Issue: LD+JSON for wiki pages

Issue Tools
- View Changes

January 24, 2024 9:03 AM

mazzly

Junior Member

LD+JSON for wiki pages

We need this to compete with other Wikis

If you go to (almost) any other wiki, their site's include LD+JSON data for the page. This is preferred by search engines and Google uses it to (sometimes) display nice-looking snippets already in the search results.

e.g. the page https://simpsons.fandom.com/wiki/Homer_Simpson has the followin LD+JSON:

Code:

{
   "@context":"http://schema.org",
   "@type":"Article",
   "url":"https://simpsons.fandom.com/wiki/Homer_Simpson",
   "name":"Homer Simpson",
   "headline":"Homer Simpson",
   "mainEntity":{
      "@type":"Thing",
      "url":"https://simpsons.fandom.com/wiki/Homer_Simpson",
      "name":"Homer Simpson",
      "image":"https://static.wikia.nocookie.net/simpsons/images/b/bd/Homer_Simpson.png/revision/latest?cb=20221226145155"
   },
   "about":{
      "@type":"Thing",
      "url":"https://simpsons.fandom.com/wiki/Homer_Simpson",
      "name":"Homer Simpson",
      "image":"https://static.wikia.nocookie.net/simpsons/images/b/bd/Homer_Simpson.png/revision/latest?cb=20221226145155"
   },
   "author":{
      "@type":"Organization",
      "url":"https://simpsons.fandom.com/wiki/Homer_Simpson?action=credits",
      "name":"Contributors to Simpsons Wiki"
   },
   "publisher":{
      "@type":"Organization",
      "name":"Fandom, Inc.",
      "logo":{
         "@type":"ImageObject",
         "url":"https://static.wikia.nocookie.net/750feb85-de88-4a4f-b294-8b48142ac182/thumbnail-down/width/1280/height/720"
      }
   },
   "abstract":"Homer Jay Simpson (born May 12, 1956) is a man from Springfield and the protagonist of the animated television series The Simpsons. He is a crude, ignorant, and slobbish individual, although he is fundamentally a good person and shows great caring and loyalty to his family, friends and on occasion, to those he barely knows or those he considers his enemies. Homer works as a low-level safety inspector at the Springfield Nuclear Power Plant in Sector 7G, although he is often incompetent and neglig",
   "image":"https://static.wikia.nocookie.net/simpsons/images/b/bd/Homer_Simpson.png/revision/latest?cb=20221226145155",
   "thumbnailUrl":"https://static.wikia.nocookie.net/simpsons/images/b/bd/Homer_Simpson.png/revision/latest?cb=20221226145155"
}

Another example from https://lotr.fandom.com/wiki/Lurtz

Code:

{
   "@context":"http://schema.org",
   "@type":"Article",
   "url":"https://lotr.fandom.com/wiki/Lurtz",
   "name":"Lurtz",
   "headline":"Lurtz",
   "mainEntity":{
      "@type":"Thing",
      "url":"https://lotr.fandom.com/wiki/Lurtz",
      "name":"Lurtz",
      "image":"https://static.wikia.nocookie.net/lotr/images/b/bf/Lurtz_2.jpg/revision/latest?cb=20110527083830"
   },
   "about":{
      "@type":"Thing",
      "url":"https://lotr.fandom.com/wiki/Lurtz",
      "name":"Lurtz",
      "image":"https://static.wikia.nocookie.net/lotr/images/b/bf/Lurtz_2.jpg/revision/latest?cb=20110527083830"
   },
   "author":{
      "@type":"Organization",
      "url":"https://lotr.fandom.com/wiki/Lurtz?action=credits",
      "name":"Contributors to The One Wiki to Rule Them All"
   },
   "publisher":{
      "@type":"Organization",
      "name":"Fandom, Inc.",
      "logo":{
         "@type":"ImageObject",
         "url":"https://static.wikia.nocookie.net/750feb85-de88-4a4f-b294-8b48142ac182/thumbnail-down/width/1280/height/720"
      }
   },
   "abstract":"The Lord of the Rings: The Fellowship of the RingSaruman! Lurtz was the first leader of Saruman's scouts\u00a0created specifically for Peter Jackson's movie The Lord of the Rings: The Fellowship of the Ring. He was portrayed by New Zealand actor Lawrence Makoare. He is non-canonical as he was created for the films, and is not in the books. Lurtz is the first of Saruman's Uruk-hai to be bred, choking the first Orc he sees to death within seconds of his birth. Attempting to intervene, other Orcs (Snaga",
   "image":"https://static.wikia.nocookie.net/lotr/images/b/bf/Lurtz_2.jpg/revision/latest?cb=20110527083830",
   "thumbnailUrl":"https://static.wikia.nocookie.net/lotr/images/b/bf/Lurtz_2.jpg/revision/latest?cb=20110527083830"
}

Most of the data needed for this should be available (already rendered for og:image, og:title & og:description, but maybe some of the fields like "author.name" and some others should be definable in the options (maybe on a per-area basis?)

Would be nice if we could get this as a feature so we can compete with the other wikis that are following those best-practice standards for SEO

Issue Details

Issue Number 6407

Issue Type Feature

Project VaultWiki 4.x Series

Category Search Engine Optimization

Status Implemented

Priority 7 - Minor Features / Enhancements

Suggested Version 4.2.0 Alpha 1

Implemented Version 4.1.8

Milestone (none)

Software DependencyAny

License TypePaid

Votes for this feature 1

Votes against this feature 0

Attachments 0

Assigned Users (none)

Tags (none)

January 24, 2024 10:51 AM

pegasus

VaultWiki Team

We already include as much schema.org data as we can on wiki pages, it's just in a different format than LD+JSON. We use microdata which places the schema around the elements it is actually relevant to. The only benefit of LD+JSON that I can see would be for sites that have simple content structures that haven't implemented a more complex format like microdata, or just to make some of the HTML slightly less verbose. Since the comparisons I've seen don't show any distinct difference in SEO performance, I see no reason to mess with this at this time.

As for "author", I believe this is filled if a page has been protected with "Collaboration Prevention" (Only the author and allowed users can make changes).

If there are particular metadata items that you think should already be available based on data in VaultWiki but are not being filled, please let us know.

Reply
January 25, 2024 2:01 AM
mazzly

Junior Member
Google recommends/prefers LD+JSON: https://developers.google.com/search...ported-formats

And there is no problem having both microdata and LD+JSON on the page. (Xenforo does this for example, breadcrumbs are microdata, while other "metadata" is LD+JSON)

From what I can tell the wiki pages are now set with microdata as "WebPage" and not "Article" (which would be more specific)

Code:

<div class="vw-blocks-container" itemscope itemtype="https://schema.org/WebPage">

And it doesn't seem like there is any microdata for the image...

Having LD+JSON would also make it easier to spot if there are some problems in the data, rather than having to look for the "spread out" meta-tags..
Reply
January 27, 2024 3:53 AM

mazzly

Junior Member

Okay did some testing with the Google Rich Results tester:

Our site, running VaultWiki (Notice it only sees breadcrumbs):

Fandom site (Finds Article item, and when expanding it has all the relevant info...):

Also, btw, it is annoying that we have to go through a 3rd party service to upload/share an image of what is happening...

Reply
January 27, 2024 9:38 AM

mazzly

Junior Member

I wrote a small add-on to change the "WebPage" type to "Article", and the Rich Results tester then started showing the relevant data:

The "non-critical" issue is: "Missing field 'author' (optional)"

So if we really want to leverage the Rich Results functionality, we should probably change the type from WebPage to Article...

Also, I noticed that the "headline" is defined, but set to empty (""), would make sense to have that one match the title of the page,

And as far as microdata for image and thumbnailUrl those would also make sense to include (use main icon image if defined)

Reply
January 27, 2024 9:52 AM

pegasus

VaultWiki Team

Google has its own rules for this structured data in order to appear in its "Rich results", which is not quite the same as for search results in general. A page that passes validation with a lot more data at schema.org (the group that comes up with the metadata stuffs) fails Google's "rich" test and it finds only some or none of the same data. From what I can tell, Google's "Rich results" only support a small subset of the various data types and keys. We'll see if we can get wiki pages performing a little better in rich results.

Unfortunately, it's not as simple as "changing the type" from WebPage to Article. WebPage and Article are distinct types, with different keys and data that are allowed under each one (something along the lines of WebPages can have children, but Articles can't). We'll see what we can do here.

Reply
January 27, 2024 12:06 PM

mazzly

Junior Member

Originally Posted by pegasus

Google has its own rules for this structured data in order to appear in its "Rich results", which is not quite the same as for search results in general. A page that passes validation with a lot more data at schema.org (the group that comes up with the metadata stuffs) fails Google's "rich" test and it finds only some or none of the same data. From what I can tell, Google's "Rich results" only support a small subset of the various data types and keys. We'll see if we can get wiki pages performing a little better in rich results.

Unfortunately, it's not as simple as "changing the type" from WebPage to Article. WebPage and Article are distinct types, with different keys and data that are allowed under each one (something along the lines of WebPages can have children, but Articles can't). We'll see what we can do here.

I've tested and at least the Page and Category pages validate fine (both schema.org validator and Rich Results tester)

Both https://schema.org/Article and https://schema.org/WebPage are inheriting from https://schema.org/CreativeWork so they should be rather interchangeable, but of course good to do due diligence and check that it works fine in all (or most) scenarios.

But seeing as both Wikipedia and Fandom (haven't checked others) are using the Article type, it should also be good enough for our needs

Sadly, Google is the king when it comes to SEO practices, so best to do as they like if we want to compete in the SERPs

Reply
June 30, 2024 1:41 PM

pegasus

VaultWiki Team

Due to new information, mainly discovery of the quirk that inline structured data does not import properties from additionalTypes, we have reversed our decision on this and LD+JSON metadata will be included in the next release. Some orphan inline metadata properties may still linger for a few releases.

All tabs now use the "Article" type, except for Feeds, which use the "Blog" type.

For wiki pages with collaboration prevention, for private feeds, and for comments, the "author" is the user who posted it.
For other wiki articles and feeds, the "author" is the "Contributors to {$siteName} {$wikiName}" which is a new phrase vw_ld_author that can be customized if needed, such as if you have a long $siteName or it sounds awkward beside $wikiName.

A side effect of populating a generic "author" is that we still needed a URL, so we will now have a separate tab to display an article's contributors with their avatars. There will now be an option to hide the contributor list from the main view and use only the tab.

All wiki content now includes the "publisher", which is the web site. We use nearly the same logic that XenForo 2 uses to fill the publisher field.

For tabs other than the main article view, we point to the "mainEntityOfPage" being the main article view, which should hopefully avoid any dings for these other tabs not actually being an article. Referring to the Article type in this way allows us to include comments or article listings in a way that is sort of semantic and that will lead to rich results.

We considered referring to comments as "DiscussionForumPosting" but it would still require similar "mainEntityOfPage" magic, as well as a lot more changes that didn't seem to add any benefit just to satisfy various required fields. In the end, a "comment" to the "Article" is closer to what is actually going on.

Reply
July 13, 2024 5:46 PM

Alfa1

Distinguished Member

I think having the Comment type for Comments makes much more sense, as it DiscussionForumPosting is something different and doesn't have an article attached. See:
https://developer.yoast.com/features...ieces/comment/
https://schema.org/Comment

Content with DiscussionForumPosting seems to have low value for Google compared to Comment.

Reply

+ Reply

All times are GMT -4. The time now is 2:36 PM.

This site uses cookies to help personalize content, to tailor your experience, and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.

Learn more… Accept Remind me later

Welcome to VaultWiki.org, home of the wiki add-on for vBulletin and XenForo!

Issue: LD+JSON for wiki pages

Issue Tools