Wikipedia:Miscellany for deletion/15 automated portals built on a single list

The following discussion is an archived debate of the proposed deletion of the miscellaneous page below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the page's talk page or in a deletion review). No further edits should be made to this page.

The result of the discussion was: delete. — JJMC89 (T·C) 01:43, 25 April 2019 (UTC)[reply]

15 automated portals built on a single list

Every one of these fifteen portals is based on a single list article. (Note: not a navbox, as was the case with e.g. the 1,390 portals deleted at WP:Miscellany for deletion/Mass-created portals based on a single navbox).

With these portals their list of selected articles (using {{Transclude list item excerpts as random slideshow}} or {{Transclude linked excerpts as random slideshow}}) is drawn solely from a single list. For example, Portal:Plums is drawn solely from List of plum cultivars, Portal:Bears solely from List of bears etc.

This makes each of these portals merely a WP:REDUNDANTFORK of the list. It simply repackages the list as a one-item-at-a-time gallery of the lede of a random subset of the pages linked from the list.

This seems to me to be much less useful than the original list, which at least shows all the page titles together and need no click-to-see-next-item. List articles also load more quickly, without the intensive Lua overhead involved in generating these portals.

Maybe some editors believe that this is a better way of viewing a list, but if so we shouldn't glorify this alternative view with the title of "portal". It's a processor-intensive alternate page view.

I propose deleting all these portals in one go because:

1. Being each built on a single list, they add no navigational utility, and are an inferior WP:REDUNDANTFORK of another navigational tool
2. Portals are not content; they are merely a means of navigating between content. So their mass deletion removes precisely zero encyclopedic content.
3. The pages have no prior history. There is no non-automated version to revert to.

Few of these topics seem to me to be broad enough to satisfy WP:POG's requirement that portals be about "broad subject areas, which are likely to attract large numbers of interested readers and portal maintainers", and many of them look like similar topics to portals which have been deleted in recent weeks. However, consensus remains unclear about where to draw the lines, and a few of them may be broad enough to support a thoughtfully-designed and properly-curated portal which used a selected article list extending way beyond the navbox, if enough editors were willing to do the sustained the hard work needed to curate and maintain it.

So, I propose that these pages be deleted without prejudice to recreating a curated portal not based on a single list, in accordance with whatever criteria the community may have agreed at that time. --BrownHairedGirl (talk) • (contribs) 23:16, 17 April 2019 (UTC)[reply]

Discussion (15 automated portals built on a single list)

add your keep/delete/comment here

• Delete. As outlined in the nom, these portals are merely forks of the associated lists. —pythoncoder (talk | contribs) 23:54, 17 April 2019 (UTC)[reply]
• Comment. This actually seems to be one of the better ideas. I looked at Portal:Bears and it seems much more user friendly than List of bears, which has no description and only one image, and it also draws content from the article bears for its introduction. (Though mysteriously the images include a seal, raccoon, polecat, ocelot and wolf?)

@BrownHairedGirl: could you comment on what kinds of portal, if any, you would be prepared to support? Because if one can't use a static list, nor the portal creator's judgement (per Wikipedia:Miscellany for deletion/Portal:Global issues), there seems to be not a lot left. Espresso Addict (talk) 00:57, 18 April 2019 (UTC)[reply]

• @Espresso Addict #1 "one of the better ideas"?. An alternative view of a single page is not a portal. It's just a WP:REDUNDANTFORK. We should have one or the other, but not both.

Anyway, please try a little experiment for me. (And @AfroThundr, please try this too.) Either log out and view List of bears, or use your web browser's "open in private window" function to view it. What you'll see then is that mouseover on any of the linked list items shows you the picture and the start of the lede.

This is hidden from editors who use WP:POPUPS to provide an editing/analysis menu in place of the preview, but most of our readers (for whom we create this stuff) are not logged in. Using the default view of Wikipedia, the bare list of bears actually gives readers much the same functionality as the so-called "portal" view, but without all the overhead of a "portal". I have tested this in Opera, Firefox, Google Chrome, and MS Edge; it works on all of them.

So, far from adding something new, the so-called "portal" is just a clumsy way of doing what the Wikimedia software already does by default. It's an un-needed, duplicative, superfluous product of the department of redundancy department. A kinda Rube Goldberg machine. --BrownHairedGirl (talk) • (contribs) 11:26, 18 April 2019 (UTC)[reply]

Thanks for this, BrownHairedGirl. It's easy to forget that readers see something quite different. I tried looking at the article logged out; the pop-ups work rather poorly for me (we have a satellite internet connection with a big latency) but they are an interesting development, which presumably does work for most readers. For some reason they seem to remove material in parentheses (eg in this case the species name, which is what I was looking for). They also cut after a fairly short extract in the middle of a sentence. (I can imagine what some MfD participants would say about a portal that did that.) Certainly food for thought about making sure the first sentence or two of the lead is informative in isolation.

Personally, I'd prefer to change the list into the tabular format where every item has a brief summary and an image, and was wondering whether the automated portal code could be adapted to provide that. Having made some tabulated lists manually they are hard work and difficult to maintain.

Re your final comment: one could say something similar about Category:Bears; redundancy is not necessarily an evil, and imo does not necessarily require deletion. Espresso Addict (talk) 21:03, 18 April 2019 (UTC)[reply]

Au contraire, @Espresso Addict. The central principle of WP:PORTAL is that a portal should be an enhanced version of the head article. This is not enhanced, so its redundancy is grounds for deletion. --BrownHairedGirl (talk) • (contribs) 13:08, 19 April 2019 (UTC)[reply]

I think we are down to arguing about may delete vs must delete. Espresso Addict (talk) 20:13, 19 April 2019 (UTC)[reply]

• @Espresso Addict #2: "what kinds of portal"?. First, let's dispose of the Portal:Global issues question. That is a unique problem, of a portal with a wholly subjective scope. Other topics may have boundary issues, but every other portal I have seen has scope with a solid core. However, as I outlined at WP:Miscellany for deletion/Portal:Global issues, the problem there is no core at all. It's a sort of Rorschach test, in which your response reveals your political preferences and priorities.

My answer to your main question draws on WP:PORTAL. It says in the lede "Portals serve as enhanced 'Main Pages' for specific broad subjects". That is expanded further down that page, but the core definition is good, and the key word there is "enhanced". Basically, a portal needs to offer a lot more than is available on any other page.

The problem is that a half-decent Wikipedia head article already offers a lot. A set of key images, an overview of the topic, list of subtopics, and one or more navboxes linking to the key articles on a range of subtopics and related topics. That's a pretty well-enhanced page to start with.

They key issue is that most Wikipedia head articles are so heavily interlinked that even a modestly well-written head article on a topic is of itself a portal. This isn't like the mid-1990s web, when web pages were mostly plain text with a few links at the top and the bottom; rich interlinking is now the norm, and Wikipedia is an exceptionally fine example of that interlinking. Compare a Wikipedia article on a recent event (e.g. Notre-Dame de Paris fire, 2016 United States presidential election, 2018 Winter Olympics) with any major news website, and you will find that the Wikipedia page has at least one order of magnitude more links:

• Notre-Dame de Paris fire: 191 links
• 2016 United States presidential election: 1,706 links
• 2018 Winter Olympics: 530 links

Try that test for yourself on a range of other major topics: e.g. Ireland, Bill Clinton, Leonard Cohen, Charles Dickens, Nelson Mandela, San Diego, Buenos Aires, Cricket, Video game, Novel, Poetry, Earthquake, Apartheid, Social democracy, Afghanistan, Bird. You'll find the same pattern: squillions of links, with the key ones well-structured through embedded lists, navboxes and sidebars.

So basically, a head article on en.wp is itself a damn good portal. That sets a very high bar for a dedicated Wikipedia portal which aims to meet that WP:PORTAL definition of an enhanced main page.

A 1990s portal merely needed to do significantly better than pages which were like a 1919 Ford Model T. That wasn't hard: just a few dozen links, and you were already miles ahead. But by analogy a 2019 Wikipedia portal has to do significantly better than the Tesla Model X, which drives itself in silence whilst massaging its occupants in climate-controlled luxury. If you wanna beat that, then you need to include things like an orgasmatron, a beauty parlour, or a food-and-drink service.

The fundamental problem with the portal-creating crew is that they haven't been asking that question. They have been taking the existence of portals as axiomatic and asking "how can we make it easy for editors to add more stuff", when they should have been asking a completely different question. They should have been looking at the pageviews which overwhelmingly show portals getting less than 1% of the pageviews of a head article, and recognising that readers clearly do not find that portals are a sufficiently enhanced version of the topic's main page. So they should have been asking "what will make portals so much better than the head article that readers will seek them out?"

I have a few preliminary thoughts on what those ingredients might be. For me, the absolute key ingredient is that the portal offers a structured selection of pages which is not available anywhere else. These list-forked pseudo-portals clearly fail that basic test abysmally. A portal which passed my test would group subtopics in sets by topic (e.g. a country portal would have separate groups for topics on govt/politics, history, arts, biography, geography, economics, demographics etc) and by quality (FA, GA, FL, etc).

Less crucially, it'd be good to have e.g.:

• featured pictures grouped by topic, with links to related articles (a set of pics without links is not a portal; it's just an image gallery).
• snippets drawn from DYK, but only if they are actually curated. The search-based selections are comically inaccurate, as illustrated by the many risible mismatches which @Legacypac has so helpfully posted in these MFDs
• An "in the news section", but again only if it actually curated. Again, search-based selection doesn't work

All that is a lot of work to set up, and it needs a lot of ongoing work to maintain it. The history of en.wp portals is that the limited editorial energy is spread over far too big a set of portals. The inevitable result is most of them are simply crap: they don't even try to offer much more than the topic's head article, and even within that limited ambition, many of them are broken and/or out-of-date.

The solution adopted last year by the portals project was to develop automated tools to reduce the maintenance burden. There was a lot of clever programming, and it brought some good changes which simplified the maintenance process, However, that was all undermined by an ideological drive to use technology as a substitute for curation ... and the inevitable result was that the utility of the portals was severely degraded. Fewer portals were actually broken, but the price paid was that fewer of them even attempted to do anything better than the head article.

On top of that, the project then allowed the new tools to be used to create about 4,000 new portals using this automated technology, the overwhelming majority of which are just automated crap, and most of which are on excessively narrow topics. (Many dozens had less than five articles within their scope; at least one portal actually had zero articles in its scope).

Shamefully, the Portals Project took absolutely zero action to restrain this crapflood. Uninvolved editors who expressed concern were shouted down, and eventually the whole thing exploded in a storm at WP:AN, the village pump etc. That led to a cleanup, which has been driven almost entirely by uninvolved editors.

My MFD nominations have tackled only the surface of this mess. I have developed a set of tools and techniques to identify the newly-created automated crap, and brought about 3,000 new crap portals to MFD, with overwhelming support.

When that lot is all processed, it will leave about 1,300 portals. They are mostly of poor quality and low ambition, and overwhelmingly with pitiful pageviews. We will need a long RFC process to decide what to do with that mess ... but until then I am only MFDing the portals which add absolutely nothing, like the 15 nominated here, plus a few occasional narrow-topic portals. --BrownHairedGirl (talk) • (contribs) 13:22, 18 April 2019 (UTC)[reply]

Thanks also for this, and the clarification re Portal:Global issues. I have a lot of things to say in response to your general points but this is not the forum. Might I respond on your talk page or should we take the discussion to the portals project talk page? Espresso Addict (talk) 21:55, 18 April 2019 (UTC)[reply]

Sure, @Espresso Addict. Copy it across to WT:WPPORT, and we can continue there. --BrownHairedGirl (talk) • (contribs) 10:47, 19 April 2019 (UTC)[reply]

• Comment - User:BrownHairedGirl - I think that your nominating statement above needs to be tweaked. You wrote: "So, I propose that these pages be deleted without prejudice to recreating a curated portal not based on a single navbox, in accordance with whatever criteria the community may have agreed at that time." But these portals are not based on a single navbox. They are based on a single list. Or is your objection to these portals only that they have not been curated? Robert McClenon (talk) 02:58, 18 April 2019 (UTC)[reply]
  • Thanks, @Robert McClenon. That should of course have read "not based on a single list". Now fixed. --BrownHairedGirl (talk) • (contribs) 10:56, 18 April 2019 (UTC)[reply]
• Comment Unlike navbox-based portals (when the navbox is the sole source), these don't seem like bad ones. I'm leaning towards keep if the only rationale is being based on a list. I also echo Espresso Addict's sentiments. — AfroThundr ^{(u · t · c)} 04:02, 18 April 2019 (UTC)[reply]
• Delete basing the selected articles on a list article links is a path to automatic errors. I picked one topic randomly and because I'd not looked at it before - Portal:Ghost towns which pulls from links at List of ghost towns by country. The list contains wikilinks to not only actual ghost towns but regions, geographic features and various other articles. Therefore the portal automatically brings in Lake Encumbene which is a redirect to Eucumbene Dam which does include a little about a town it submerged. That might be a reasonable inclusion but Calabria (the toe region of Italy population nearly 2 million people) is NOT a ghost town and contains no ghost town content. Automated selection of articles via a list (or any other) article does not work, period. See discussion on Portal:Kathmandu for another example. There I proved an innocent edit to add a wikilink in the base article introduces inappropriate inclusions into the portal. Legacypac (talk) 05:52, 18 April 2019 (UTC)[reply]

It seems to work better for very simple lists such as List of bears. Espresso Addict (talk) 08:20, 18 April 2019 (UTC)[reply]

• Delete. Portal:Tells (archaeology) caught my eye, and is a particularly silly example of automated portal creation. We barely have a single article on tells, so there isn't any content for this to be a "portal" to other than the random collection of tell sites copied from list of tells, which have little in common except for their morphology. It's akin to having a Portal:Buildings with cellars. – Joe (talk) 09:09, 18 April 2019 (UTC)[reply]
• Delete 15 - BrownHairedGirl has provided very useful statements about why this list of portals is not useful and about why portals are often problematic. I have checked a few of the portals in her list of list-based portals, and I concur with her judgment and that her judgment should be accepted for the 15 portals listed. Robert McClenon (talk) 15:02, 18 April 2019 (UTC)[reply]
• Delete all these 15 portals

Pldx1 (talk) 16:30, 18 April 2019 (UTC)[reply]

• Comment. Portal:Seas looks like an old portal, but isn't. The 2013 edits were rather related to Portal:Nautic. Pldx1 (talk) 16:30, 18 April 2019 (UTC)[reply]

Pldx1 (talk) 16:30, 18 April 2019 (UTC)[reply]

• Speedy delete per G7 for Portal:Board games Gazamp (talk) 09:52, 22 April 2019 (UTC)[reply]
• Delete all per nominator and WP:POG. Even the creator of one of them has asked for it to be G7'd. SITH (talk) 12:27, 24 April 2019 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the page's talk page or in a deletion review). No further edits should be made to this page.