User talk:Guinnog/date linking
From Wikipedia, the free encyclopedia
Contents |
[edit] Ground rules
Beyond obvious things like WP:AGF (and I'd like that to be extended to me as well!), I think it would be helpful to focus on utility rather than precedent or policy. The policy is muddy, and the precedent is unclear. When debating, try to focus exclusively on how the page can best assist the end-user. I know that may be difficult, but I explicitly don't want to go for any sort of straw poll approach here unless we have to. Maybe that way we can turn the unclear policy to our advantage in resolving the dispute, although ultimately it would be nice to end up with a clearer and better policy.
[edit] Proposed framework
If we could generate a list of examples where one should:
- Always link
- Usually link
- Seldom link
- Never link
I think that might be helpful. --Guinnog 04:34, 3 November 2006 (UTC)
All right, I'm going to enact this. --Guinnog 14:11, 3 November 2006 (UTC)
Done! I made a table. See if you can add to it. --Guinnog 14:37, 3 November 2006 (UTC)
[edit] Summary of progress so far
Hmains and Rebecca have been good enough to do their bit with the table. I accept wholeheartedly their reservations about the rather artificial premise of my criteria. It was truly interesting for me to examine my own assumptions about the relative value of linking different dates. The next step I propose is to see how their differing perspectives shape up in editing a test page. I suggest that detailed discussion of the various changes should take place here on this talk page after they have completed the test exercise, and that we leave the main page for any resolutions we may arrive at. --Guinnog 12:41, 5 November 2006 (UTC)
Hmains and Rebecca have now completed this exercise, in the latter case by accepting an edit I did on her behalf according to the principles she laid out in her previous input. The next step will be to analyse the different edits the two users made to see if we can refine the table towards a consensus we can all accept. I will get on with that and will complete within 24 hours. In the meantime, any comments are welcome in the space I made for them below. --Guinnog 02:44, 6 November 2006 (UTC)
[edit] Discussion
[edit] Principles (table)
I suggest that Hmains begin by explaining what his principles are with regard to date delinking. Provision of diffs and examples would obviously facilitate discussion. (Guinnog)
I am afraid my 'principles' are not much more than what you yourself have just written. I too know what guidelines say. While I am doing my copyedit work, I see the linked dates, I look or have looked before at the 'year' articles and do not see that they add any context. I see context in timelines that relate to the subject of the article, but not just links to 'year' articles. These links provide no no discernable benefit to WP readers that I can see so I do my editing just as I do with anything alse that has no beneift. Hmains 05:21, 3 November 2006 (UTC)
- Can you see any merit in retaining any linked dates at all? The ones that come to mind are in articles about history. For example, I retained a link to the 19th century when I copyedited Flashman a while ago. I also often leave dates where they arguably might provide a context. An interesting effect of that is that I almost never let 2006 stand as a linked fragment, but I would often let 1066 stand. In other words, more recent dates seem less useful to me, as the reader (I would assume) is more likely to know enough to place the year in context. --Guinnog 05:29, 3 November 2006 (UTC)
-
- I've added the table above to see if we can end up with some rules of thumb here. Obviously if you are adding to it, it would help if you could give specific examples.
- I don't know what to make of all this. I am not sure that my ideas on year linking are all that important. I am just trying to follow the WP Guidelines as I understand them. I do not want to be implementing something that is my own personal taste or that of some other few editors--that would be a wasted effort. In the Guidelines and prior discussions, I don't see anything special about 'century' linking/delinking or about '1900' or '1800' or whatever as cut off points. Hmains 16:59, 3 November 2006 (UTC)
- I appreciate your comment. It is important to realise that Wikipedia is driven by the needs of its users, not its editors or admins, and that policy and guidelines are useful only inasmuch as they aid this. In a case like this, it is the looseness of the policy we have which seems to have caused so much trouble, and I would hope that any productive conclusions we can come to here may play a part in clarifying policy in the future. --Guinnog 17:20, 3 November 2006 (UTC)
- I did also stress that these were only my guidelines when copyediting. What are yours? --Guinnog 17:22, 3 November 2006 (UTC)
Guinnog, what you put in the table looks fairly reasonable to me. At least, I would not mind if it were followed. Really, I think the whole date-linking issue could be settled by adopting the principle used for links in general: Link only the first occurrence or, for longer articles, perhaps also link a few further occurrences, but no more than are needed to keep one link to the given item in the reader's view. (For whatever difference in makes, there is an actual guideline, somewhere, saying something like that -- for links in general, though not for years in particular.) I suspect that if that principle were followed for bare years then the number of links to them should fall low enough that the links would cease to be an issue for all but a zealous few editors.
I don't quite agree with you about dates in tables, though. Part of the point of tabulating is that a user can find the line (or column, depending) holding the item of interest, and then simply scan across (or down), without needing to refer to rest of the table. If not every occurrence of a year is linked, a user wishing to use a year-link most often must go back to the table at large to find one, partially defeating the point of having the table. Also, links in a table can scarcely be accused of breaking textual flow, given that tables by their very nature break information into separate bits. Lonewolf BC 19:15, 3 November 2006 (UTC)
- Very good points. I'll amend my entry to reflect that. --Guinnog 19:33, 3 November 2006 (UTC)
[edit] Individual points (from example pages)
[edit] Links removed by both
- Repeated links, especially close together in article. NB that medieval is a redirect to Middle Ages, and that eg 1400s is often misused in place of 15th century. Whether linked or unlinked, this is a serious error and should be checked and corrected wherever seen. Note also that forms like "twelfth-century" and 1980's (for the decade) are considered wrong and should be corrected, and that 1990s is not a link to the decade but to the single year.
- Recent links are seen as low-value by all three of us. I thought roughly the last 100 years, Rebecca is more conservative and says the last 20 years. Obviously both of us would use an element of discretion here when delinking.
- Easter-egg links of the type [[1980 in stamp collecting|1980]] are deprecated.
[edit] Links removed by Hmains but retained by Rebecca
(in order of appearance in the article)
- 15th century; some good background about the historical era.
- 6th century BC; much less useful
- 1st century; another fairly poor article. Peripheral interest only.
- 3rd; thin article but contained one interesting fact slightly relevant to article
- 7th; peripheral interest only.
- 240 BC; one interesting fact; but this article is tiny. Little benefit.
- 1st century BC; thin article but contained one interesting fact slightly relevant to article
- 12th century; a slightly better article. One or two interesting facts. Some interest.
*1st century (second instance, but well down a long article from the first)
- second century; another thin article. One slightly relevant fact.
- 4th century; nothing here of any relevance to the article at all.
- 5th century; peripheral interest only.
- 354; article hardly exists. Nothing whatsoever here.
- 430; Small article. Couple of peripherally relevant facts.
- 245; Small article. One peripherally relevant fact.
- 325; article hardly exists. Nothing whatsoever here.
- 315 Little more than a stub. No interest here.
- 386; peripheral interest only.
- 344 Little more than a stub. No interest here.
- 408; peripheral interest only.
- 394; peripheral interest only.
*408 (second instance)
- 547; stub (unmarked). No utility at all.
- 329; stub (unmarked). No utility at all.
- 379 Better article. Still peripheral interest only.
- 9th century Decent article. Good historical context.
- 560; peripheral interest only.
- 636 Stub-class. Nothing, or very little that was pertinent
*12th century (second instance, in image caption)
- 480 Slightly better. One interesting (though still peripheral) link, to Anicius Manlius Severinus Boethius
- 524 Stub. Another mention of Boethius.
*9th century (second instance)
- 672 Little here. We did get the birth of Bede, slightly relevant.
- 735 Sub-stub article. Bede's death.
- 700; peripheral interest only.
- 784 Sub-stub article. Nothing here at all
- thirteenth century Decent article. Some good context.
- eighth century Less good, but some peripheral interest.
- 1550 Some minimal peripheral interest
*13th century (second instance)
- 11th century Decent article. Good historical context.
- 1070 Stub. Little or nothing of relevance here.
- 12th century Decent article. Good historical context.
- 1013 Barely above a stub. No interest.
- 1054 Mention of SN 1054 supernova slightly interesting and of peripheral relevance.
- 1225 Birth of Thomas Aquinas. Very peripheral relevance.
- 1274 Less poor article. Little of relevance though.
*13th century (third instance)
- 1400 Little here of interest. Less poor than some of the earlier year articles. Peripheral interest only.
- 1120 Stub article. One very interesting and relevant fact. Unfortunately Welcher of Malvern is a redlink. Could be improved though.
- 15th century Good historical context.
- 1828 Nothing much here. Peripheral interest for those who like random facts.
- 19th century Very respectable article. Good historical context.
- 16th century Reasonable article. One very relevant reference.
- 1888 Nothing much here. Peripheral interest for those who like random facts.
- 1898 Nothing much here. Peripheral interest for those who like random facts.
- 1830 Nothing of relevance.
- 1896 Some minimal historical context.
- 1816 One interesting fact.
- 1885 Some minimal historical context.
[edit] In chronological order, by type
| Century or year | Grade for article quality (1=high, 5= low) (=x) |
Grade for relevance (=y) |
Grade for utility (=x*y) |
Comment |
|---|---|---|---|---|
| 6th century BC | 3 | 2 | 6 | Decent article, provides some historical context to when Pythagoras lived |
| 1st century BC | ||||
| 1st century | ||||
| 2nd century | ||||
| 3rd century | ||||
| 4th century | ||||
| 5th century | ||||
| 7th century | ||||
| 8th century | ||||
| 9th century | ||||
| 11th century | ||||
| 12th century | ||||
| 13th century | ||||
| 15th century | ||||
| 16th century | ||||
| 19th century | 2 | 3 | 6 | Very interesting article; lots of peripheral historical articles to browse to, if you were finished reading the original article. |
| 240 BC | ||||
| 245 | ||||
| 315 | ||||
| 325 | ||||
| 329 | ||||
| 344 | ||||
| 354 | ||||
| 379 | ||||
| 386 | ||||
| 394 | ||||
| 408 | ||||
| 430 | ||||
| 480 | ||||
| 524 | ||||
| 547 | ||||
| 560 | ||||
| 636 | ||||
| 672 | ||||
| 700 | ||||
| 735 | 5 | 4 | 20 | Year identifies the death of Bede, but article is so poor it provides only two other links, plus the turgid List of state leaders in 735 as context. |
| 784 | ||||
| 1013 | ||||
| 1054 | ||||
| 1070 | ||||
| 1120 | ||||
| 1225 | ||||
| 1274 | ||||
| 1400 | ||||
| 1550 | ||||
| 1816 | ||||
| 1828 | ||||
| 1830 | ||||
| 1885 | ||||
| 1888 | ||||
| 1896 | ||||
| 1898 |
[edit] Analysis (preliminary)
I will have more to say on this. For now, let me explain what I have just done. I tried to put myself into the role of a reasonably intelligent but non-expert person reading the article on Flat Earth (the subject of the test piece). If we can take as read that the repeated, easter egg and very recent year links are regarded by both editors (and by me) as low-value links, it is interesting to click on each of the links that Hmains would have deleted and Rebecca would not, with a view to finding what information if any related to, or even gave meaningful context to, the events described in the article.
Some rather surprising (at least to me) things came up. The century articles seem more valuable for providing historical context than the individual year links for the earlier periods. Many of these early individual year links are so poor in my opinion as not to be worth linking to at all at present. It is also rather hard for me to see how they could ever plausibly be significantly improved. Look at 735 for example.
If you review the links in question, it is hard to avoid the conclusion that many of the year links added nothing at all in terms of direct relevance to the article. I do recognise there are some Wikipedia users who will value the semirandom possibility in, say, clicking on 1885 and thus navigating from Flat Earth to LaMarcus Adna Thompson, the roller-coaster pioneer. I had not been aware though that many of the early year articles are so poor as not to give any real possibility for this activity. I'll think about it some more and try to take this onward this evening. For now, anybody else want to comment? --Guinnog 13:02, 6 November 2006 (UTC)
- Well, I've made a table with the links that Hmains would have deleted and Rebecca would have kept (notionally) in chronological order. For my next trick, I'm going to grade each of the articles (which I've linked from the data table; you were right, Lonewolf BC, thank you). I've chosen to grade each article separately for article quality and relevance, from which I'll make up a utility score by multiplying the two. Like everything else here it might seem arbitrary, but I think it is fair. In the wiki spirit, I'll only do them one or two at a time, and will be happy if anyone else wants to join in. We should, especially if Rebecca and Hmains (and potentially other people) will take part, end up with a very interesting dataset, showing what links are better than others, for this randomly chosen article. I'd still like anyone taking part not to edit the original article though. I'll start by grading a couple tonight, and then leave it to see if anyone else will have a go. Enjoy, and remember we are evaluating the links not on principle, but from the point of view of being valuable to a hypothetical person doing research for a project, say. --Guinnog 23:24, 6 November 2006 (UTC)
[edit] Desired outcome
I thought this might be a good place to consider what we want from this, beyond a resolution of the immediate dispute between these two respected editors. It has become obvious that the existing policy is flawed by its well-meaning ambiguity. I'd like to propose that we aim to come up with a list of examples along the model of the table I made up; something like Always link, Usually link, Seldom link and Never link. I would stress that the criteria I used myself, which largely depended on the period of the date being linked to, were only my own ones, and need not be seen as a model for whatever we will end up with. I hope that the exercise I have asked Hmains and Rebecca to do may allow us to discuss towards establishing such principles. Any comments would be most welcome. --Guinnog 13:25, 5 November 2006 (UTC)
- A resolution of the immediate conflict would be good. An improvement of the guidelines to prevent recurrence and improve clarity would be even better. Support your efforts. ++Lar: t/c 22:38, 6 November 2006 (UTC)
[edit] President Lethe's view
I see three aspects to the question of how to write dates at Wikipedia: personal/national custom, logical writing style (specifically, punctuation), and cross-referencing.
Encoding a date so that it is a link results in three things:
1. The date appears in different ways to different readers; which form of the date appears is the result of user preference (or default).
2. The date is punctuated logically or illogically (and thus, according to my prescriptivism, properly or improperly); this depends on the combination of the original writing and the displayed writing.
3. The date appears in a different color, with underlining, and acts as a cross-reference to another article.
My personal opinion about (1) is that user date preferences should be removed from Wikipedia's programming, and that the Manual of Style should require exactly one style for dates that aren't part of direct, verbatim quotations.
If Wikipedia were to become as I recommend, then points 2 and 3 would be simplified drastically.
If there were only one way of writing a proper date link, and only one way of displaying a date link, this problem would be gone.
(3) would become only a matter of relevant cross-referencing, instead of also a matter of (im)proper punctuation and personal/national custom, if my suggestion were implemented.
My personal preference about when a date should act as a cross-reference (which I see as the worthy point of encoding a date as a link (I think this matter of personal/national custom is unworthy)) is that, if a date appears on the screen, I be able to click on that date to see what else happened on that date in history, regardless of which century it's in. The flow of my reading is not significantly distracted by the different appearance (color, underlining) that cross-references have. If the same date appears five times in a window without any scrolling, not all five occurrences should be cross-references. Also, any cross-reference that seems to lead to an article about a year or date in general (rather than, say, an article about pop music in that year) better lead to what it seems to lead to.
I understand that people will still argue about how relevant a cross-reference is in a certain space. But the argument about links should be only about cross-referencing—not about this waste of time, effort, emotion, and computer resources, not to mention this imperfector of punctuation, that is encoding dates for personal/national preference. My main concern is the combination of (1) not having an imperfect date-rendering program botch the punctuation, (2) not having people fight over personal/national style when all we have to do is say "There is only one preferred style (in terms of punctuation and ordering) of writing dates at Wikipedia (except in verbatim quotes from other sources), and all editors have the right to make dates conform to that style, and no editor should cause a date to stop conforming to that style", and (3) having the link-encoding argument be reduced to the sole question of cross-referencing.
I think this covers my view, in terms of the changes that I advocate, my reasons, and what aspects I see as most worthy of the passion and effort.
President Lethe 03:22, 7 November 2006 (UTC)
