Talk:Opinion polling for the 2015 United Kingdom general election/Archives/2019/February

Splitting

The following is a closed discussion of a split request. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section.

The result of the split request was: No article split. LightandDark2000 (talk) 22:26, 20 May 2018 (UTC)

---

Wikipedia:Article size This is one of the longest pages on this site and needs to be split for compatibility with browsers. I would suggest splitting it by year, since that is how sections are in the article already. E.g. 2010 Opinion polling for the 2015 United Kingdom general election, etc. ―Justin (koavf)❤T☮C☺M☯ 16:42, 3 February 2017 (UTC)

I think splitting off the sub-regional polls might be a better option? Absolutelypuremilk (talk) 20:31, 3 February 2017 (UTC)

Justin, can you specify which browsers have compatibility issues? Rami R 20:52, 3 February 2017 (UTC)

@Rami R: I can't tell you anything more than the guideline above but at the very least, I would imagine that excessively large/long pages would be difficult to display on mobile devices (if not impossible) and of course older browsers or persons with spotty Internet access (which are about 2 billion). ―Justin (koavf)❤T☮C☺M☯ 20:55, 3 February 2017 (UTC)

[Citation needed]. Rami R 21:40, 3 February 2017 (UTC)

Potentially, Opinion polling for the 2015 United Kingdom general election in Scotland, Opinion polling for the 2015 United Kingdom general election in Wales, etc. could be used to reduce the length of this article. Clyde1998 (talk) 18:39, 19 April 2017 (UTC)

I believe this page is more of an archive than an article meant to be read. It would likely suffer more from fragmentation than there is to gain with slow browsers by splitting it. Perhaps the long tables could simply be set to collapse by default? Nicolas Perrault (talk) 00:02, 24 August 2017 (UTC)

• Oppose - The largest factor in the huge article size (in kilobytes), are the charts and graphs, which are definitely not an issue. Also note that the majority of this article consists of the graphical features. Unless the prose content of this article is going to be substantially expanded, this article should not be split up, as it would likely result in unnecessary fragmentation of the topic. LightandDark2000 (talk) 07:28, 11 May 2018 (UTC)

I think it's clear the proposal didn't gain consensus. --Dweller (talk) Become old fashioned! 09:52, 11 May 2018 (UTC)

---

The above discussion is preserved as an archive of a requested article split. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.

Redux

This page's source code is currently 656,160 bytes. That's ridiculously long. The downloaded page size (desktop) is 650.9 KB. That's also ridiculous. The question should not be whether to split the page, but how. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:36, 21 December 2018 (UTC)

I support to split the articles into the following format:

2010 → 2010 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2010

2011 → 2011 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2011

2012 → 2012 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2012

2013 → 2013 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2013

2014 → 2014 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2014

2015 → 2015 opinion polling for the 2015 United Kingdom general election or Opinion polling for the 2015 United Kingdom general election made in 2015

Expected article size per page is between 80-100KB ‑‑V.S.(C)(T) 06:56, 22 December 2018 (UTC)

Support in theory, but these titles are a problem. Opinion polling for the 2015 United Kingdom general election (2010)? I think there might be some technical change that can be made to the tables to reduce their size but I'm not an expert in that area. That could be much better than splitting the articles, but at this size it definitely needs to be. Onetwothreeip (talk) 07:12, 22 December 2018 (UTC)

As a start, can we split off the regional polling? Keeping the national polling for different together years makes it far more readable. Absolutelypuremilk (talk) 09:54, 22 December 2018 (UTC)

I support splitting off the regional polling, but oppose splitting different years throughout different articles. Particularly, when such a split would not be natural and would basically require users to visit different articles to check opinion polls that essentially constitute part of one same trendline (years were split off into different tables to ease editing due to the amount of opinion polls, not because they were different topical areas, which they aren't). The article's size in itself is not particularly troubling, as under WP:AS and WP:SIZERULE, it is the size of readable prose, and not the article's source size, what counts in terms of splitting requirements (in this sense, this article has barely above 4 KB of readable prose, which is ridiculously small to require a split). And it should also be noted that a great deal of the wiki markup size is taken up by the referencing (i.e. the external links and the proper references), not by proper article content. In terms of page navigation, making the tables collapsible would do the job. Impru20^talk 10:07, 22 December 2018 (UTC)

It's more than simply the amount of readable prose, especially since that is a rule that applies to articles that are not lists. WP:SPLITLIST and WP:ARTICLESIZE are more relevant for lists. Readers are reading the tables primarily and the prose secondarily, not the other way around like a typical article. All the policy points to splitting lists if and how they are naturally split already in the article, which has already been done so by year. The regional polling has now been split off. Can anyone come up with a technical method on reducing the size of the tables? This would possibly prevent the need to split the article. Onetwothreeip (talk) 10:36, 22 December 2018 (UTC)

As said, much of the size is taken up by the referencing, with each opinion poll having its own source (I made a rough calculation, with each external link being roughly 100 bytes in size, and there being about 2,000 opinion polls. This means that about 200,000 bytes or so are taken up just by the external links in each opinion poll; basically, ~40% of the article's source code size). The issue is that the term "readable prose" also excludes such sources from the count (WP:RPS), so in terms of "split need" you would still need to exclude them from the count, no matter if these are in the tables themselves. Unless you suggest removing the sources altogether (which would surely decrease the article's wiki markup size by a great deal but is definitely not a solution) I don't see any need for the splitting, or even what would constitute a practical way to split this (as said, splitting it by years would generate more navigation issues than it would solve, specially considering there is a chart showing the whole trendlines from the 2010 to 2015 period which would lose its sense if split). Impru20^talk 10:41, 22 December 2018 (UTC)

It would still be possible for this article to be split by years and include the entire chart in every split article. I would rather this article was retained as one, but that would require reducing its size by some technical means and may still not be possible. This article is most definitely too large no matter which way you want to look at it or what you want to exclude. We can't pretend there isn't a practical way of splitting this list, as we already have this list split into years, it's just something we may prefer not to do. Onetwothreeip (talk) 20:35, 22 December 2018 (UTC)

Your own reply shows there isn't any actual need at all for splitting this. Previous discussions for this very same article (and I'm not referring just to the above one) have all ended up the same way: to not split it. As has been shown, policy doesn't require the article being split (indeed, that is the case for SPLITLIST and ARTICLESIZE), as roughly half of the size is not attributable to actual content but to referencing in the source code (which is necessary, as you require a source for every opinion poll). If it isn't required, and it is acknowledged that such a solution is not only undesirable, but may actually create more issues than it may solve, why to do it?

Surely, you could split it by years. You could also split the tables themselves and have January to June polls in one article, July to December ones in a different one, for every year. You'll have nicely small articles. Would it be practical? No. Impru20^talk 20:46, 22 December 2018 (UTC)

I'm not clear why you think "roughly half of the size is not attributable to actual content but to referencing" is relevant, given the figures I posted above. Anyone wanting to read, or edit, the article must contend with both. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:57, 22 December 2018 (UTC)

I don't think splitting it is undesirable. Splitting the article wouldn't really create problems either. Looking through the source text it looks like there is a lot of redundant and needlessly long technical code but I'm not adept at reducing it. References also are absolutely relevant to the article's size, since it's not only readable prose that is the reason for splitting articles. I think we may have to split the articles by year and then we can transclude them back into this article. Onetwothreeip (talk) 21:07, 22 December 2018 (UTC)

I concur with Impru20. This isn't a regular article: it's useful for reference. I'd leave as is. Bondegezou (talk) 21:19, 22 December 2018 (UTC)

If this is not a "regular article", then it probably fails WP:NOT, and should be deleted. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:24, 22 December 2018 (UTC)

It is not a "regular article" in the sense that, as has been pointed out, every opinion poll requires a reference; a large number of opinion polls mean a larger amount of referencing, which in turn means much more source code space will be taken by the sourcing. I don't see how it fails WP:NOT, so you should please elaborate on it. Nonetheless, a formal proposal for splitting/deleting should be done according to WP:RFC/WP:AFD procedures, seeing how there has been obviously a very strong and consistent consensus in the past against splitting it (much less deleting it). Impru20^talk 21:30, 22 December 2018 (UTC)

Sometimes editors who have been highly invested in an article justify its existence and/or length by saying its purpose is different than the purpose of most Wikipedia articles. This article doesn't fail WP:NOT but if it was true that this article wasn't an encyclopaedia article like the rest of them, it would most certainly need to be deleted. As for references, it is not a given that an article needs thousands of citations, so it's possible that these could be cut down and that could go some way to reducing the size of the article. Splitting articles does not need any formal procedure and can be done as any edit can be done per WP:BOLD and WP:BRD. There aren't special articles which are free from Wikipedia's manual of style and policies. Onetwothreeip (talk) 23:08, 22 December 2018 (UTC)

Well, this discussion by itself is titled as a "redux" of a previous discussion (just one of several throughout the years) which was closed as no consensus for a split. Indeed, if this could be done so easily as per BOLD and BRD, I guess we wouldn't be here discussing on the opportunity of the splitting, right? This is controversial, which is why no one did the splitting outright. And seeing how the issue was repeteadly brought down in the past, I can't see how an eventual RfC wouldn't be helpful to ascertain the wider community's consensus if that's the case. I should note that there are dozens of opinion polling articles like this one throughout WP, all of which are affected by the same issue: a large Wiki markup size due to referencing. I don't think a pair of users should unilaterally chose that these articles should be split (thus hampering navigation and requiring to split information which is shown better if kept together) just because they are some "source code size"-need which does not exist (again, ARTICLESIZE and SPLITLIST, together with RPS).

... and, a deletion proposal was brought forward, which would indeed require going through AfD, as I can't think of any scenario where deleting such an article could be seen as uncontroversial. So not under BOLD either.

As for references, it is not a given that an article needs thousands of citations Well, opinion polls, just like every other information in WP, are not exempt of verification. I don't actually care if it's thousands or one (if that's even possible), but you can't leave them unsourced. Still, under the same premise it would not be a given either that this article needs to be split, so... Impru20^talk 23:29, 22 December 2018 (UTC)

The last discussion was "closed" absolutely nonsensically, it was by someone who expressed a strong opinion in the discussion, and there wasn't an RfC on it anyway. There is the second largest opinion polling article, so it's not something intrinsic to polling articles to be this big, and it's currently the 38th largest article overall. If it was more urgent I would probably boldly split the article but transclude the constituent articles into this article, so that the entire data set is preserved here. Pigsonthewing was not proposing this article be deleted, they were only saying that is a logical conclusion from Bondegezou's claim that this article is not supposed to be a normal article. Of course every entry needs to be cited, but the current system is probably the largest possible way of doing that, even if it's the only way. However, it is pretty well given that this article should be reduced in size, whether by splitting or by other means. Onetwothreeip (talk) 23:38, 22 December 2018 (UTC)

True, but size of opinion polling articles typically depend on the number of polls, and the sheer size of the source code is not a justification alone for splitting (precisely, splitting tables into sections by year solves any editing issues by making it simpler, unless (as I stated) someone is proposing to actually split the tables themselves further). Transclusion is a choice, but I don't find it particularly helpful or coherent to split yearly tables into their own articles with the same charts for each of these, then transclude them here together with constituency and regional polling (which, IMHO, are better left into their own articles). There is a sensation to me that the splitting (or eventual size reduction) is merely sought for the sake of splitting, but that no coherent way forward is put behind it.

Nonetheless, as said, the biggest issue with size comes with sourcing. If anyone knows a better way to source opinion polls without it taking up so much space... Impru20^talk 23:49, 22 December 2018 (UTC)

I wouldn't want to transclude the constituency or regional polls either, only the national polls. Onetwothreeip (talk) 23:59, 22 December 2018 (UTC)

Break

The standard script estimates the "readable prose size" at 4394 B (729 words), well within what WP:AS allows.

File size: 872 kB
Prose size (including all HTML code): 6884 B
References (including all HTML code): 180 B
Wiki text: 493 kB
Prose size (text only): 4394 B (729 words)
References (text only): 18 B

Bondegezou (talk) 10:41, 24 December 2018 (UTC)

Again, utterly irrelevant to the figures I quoted above. Here they are again: This page's source code is currently 656,160 bytes. The downloaded page size (desktop) is 650.9 KB. And WP:AS does not support your argument. Its very first section says quite clearly: "There are three related measures of an article's size:

• "Readable-prose size: the amount of viewable text in the main sections of the article, not including tables, lists, or footer sections
• "Wiki markup size: the amount of text in the full page edit window, as shown in the character count of the edit history page
• "Browser-page size: the total size of the page as loaded by a web browser"

You're only looking at the first. WP:AS also says (emboldening mine) that section editing "should work as long as none of the sections are longer than 32kB [of wikicode], which they really shouldn't be.". Every one of the year sections on this page greatly exceeds that length. The longest is well over three times that size. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:09, 24 December 2018 (UTC)

"This page's source code is currently 656,160 bytes. The downloaded page size (desktop) is 650.9 KB." We already had this discussion at Talk:Sub-national opinion polling for the 2015 Spanish general election: you must not take a page's source code size to determine whether a page should be split or not, and AS (specifically, WP:SIZERULE) is very explicit on it: Please note: These rules of thumb apply only to readable prose (found by counting the words, perhaps with the help of Shubinator's DYK tool or Prosesize) and not to wiki markup size (as found on history lists or other means). This is, the wiki markup size and the browser-page size are indeed two measures of an article's size, but only the readable-prose size is determinant when it comes to article splitting. WP:SIZESPLIT is also explicit when referring just to readable prose size when it comes to splitting based on size reasons. Impru20^talk 11:11, 24 December 2018 (UTC)

You are misinterpreting that text. It means that the figures quoted apply only to readable prose, not to markup. It does not mean that other figures (such as the 35Kb I quoted) can not apply to the latter. If it did mean that, it would be condoning tables of infinite length, which is clearly nonsensical. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:18, 24 December 2018 (UTC)

I don't think I'm misinterpreting anything, as I just quoted the rules applying to article splitting. On the other hand, what you have quoted on the 32Kb does not refer to splitting (in fact, that whole section does not even mention splitting at all), but to "problems editing a long article". And the solution proposed for that is in the first line of the next paragraph: you should be able to edit the article one section at a time by using the "Edit" links you see next to each header in the article. This should work as long as none of the sections are longer than 32kB, which they really shouldn't be. This means that, in order to avoid editing issues deriving from a large article size, you can edit the article by section editing. And if there are size issues that make it harder to edit, more sections or sub-sections can be added if needed (which is what was done when the tables were divided by years). That has absolutely nothing to do with article splitting, which is what this discussion is about, and you shouldn't really mix up different and unrelated quotes to try to reach a conclusion not stated in the guideline itself. In order for an article to require splitting, the readable prose size must exceed one of the figures shown, which this article very clearly does not exceed. Impru20^talk 11:27, 24 December 2018 (UTC)

Btw, just as a note, the article's source code is currently at 504,753 bytes, not 656,160 bytes. Of which roughly 40% or more is attributable to sourcing alone. Impru20^talk 11:33, 24 December 2018 (UTC)

Andy, your "utterly irrelevant" comments seems a tad harsh. WP:AS uses three measures. Leaving aside debate over the exact meaning of some phrases within WP:AS, it is clear that one needs to consider all three measures. I'm not saying readable prose trumps browser-page size, but that any decision should not be based on just one of these.

With most articles, all three measures track closely with each other. This is one of those articles where they do not (which was my point about this not being a typical article). Bondegezou (talk) 11:54, 24 December 2018 (UTC)

Impru20 Bondegezou The guideline for only counting prose applies to prose articles. This is not a prose article, this is a table article. Readers are reading these tables primarily, and not seeing them as additional elements like in prose articles which feature tables. Otherwise we could just have infinite tables. Onetwothreeip (talk) 12:59, 24 December 2018 (UTC)

That's an interpretation of WP:AS. I couldn't see anything that says that explicitly. Either way, what I think we agree on is that readers interact with tables differently to prose. So, again, I think we have to see this article as being somewhat different to a 'typical' article that is mainly prose. What WP:AS does say about tables is, "If there is no "natural" way to split or reduce a long list or table, it may be best to leave it intact". I take that to mean there is more leeway with tables to go longer when it makes sense to keep their contents together. I believe that is the case here. Bondegezou (talk) 13:05, 24 December 2018 (UTC)

You may be right. The article has five very long tables; there is no need for them all to be on the same page. That can be achieved without splitting any of them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:10, 24 December 2018 (UTC)

Given that WP:AS urges splits at >100kB, whatever "leeway" that is extended to tables has probably been exceeded when this article is five times the size of that number. Onetwothreeip (talk) 13:14, 24 December 2018 (UTC)

Onetwothreeip No, the linked guidelines (i.e. WP:SIZERULE and WP:SIZESPLIT) apply to the splitting of any kind of article. What you linked before (WP:SPLITLIST) applies to table splitting (but not necessarily splitting them into various articles, as it also covers splitting tables within a single article to avoid too-long tables; this was already done when the tables in this article were split by year). I think there's a lot of confusion with guidelines here. Not the whole of the WP:AS guideline involves article splitting, just the "Splitting an article" section of it. Despite this, I'm seeing various sections of AS being used, even if unrelated to each other, to try to justify that splitting the article is an actual must (remember how the discussion started with the claim that, under the article's size, "The question should not be whether to split the page, but how"), but AS also involves other solutions, such as section division or content reduction.

Splitting that is motivated out of the article's size is regulated under SIZERULE and SIZESPLIT, but under such guidelines splitting requirements do very clearly refer to readable prose without making a distinction on whether the articles are prose articles or table articles (with the exception that, actually, under both SIZERULE and SIZESPLIT, splitting is not even formally required for list articles, when they state that the requirements apply less strongly to list articles, especially if splitting them would require breaking up a sortable table). This is why I'm contesting there any actual "need" to split this article into other smaller ones. Of course it can still be split, but such a need, as a requirement or obligation as it was seemingly put, does not exist for this article.

Pigsonthewing On the splitting itself, we have several issues which make splitting impractical: 1) a great deal of the wiki markup size comes because of the sourcing required for each opinion poll, which makes the source code size to be deceiving, as actual readable content (either in prose or in tables) comprises much less than 500 kB; 2) as for the tables, they are already split in the most logical way possible; any further split would not be natural, specially considering they are sortable; 3) the article has been already split as much as it was logically possible to, moving subnational polling into various articles and leaving here only the nationwide polls. Any further article splitting would require the tables being moved into different articles, which would break the article's topical scope (indeed, transclusion could be a choice, but then you would end up having four or five separate articles which would be already large enough on their own (if counting wiki markup size) than just one, while also requiring people to visit/edit several separate articles to make any meaningful edit to the tables. Not practical).

(I hope I addressed all the points; I do not wish to be verbose but I think the issues with the guidelines must be clarified xD). Impru20^talk 13:32, 24 December 2018 (UTC)

The distinction between prose articles and table articles is made from the reasoning for taking into consideration the prose only. Categorising articles into prose articles and table articles is a false dichotomy though, since the content in the tables like in this article is effectively the prose of the article. I'm honestly not sure why you continue to say that much of the article's source is referencing. That is something being taken into account for splitting the article, it's not like that's accidentally being included in the "true" article size. There are six tables here, so really they might as well be their own articles. It's a bit silly making someone load five years worth of lengthy and detailed data when they only want data for a specific time of the campaign. Onetwothreeip (talk) 13:42, 24 December 2018 (UTC)

Ok, then that it's your own interpretation of the guideline, not what the guideline actually says. I remind you that you yourself made such a (false?) dichotomy to argue that readable prose-requirements were not appliable to this article (and could, thus, be split), when the guidelines pretty much hint to the contrary: if any distinction is to be made between the two types of articles, it would actually be to harden requirements for splitting list/table articles, not to ease them (They also apply less strongly to list articles, especially if splitting them would require breaking up a sortable table). And I mentioned sourcing size because verification is a true obligation of WP, so you can't dispose easily of it. Opinion polling articles are prone to have a large size due to sourcing, and applying AS as you intend would mean that opinion polling articles would be required to be split much more often than other articles (which, as said, does not seem to be what the guidelines do hint to).

Further, it is a false consideration to take the tables as if they were unrelated and could be treated differently; all the tables are inter-connected parts of the same polling trendline. The only reason they are shown separately is because of the editing and readability issues resulting from a single, never-ending table; but such issues were already solved with the table splitting and the section division. What it is suggested here is that, aside from such solutions, the article should be split anyway out of size reasons alone (based on a wrong interpretation of AS as a whole), without considering any other reasons. Nonetheless, I already stated how article splitting would be impractical here, so I won't repeat myself in that sense. Impru20^talk 13:58, 24 December 2018 (UTC)

We're going round in circles, I fear. Since this discussion was begun, the article has been cut down somewhat, particularly with some useful edits by Onetwothreeip, so we're 154k down. That's good. Beyond that, we're where the previous discussion ended: no consensus. Bondegezou (talk) 20:20, 24 December 2018 (UTC)

I think you do agree with me that table articles and list articles can be longer than the 100kB guideline, but this article is 500kB. That's a lot more than 100kB. The guideline for "readable prose", as we understand that to mean content that isn't included in elements like tables, is for articles where most of the reading is done by reading that prose. In this article, the effective prose is the data of the tables. Seriously, if we take "true size" to be equal roughly to the prose size as generated by measuring tools, we get an absurdly small size for an article like this. If we used that measure to determine when to split this article, how big would it have to be? How many decades or centuries of opinion polling data would it take until the prose size tool guides us to split this article?

The reasons for splitting this article is because it's too large, it's not because WP:AS or any other guideline says so, and the issues that arise from having such a large article. They are just guidelines that confirm the principle that an article of this size should not exist at this size. While splitting the data into six tables is commendable in making editing easier, the problem is still the same with the visual editor as if it was still one table. As for how the article is already split, it matters not why it was. The list is split already into six tables and if someone only wants to see one table they are forced to load the entire set of tables, one of the largest articles on Wikipedia. Onetwothreeip (talk) 20:36, 24 December 2018 (UTC)

We're going round in circles here. Yes, the article is 500 kB; the vast majority of which is not readable prose, and most of which is attributable to sourcing alone, with the average table article not requiring as much sourcing as opinion polling articles do (which is one of the reasons this article or other opinion polling articles are not "regular" articles. Yes, article size-splitting is regulated under the specific sections at AS and SPLIT, so you can't say that this must be split for reasons other than those provided there. This said, we can effectively agree that no need for splitting is set out in the guidelines for an article like this one, and none of the issues I brought above advising against splitting have yet been contested. Currently, the article's layout may be somewhat troubling for visual editing and navigation, but that can be solved out through the use of "collapsible collapsed" tables (such as done here). In the end, the alleged need to split the article comes down to it being "too large", yet under WP guidelines this would constitute one of the exceptions to such a need.

"if someone only wants to see one table they are forced to load the entire set of tables" The issue is that opinion polling tables aren't really meant to be split. They are split here because of editing issues (because it is easier to edit six smaller tables than a single one), but all six tables are meant as part of a single whole. This is easily verified when you check out that opinion polling tables elsewhere in WP are not typically split, unless too long. So no, people are not supposed to "see one table" and not the others, or to the very least, being required to visit/edit other articles to check the whole of opinion polling. Nonetheless, I think this would be part of a whole different discussion on content itself. Impru20^talk 11:40, 25 December 2018 (UTC)

We are largely going around in circles so I'll keep it concise. Even if much of the article is just the referencing, it is still part of articles where the >100kB rule applies. References are included in article size and if we are ignoring references then the >100kB benchmark would be lower.

Prose is meant to mean what the reader primarily reads, which in this case isn't what is normally considered prose but the table data itself. Not counting the table data when judging if the article is too long is clearly absurd, and it would mean that an article for all the opinion polling that has ever been in the United Kingdom on one article wouldn't be too long.

The easy solution here is to split and transclude them back, at least into two articles but more likely six, one for each year. The difference between is that if they want to see one particular year on a shorter opinion polling article, there is not as much else to load that they don't care to read than there is on this article.

There is great reason as I'm sure you would agree that the size of the tables and references to be reduced by technical means. I don't have any plans to split this article or do anything drastic for now, so these discussions are largely academic and I don't want this to distract you or others from more constructive editing. Cheers, Onetwothreeip (talk) 20:50, 25 December 2018 (UTC)

Again, the >100kB rule applies just to readable prose, and referencing is not readable prose under WP:RPS. Seriously, we're going round in circles, so I'll be brief: you keep telling me of your own interpretation of the >100kB rule and I keep telling you how the guidelines say that such rules should be applied. You are re-interpreting what "prose" should be, in your view, under size splitting standards, which is not what it actually is (again, RPS). You may see it as absurd, but it is what the guideline says, and I think we may at least agree on that because it is literal: Readable prose is the main body of the text, excluding material such as footnotes and reference sections ("see also", "external links", bibliography, etc.), diagrams and images, tables and lists, Wikilinks and external URLs, and formatting and mark-up. This idea is repeated several times throughout AS and SPLIT. I do not agree that the article should be split, as I see it would create more issues than it would solve, as I pointed out above (including the issues resulting from transclusion or the thinking around the "what-if-people-wants-to-see-just-one-year", so I'll not repeat myself). Impru20^talk 21:24, 25 December 2018 (UTC)

I've come to the conclusion that while reducing the size of this article is necessary, splitting this article into six is not necessary, at least immediately. This articles details an unusual length of time and an abnormally high density of polling organisations releasing data, so we should be able to split this into two articles which ought to be easy given this spreads across six calendar years. Onetwothreeip (talk) 00:54, 22 January 2019 (UTC)

I think the best course of action now would be to split the 2010 to 2012 tables out, which I intend to do shortly. This would leave 2013 to 2015 together, which is the far more relevant half that readers are looking to read given that this article is ostensibly about the 2015 general election. Because there are six calendar years here, splitting in half doesn't cut any year in half. Onetwothreeip (talk) 04:32, 9 February 2019 (UTC)