User:Shereth/MOSNUM

RFC on deprecated date-linking for auto-formatting purposes

The linking of dates for the purposes of auto-formatting has been deprecated and is no longer in practice; see MOS:SYL for the pertinent section in the Manual of Style. This RFC is intended to clarify what impact this has on the use of linked dates and years in Wikipedia articles.

In order to facilitate future discussion pertaining to linked dates and their use in Wikipedia, it has been requested that the community consider the following points of concern.

Should linked dates (for auto-formatting purposes) be systematically removed?

By deprecating the linking of dates for the purpose of auto-formatting, the practice of adding this kind of formatting to articles has been eliminated project-wide. However, there is some question as to whether this discontinuance should also warrant systematic removal of existing linked dates, or whether existing linked dates should be grandfathered in and any removal thereof left to the discretion of article maintainers.

For the purposes of this discussion, dates linked for the purposes of auto-formatting are dates that appear in any of the following formats, as per Wikipedia:Manual of Style (dates and numbers)/Date autoformatting:

• [[May 15]]
• [[15 May]]
• [[May 15]], [[2005]]
• [[15 May]] [[2005]]
• [[2005-05-15]]

As such, it is proposed that MOS:SYL be modified to the following:

Proposed Wording

• Autoformatting: Dates should not be linked purely for the purpose of autoformatting (even though in the past this was considered desirable). Existing links for the purpose of autoformatting dates should be removed when encountered.

Discussion

Please discuss your support or opposition to the above proposed wording here.

• Should arguments agains this proposal that rely on future improvements to the autoformatting program carry any weight, since those participating in the discussion have no power to make such changes? --Gerry Ashton (talk) 18:26, 13 November 2008 (UTC)
  • I don't think you can predetermine the validity (or lack thereof) of arguments prior to the filing of an RFC. Determination on whether or not the arguments carry any weight should be made upon review when closing a discussion. Shereth 18:30, 13 November 2008 (UTC)
    • No, but we're intent on discussing how the guideline should work based on currently available methods and technology. To do so based on theoretical technological enhancements would fall into the realm of WP:CRYSTAL. Ohconfucius (talk) 02:16, 14 November 2008 (UTC)
• "Should be removed" is perhaps unnecessarily strong; it's not like editors are under any obilgation to do this - the point is that they can be removed, and other editors should not then recreate them.--Kotniski (talk) 18:50, 13 November 2008 (UTC)
  • That may be a point of discussion at RFC. Some editor (perhaps even I) will assuredly contend that human judgement to relink specific cases should trump scripted bot delinking.LeadSongDog (talk) 21:24, 13 November 2008 (UTC)
• The most immediate question which springs to mind is: how is one to know if the date linking in any given article is "for the purpose of autoformatting"? Ohconfucius (talk) 02:10, 14 November 2008 (UTC)
• This wording seems unclear. There is no autoformatting going on with [[2005]]; it appears that way for everyone. Also, [[May 5|May 5]] is not autoformatting but it is linking. I don’t understand what the precise intended effect of this new wording was supposed to address. That a bot will not be involved? Greg L (talk) 05:39, 14 November 2008 (UTC)

1. I second Kotniski's concern about the apparent compulsion in "should be removed"; "may be removed" is the appropriate wording.
2. I second Ohconfucius's point about the critical ambiguity between a date that is linked (i.e., to connect to the date fragment page or pages) and a date that is "autoformatted". Confusing linking and autoformatting is a fatal flaw in the wording, and arises because both use the same syntax (which in itself was an appalling misjudgement by the programmers in the first place). Technically, it's impossible to link a date to its component day-month and year pages without autoformatting it. The confusion and ambiguity occurs twice in the wording. Many people will not understand it. Date autoformatting needs to be treated by itself, first. In any case, the notion of adding square brackets to dates for any purpose other than autoformatting them is completely undermined by this statement, in the explanation: "For the purposes of this discussion, dates linked for the purposes of auto-formatting are dates that appear in any of the following formats,...".
3. Where is the text of the reference ([1])?
4. What does "when encountered" add, apart from fuzziness? Who would remove date autoformatting that they hadn't encountered? These two words should be removed.
5. "there is some question" is ungrammatical.
6. "Existing links for the purpose of autoformatting dates should be removed when encountered."—Why in italics? But more importantly, if people approved of this wording, it would mean that the same wording would have to be enacted for every single point in every style page; this would blow out the size, somewhat, and make the style guides cluttered and tiresome to read. Permission does not have to be explicitly granted to users to change text so that it complies with style guides (as much as Tennis expert would like this to be the case).
7. "Grandfathered" will require that most readers hit the link; I'm unsure that I understand it having consulted the linked article (it's exclusively US English, it says). You need to spell out what it means in this context, here. Why is this raised in your explanatory text, but not mentioned in the proposed wording? Either don't mention it or pose it as a separate question for people to respond yes or no to. Number the questions.
8. If LSD's suggestion is to be taken seriously, you need to ask whether articles may be inconsistently autoformatted/not autoformatted. That is the principle at issue.
9. The full US format needs to be exemplified without the comma, too; a large proportion of mark-ups are thus (the system adds the comma). Tony (talk) 12:31, 15 November 2008 (UTC)

Should bare linked years be systematically removed?

The current wording of MOS:SYL regarding date linking is disputed. The linking of bare years in Wikipedia articles, for example [[2002]], was not covered in the undisputed version of MOS:SYL. Discussion regarding whether or not to treat these types of links as deprecated, as well, often arises in tandem with questions regarding the de-linking of dates for auto-formatting purposes.

In order to facilitate future discussions on this matter, it is proposed that the following wording be added to MOS:SYL:

Proposed Wording

• Linking to years: Years should not be linked, unless there is a reason to do so. More information can be found at WP:CONTEXT#Dates. Unless it is determined that a link to a year provides useful context to an article, such links should be removed from articles when encountered.

Discussion

Please discuss your support or opposition to the above proposed wording here.

• Even relevant linked years should only be linked once per article. --Gerry Ashton (talk) 18:23, 13 November 2008 (UTC)
• Same comment as above.--Kotniski (talk) 18:52, 13 November 2008 (UTC)
  • This is of course covered by WP:OVERLINK.—MDCollins 23:32, 13 November 2008 (UTC)
• The problem with this wording is that it sounds like deprecation of systematic bare year linking were something new. This deprecation has been the guideline for years though (which is the reason the MOS:SYL and MOS:UNLINKYEARS shortcuts exist in the first place), and was only removed in August, without discussion, probably because Greg didn't think it would be needed anymore. I note that this deprecation has now surfaced as disputed, so I don't have a problem with asking whether status quo should be changed, but the wording of the question should preferably not lead people to the wrong conclusion as to what the status quo is. -- Jao (talk) 19:08, 13 November 2008 (UTC)

  Yes, I don't understand the statement about the "undisputed version of SYL". What version is that, and on what basis is it determined that it is "undisputed"? Date linking was previously covered elsewhere in MOSNUM than in SYL, so it's hardly surprising that this point was absent from SYL at some time.--Kotniski (talk) 19:32, 13 November 2008 (UTC)

• I'd suggest removing the double negative conditional: "Dates should only be linked if..."LeadSongDog (talk) 21:31, 13 November 2008 (UTC)
• The proposed wording as currently framed addresses the point from my point of view. Greg L (talk) 05:41, 14 November 2008 (UTC)

  So can you answer Jao and my questions above about versions of SYL?--Kotniski (talk) 08:41, 15 November 2008 (UTC)

• Yup, it involves WP:CONTEXT and WP:MOSLINK too. And this is old old old. WP has well-established, mandated guidelines on this. Come to think of it, those two pages and MoS are involved in the previous issue, too. The wording is repetitive. No exception is granted for chronological pages themselves (this I'm willing to concede, even though it makes better sense to rely on the very prominent, user-friendly nav boxes at the top of those articles—MOSLINK's rule against adjacent "links" is broken all over these pages currently, and the day-month or year openings are much easier to read and more attractive when bolded—I've seen it a few times). Tony (talk) 12:42, 15 November 2008 (UTC)

Enforcement of the removal of deprecated links

A third question encountered in discussions regarding these types of links is whether or not the use of bots or other automated tools is appropriate. It has been pointed out that there are both benefits and costs to the use of such tools when dealing with potential exceptions.

Please note that this section will be largely irrelevant if the above proposals fail to achieve consensus.

If the community concludes that bot-assisted removal of deprecated date links is appropriate, a mechanism to designate certain date links as exceptions that the bot should leave linked, will be developed.

Discussion

Please discuss below whether enforcement of the above proposals (should they gain consensus) should be undertaken by bots or fully automated scripts. Script-assisted edits overseen by editors are presumed to be allowed in all cases, provided they are not fully automated.

• There should be a distinction between script-assisted edits compared to fully automatic bots. --Gerry Ashton (talk) 18:24, 13 November 2008 (UTC)
  • I cleared it up per this request. Shereth 18:29, 13 November 2008 (UTC)
• "...irrelevant if the above proposals fail to achieve consensus..." seems presumptuous. The above proposals merely clarify what has already achieved consensus, so it will reasonably be argued that if there is no consensus for a change now, the current guidance should remain. And there is nothing in the current guidance to prohibit enforcement with the help of bots or any other tools.--Kotniski (talk) 19:40, 13 November 2008 (UTC)
  • "Enforcement" of a guideline is a non-starter. Guidelines suggest and persuade, not coerce. If you want enforcement, get a policy.LeadSongDog (talk) 21:38, 13 November 2008 (UTC)
    • Agree with Kotniski and LeadSongDog. Also, I believe this section is irrelevant, as it is adequately covered by WP:BOT. Ohconfucius (talk) 02:21, 14 November 2008 (UTC)
• The activity of a bot is going to bring new editors with their hair on fire over this. If we are going to presenting bot activity as something to vote on, we have to be prepared to tell these editors “it’s been discussed, it’s settled, and you missed out.” Are we prepared for this? Greg L (talk) 05:43, 14 November 2008 (UTC)

• this seems like a more neutral way of stating it:

It has been pointed out that there are both benefits and costs drawbacks to the use of such tools. when dealing with potential exceptions. Please note that this section will be largely irrelevant if the above proposals fail to achieve consensus.

• and this would i think be clearer/more accurate:

If the community concludes that bot-assisted removal of deprecated date links is appropriate, a mechanism ways to designate certain date links as exceptions that the bot should leave linked will be developed.

in fact ways to designate exceptions are needed so that humans can recognize them as well, but maybe it's better to bring that up during the RfC itself ... Sssoul (talk) 08:19, 14 November 2008 (UTC)

1. "A third question encountered in discussions regarding these types of links is whether or not the use of bots or other automated tools is appropriate." Specify at the opening, please, or move the discussion to the use of bots/scripts in any situation. Are you proposing that they be wiped from the project altogether?
2. I agree with Kotniski's point: consensus has already been demonstrated.
3. Sssoul's first point also needs to be pinned down. Sssoul's second point—this is confused. Style guides determine the recommended practice; bot permission pages et al. determine permission for the operation of bots. If you want to suggest that bot permission should look favourably on proposed functions that comply with the style guides, I think that should be proposed further up the chain, not at MOSNUM. Tony (talk) 12:52, 15 November 2008 (UTC)

if "[my] second point" is the one about "ways to designate certain links as exceptions", i was thinking of techniques like introducing a non-date word - eg "[[1981|Other notable events of 1981]]" so that both bots and humans would recognize the link as being valuable to someone; i believe Masem has suggested using some kind of template to achieve the same thing. anyway my main point was about the wording: the ideas that have been suggested for designating some links as exceptions aren't "mechanisms". Sssoul (talk) 13:13, 15 November 2008 (UTC)

ps: i believe Shereth's point in mentioning this in the (proposed) RFC is to clarify right away that bots can be programmed to leave some links linked. i agree that that's worth clarifying for newcomers to the discussion, to reduce the amount of time/space that will need to be devoted to pointing out that there are ways to "exempt" some exceptional date links. Sssoul (talk) 11:14, 16 November 2008 (UTC)

Additional points that might be worth adding

Preserving metadata around dates

Several editors involved in this discussion are not opposed to delinking per se, but are concerned that we're losing some specific information about these dates which will be hard to recreate after all the existing links have been removed. It might be too late for this already since so many have been removed. A template seems to be the preference to preserve the metadata and could be included as part of whatever is decided above. dm (talk) 05:53, 14 November 2008 (UTC)

• That whole discussion appears to be exclusively about DoB and DoD. It is easily solved by a concerted effort to insert the {{persondata}} into biographies throughout. Ohconfucius (talk) 08:35, 14 November 2008 (UTC)

• I dont agree with that sentiment, though those two dates in particular are certainly often discussed. dm (talk) 14:10, 14 November 2008 (UTC)

• It is an unreasonable burden on editors to expect them to add multiple redundant versions of metadata. For example, the hypothetical markup {{Persondata|NAME=John Doe|DATE OF BIRTH={{date|January 1, 1999}}. . .}} is excessive. So it would be necessary to develop a list of recognized templates, and including a date within any of those templates would count as marking it with sufficient metadata. (This assumes there is really any value in this metadata idea; whether marking every random date is worthwhile is doubtful.) --Gerry Ashton (talk) 15:05, 14 November 2008 (UTC)

• If a proper working DA is given to us by MediaWiki, then every date should be kept with metadata. This should be a simple one letter template ("d"?) and presently spit back out what it was given. --MASEM 15:10, 14 November 2008 (UTC)

Why? If there's a feature in the software that we don't want, we can ask for it to be turned off for English WP. It certainly isn't axiomatic that we want such a feature. Those asking for metadata to be kept should say whether they think editors should be asked to mark up their new dates in the same way. If not then it seems a bit pointless to mark up the old ones; if yes then this unnatural burden on editors should be justified with reference to some kind of real value to the project.--Kotniski (talk) 08:38, 15 November 2008 (UTC)

I agree with Kotniski and Gerry Ashton; I find nothing to recommend the retention of metadata (it's a non-issue, and dates in any case can be re-recognised by scripts/bots, as easy as pie). Some editors tried the "STOP, we want to gather statistics using the existing square-brackets" argument, but failed to say how, what and why, and did not in fact deliver any stats that were even vaguely useful or reliable. Some editors have said "STOP: we might want to re-introduce date autoformatting in the future". I hope not; it's been a complete and utter failure. Again, dates can easily be recognised by scripts and bots. Ohconfucius, I'm very uneasy about the use of many templates, and this is one of them. Exactly what is the benefit, does it necessarily render dates autoformatted, and if so, can the DA be removed by bot or script? Tony (talk) 13:00, 15 November 2008 (UTC)

• You are right. However, some users insist that it is essential to preserve the metadata. They disagree that bots can put the 'metadata' back just as easily as taking it away. I merely thought that if they feel it is so vitally important that they were prepared to disrput the de-linking work already under way, they should they should put their money where their mouths are - I felt that doing so in such a way which is already established by the Biography project would be best, irrespective of whether I agree with the general practice of templating. At least this form of templating is invisible to the reader, and those editors who couldn't give a shit about it could easily ignore it. Ohconfucius (talk) 13:00, 16 November 2008 (UTC)

Approaching this from the wrong direction

Having thought about this a bit, I think this RFC is coming from the wrong direction. See my comments above about what happens if there is no consensus. All the first two points are about is making some proposals to clarify what we already have, effectively just adding sentences that mean the same as the sentences that precede them. Obviously if we have guidance that says how articles are to look, it doesn't just mean text written after the date of the guideline, but existing text too, so it is completely obvious that it is appropriate to edit articles to bring them into line with that guidance. The proposal therefore amounts in effect to a minor cosmetic change of the sort that we normally don't even bother discussing. To initiate a major discussion based on this is a waste of time, since whatever the outcome is, the result will not change the substance of the guideline anyway.

The question about bots and scripts and so on ought not to be an issue considered together with this one. Anyone is entitled to use semi-automated tools to make their editing easier and faster, as long as that editing conforms to the guidelines. Bots need permission, but there are special fora for that. Having agreed on a guideline, it is clearly desirable that articles be brought into line with it, so bots which do that (if the known technical problems have been addressed in particular cases) are beneficial to the project. Guidelines don't need to explicitly give permission for their being upheld by bots. It certainly isn't the job of a Manual of Style to address the technical issues of automated tool use.

What ought to happen here, AFAICS, is that those who claim they "dispute" the present wording of the guideline - wording that (in substance) has very long established consensus in the case of years, and less long but clearly and explicitly established consensus in the case of autoformatted dates - to actually make a proposal to change that wording in some substantial way. I don't see how we can allow a disputed tag to remain on a section when there is not even an explicit proposal on how to change it. The impression created at the moment (perhaps it will be proved wrong) is that a small group of editors who don't like something - but have no reasonable arguments against it - are bypassing the normal procedure for creating and changing guidelines, and claiming that the noise they are making and the red question marks they are slapping on it serve as some kind of evidence that it doesn't enjoy consensus.--Kotniski (talk) 08:30, 15 November 2008 (UTC)

As usual, Kotniski puts things into very clear perspective. Tony (talk) 13:02, 15 November 2008 (UTC)

First of all, it's retarded that this is on a user subpage. But second of all, regarding the dispute tag, I was quite explicit in my suggestions on potential ways forward: 1) My first proposal was a community wide straw poll on whether or not we should have linked dates/autoformatted dates. This was met with general disapproval from the MOSNUM regulars because they (incorrectly) believe they already have consensus. 2) My second proposal was to provide a mechanism to format dates without linking them, and to make this formatting available to non-logged in users/readers. Again I was met with apathy from the MOSNUM regulars who preferred to maintain the status quo of mass delinking despite obvious objections (and that niggling issue of no consensus). Now we have this new userspace subpage which appears to be asking the wrong question (do we want to unlink deprecated date links or let them slowly fall away as editors proceed with normal edits) when the right question is to determine if the community wants to deprecate date links/formatting at all in the first place (though I admit the current question would serve as an obvious secondary question so we can decide if we want bots/AWB/scripts running around performing edits which may consist only of edits removing date links.

Now, first things first, this needs to be moved someplace in Wikipedia space. Second, we need to decide if any of the options I've presented previously would be worth following (rather than ignoring and allowing to be archived). Once we've done these things we should be in a better place and everyone should be happier with how MOSNUM is operating. —Locke Cole • t • c 09:48, 16 November 2008 (UTC)

I think the reason it's in user space is that it's not yet a Request for Comments per se, but a draft to make sure that a soon-to-be-live Request for Comments is asking the right questions - a request for comments on the request for comments, if you will. And if your proposals failed to convince people, then that probably means that WP doesn't want to go in that direction. As to (1), there was a very wide discussion on the issue, in which it was clearly established in the end that autoformatting was a misguided idea. To initiate another such discussion or poll so soon after the other one, when no new arguments have apparently been put forward, would tend towards the disruptive. As to (2), this would be, as I've said before, an unnecessary and unwelcome solution to a non-problem, and the fact that there wasn't any enthusiasm for the idea probably indicates that most people tend to agree with me. I'm not saying you shouldn't continue to make proposals if you want to (I'm doing that all the time, and most of them don't get anywhere, which is just too bad), but if you can't get support for them, and indeed serious objections are raised to them, then maybe you should consider letting them go and moving on to other issues.--Kotniski (talk) 10:17, 16 November 2008 (UTC)

Locke Cole, here's where Shereth explained why it's not "retarded" to have this in user space for now. the crucial bit is:

"I have drafted a formal request for comment on the matter; the draft is currently User:Shereth/MOSNUM. I would like to get input from the folks here at MOSNUM on how well formulated the questions are and how it might be improved."

hope that helps Sssoul (talk) 11:22, 16 November 2008 (UTC)

With that limitation in mind, and the fact that it's asking the wrong questions given the ongoing dispute, this RFC will fail to achieve anything of value. —Locke Cole • t • c 22:21, 16 November 2008 (UTC)

• Welcome to the twenty-first century, Locke Cole. User:Shereth has been with you for quite a step of the way. The reason why he appears to be parting company with you and so phrasing the questions being discussed is he also accepts that date-linking is deprecated. I guess he is more open to accept that date-linking should be removed altogether now that its raison d'être is gone. I know I'll be mixing my metaphors, but will say that the boat has sailed on your dream, the parrot is dead, so it's high time you opened your eyes to the reality and stopped flogging the dead horse. Your gripes are getting very tired and no-one is listening to you any more. Your preparedness to enter into edit-warring on de-linking is a pathetic attempt to drag everyone through the Time Tunnel. It's time to turn the page and move back to the future. What's more, you seem pretty close to it on occasions and could get yourself banned per WP:3RR. Ohconfucius (talk) 13:16, 16 November 2008 (UTC)

• I'll be ignoring you, your contributions, and anything else you should do to antagonize me in this dispute. —Locke Cole • t • c 22:21, 16 November 2008 (UTC)

This is not an RfC

There is no request for comment; there is no provision for getting comment from outside the same circle of users who have been repeating themselves for months. Since this is in user space, there is even less likelihood that this conversation will be overheard than at WT:MOSNUM; which may account for the usual bores making even more doctrinaire claims, with even less evidence, than elsewhere.

We have now reached that late stage in the bankruptcy of an ideology: the proclamation of the Wave of the Future.

It was a nice idea, Shereth, but it isn't working. Septentrionalis PMAnderson 20:10, 16 November 2008 (UTC)

I think the idea that Shereth was trying to do is make that the language of the RFC was in place before hitting the wider group of editors to make sure the clarification that is needed is obtained. That is, once this was finalaized, it would be moved to a subpage of WP:MOSNUM and a formal RFC issued. This approached with previously contentious topics (WP:N for one) to make sure that the final RFC was well focused to get to the point. --MASEM 20:31, 16 November 2008 (UTC)

Agreed, this page has failed as the proponents of date unlinking are unwilling to put their primary question up for wider community discussion (the one that twelve editors decided on an obscure subpage of the MoS). It's unfortunate that it's come to this, but given the attitudes of the MOSNUM regulars it's not altogether shocking. —Locke Cole • t • c 22:24, 16 November 2008 (UTC)

Getting a bit tired of both sides here, but I have to agree with you on this point. While I fully side with Tony, Greg, Ohconfucius et al. on the actual issue, I certainly don't share their view that the consensus established on WT:MOSNUM, without an RFC, cannot be questioned or that it cannot have changed since then. As a number of people (apart from yourself, Arthur Rubin and Tennis expert come to mind) actually have questioned that consensus, why not include the question in the RFC? I'm pretty convinced, for reasons already stated by Tony and the others, that the community at large agrees with the deprecation, but what can possibly be the harm in actually and explicitly asking the community whether it wants the MOSNUM change reverted? -- Jao (talk) 22:57, 16 November 2008 (UTC)

You are the first I've seen other than the disruptive foursome who actually agrees with them. (And it is not clear that you agree on the key question of entrusting this to automated or semi-automated editing.) So I doubt the consensus; which is the usual reason why Faultless Leaders cannot have their decisions questioned. Septentrionalis PMAnderson 23:52, 16 November 2008 (UTC)

right, PMAnderson: this is not an RFC, it's a draft of one, as Shereth explained here. the crucial bit is:

"I have drafted a formal request for comment on the matter; the draft is currently User:Shereth/MOSNUM. I would like to get input from the folks here at MOSNUM on how well formulated the questions are and how it might be improved."

now i'm getting curious about how people find this discussion without reading that bit, but (smile!) i reckon that (like missing the same confusion + explanation above) is no doubt symptomatic of some late phase of something-or-other ... anyway i hope it's clearer to you now Sssoul (talk) 20:51, 16 November 2008 (UTC)

Can we PLEASE get this RfC going?--User:2008Olympian_chitchat^seemywork 18:36, 18 November 2008 (UTC)

On the grounds that since its wording pleases no-one on either side, that must mean it's just right?--Kotniski (talk) 19:38, 18 November 2008 (UTC)