 2009 (UTC)Reply
Well, I did have problems with his summary. Vagueness was one issue. I've looked only at the second of the "three" RfC analyses properly here. Tony (talk) 06:56, 14 February 2009 (UTC)Reply
Mine is quite vague; I intended to expand it some but other Wiki things caught my eye and real life became busier. Sometime in March (and after the Rfarb finishes), I may pick the RfC apart—response by response. Dabomb87 (talk) 15:24, 14 February 2009 (UTC)Reply
Nevertheless, it's dabomb's work that demonstrated just how complex these RfCs are to interpret—particularly the "detailed" ones. WP would benefit from a set of guidelines on how to construct and interpret RfCs, since this is the kind of exercise that typically lies at the centre of the reliability of PhD projects in the social sciences. The most obvious structural issue to come out of this exercise is the need to avoid "funnelling". The concept is by no means my invention, although the term is in this context. Tony (talk) 15:29, 14 February 2009 (UTC)Reply
Right, I mean that's part of the reason I didn't hunker down myself and write an RfC back last fall on this. It's not hard to end up basically push polling when trying to simplify a complex issue for a random audience of editors. The manner in which I would have tried to simplify it would have naturally reflected my own point of view, and if I had just presented it in its whole complexity I don't think I would have gotten particularly high quality answers. -- Kendrick7talk 19:40, 14 February 2009 (UTC)Reply

Comments

edit

If it comes to that, I don't like the last three bullet points under when years should be linked. I think this might lead to overlinking. But, even the first few, if accepted, make delinking bots impossible. — Arthur Rubin (talk) 22:21, 14 February 2009 (UTC)Reply

Well, I tend to hold to the original meaning of "overlinking" meaning linking from one article to another article multiple times, and linking out common words. I don't think I'm promoting that. -- Kendrick7talk 22:58, 14 February 2009 (UTC)Reply

(outdent) "In this sense, we might as well link every noun and adjective, because they all have some tangential relationship to the subject." No, that's not an accurate comparison at all. Links to those would only reveal information about the English language.

Context: if you reach 1170 from Strongbow, you find out that while the Normans were invading Dublin, Madoc of Wales, just across the Irish Sea, was allegedly discovering America. That's local context. Expand to regional context and you get the Danes attacking Estonia. Expand to global context and you find Fes becoming the biggest city in the world. Other, more detailed year articles (1345, anyone?) will tell you about everything that was going on everywhere.

Events are islands. History is a sea. Let our readers sail freely. — Hex (❝?!❞) 19:35, 15 February 2009 (UTC)Reply

Yes, and we have a search box for that. Links are supposed help readers understand the topic at hand, not to act as a discretionary browsing tool. All those factoids are interesting. They also have nothing to do with cheddar cheese or its history. Dabomb87 (talk) 19:41, 15 February 2009 (UTC)Reply
I think we have found the problem here - your extraordinarily blinkered understanding of hypertext. — Hex (❝?!❞) 21:44, 15 February 2009 (UTC)Reply
And what is your un-blinkered understanding of links? Dabomb87 (talk) 22:23, 15 February 2009 (UTC)Reply
That they are a tremendous tool for the unfettered expansion of knowledge, encouraging free exploration of the realm of information, and serving as a giant source of serendipity. They make Wikipedia the free school and Montessori of encyclopedias. To willingly strip articles of some of their potential for learning is butchery. — Hex (❝?!❞) 23:49, 15 February 2009 (UTC)Reply
Before anyone suggests it, I am not suggesting linking everything on the page, nor even every occurrence of a term on the page, or even necessarily every potential linkable term - I've said time and time again that presence of links should be down to editorial judgment on a case-by-case basis. What I am arguing against is the blanket prohibition of a whole class of link on the mistaken belief that they are not "useful". — Hex (❝?!❞) 00:01, 16 February 2009 (UTC)Reply
Well, as an editor, may I suggest that this specific link in this specific article (1170 in cheddar cheese) is not useful? Dabomb87 (talk) 00:04, 16 February 2009 (UTC)Reply
Except to someone who reads the article and is curious about the year 1170. God, it's not rocket science, is it? — Hex (❝?!❞) 12:37, 16 February 2009 (UTC)Reply
And the chances of that are? Dabomb87 (talk) 13:09, 16 February 2009 (UTC)Reply
Perhaps you'd like to suggest a scientific method of calculating those chances. While you think about that, let me note that I often follow year links. It would be an act of supreme arrogance to imagine that I am somehow unique amongst all our millions of readers. — Hex (❝?!❞) 13:20, 16 February 2009 (UTC)Reply
I've put some comments below. I do wish those on the other side would realize that simply unlinking chronological items (years in particular here, but dates as well) isn't the only option. —Locke Coletc 14:15, 16 February 2009 (UTC)Reply
  • You say I should read the article on 1170 ... which bit in particular is relevant to cheddar cheese? Run it past us, finally. Tony (talk) 03:00, 16 February 2009 (UTC) "if you reach 1170 from Strongbow, you find out that while the Normans were invading Dublin, Madoc of Wales, just across the Irish Sea, was allegedly discovering America. That's local context. Expand to regional context and you get the Danes attacking Estonia. Expand to global context and you find Fes becoming the biggest city in the world. Other, more detailed year articles (1345, anyone?) will tell you about everything that was going on everywhere." ... But we don't want to ruin our wonderful wikilinking system with these distant generalities. If they are even vaguely relevant to the topic, put them in the article. They are not relevant to understanding Strongbow or cheese. Nor is the fart of a French peasant at 11.02 am on a certain day in 1170. You draw the circular boundary so widely is dilutes the power of the linking system. Serious readers soon tire of it when they meet false negatives put up as relevant by WP. Yet there are more closely relevant articles: link them alone. Tony (talk) 03:05, 16 February 2009 (UTC)Reply
  • Sorry, Tony. Some of us are "big picture" people and perhaps you are not. Perhaps you do not share the love of history which has drawn many editors to the ideals of this project. I, for one, read my Funk and Wagnalls from A to Z when I was a kid, and not necessarily in that order, so don't you dare lecture me about what "serious readers" want from this encyclopedia. -- Kendrick7talk 03:22, 16 February 2009 (UTC)Reply
Ooh, anything you disagree with is now framed as a "lecture", is it? You degrade your argument with such spin. Don't apologise sarcastically. I'm big picture too and love history, but you don't seem to get the point. And if you like discretionary browsing, fine, but key the information into the search box rather than diluting the high-value, more specific links. As I said, what has it got to do with cheese? Tony (talk) 03:29, 16 February 2009 (UTC)Reply
Those last two sentences from Tony are basically the core of the argument here, I have yet to see a true argument as to how a link to 1170 improves the reader's understanding of the history of cheddar cheese. Dabomb87 (talk) 03:31, 16 February 2009 (UTC)Reply
Did someone move you cheese? This is a hyperlinked project, quit with the Whambulance -- Kendrick7talk 03:38, 16 February 2009 (UTC)Reply
It is important to remember that this is a hyperlink project, so really, please consider that there should not be any hindrance to clicking on anything that might one day be relevant to even one reader out there.  HWV258  04:54, 16 February 2009 (UTC)Reply
That particular straw man is also a dead horse. Please don't insult the other participants in this discussion by using it again. — Hex (❝?!❞) 12:37, 16 February 2009 (UTC)Reply
On the contrary, my point goes to the very heart of the issue. Who is to say which links are important? If it's "no one", then there will be never-ending debate and misery (as has currently been seen with the date-delinking fiasco).  HWV258  21:49, 16 February 2009 (UTC)Reply

← Further comments:

  • RfC 1 #1: please strike "and at that, very rarely," and the final sentence "There will be very few cases when a date link is relevant to the context." It doesn't seem supported by the actual responses.
  • When to links / Year links: I still think many of those are inappropriate for linking; the question for each of them should be brought separately in the new RfC.
There is, at least, some support for each of the bullet points, so they should go to the new RfC. — Arthur Rubin (talk) 00:02, 20 February 2009 (UTC)Reply

Other options

edit

This has been discussed before, but why don't we simply continue to have years/dates linked, but change the color of those links? They would still underline when hovered over, but they wouldn't "devalue" other context links (due to their different color). UC Bill has developed code which addresses a number of major concerns of those against date/year links, I'm certain it wouldn't be hard for him to add code to emit something similar to <a class="chronologicalLink" href="http://en.wikipedia.org/wiki/January_1">January 1</a>, and then we could override the color and text style of date links for all readers (at common.css or monobook.css), as well as on a editor by editor basis on our userspace monobook.css. This seems more reasonable than crippling the site for some at the insistence of others. —Locke Coletc 14:15, 16 February 2009 (UTC)Reply