Which metadata should Tabi include and how?

[Almost-Senseless-Coder]

In the spirit of getoutofmyhead.dev I've been contemplating what the best approach to adding (custom) metadata to Tabi might be.

Right now, Tabi does a fairly good job of only including metadata which is absolutely necessary. However, to support correct importing of blog posts into Zotero, Tabi would (at least) have to add <meta name="DC.type" content="blogPost"> as well as a <link> importing the necessary Dublin Core namespaces and vocabularies.

It would also have to include the publication date nof the blog post, either through article:published_time (or something similar, I'm not 100 % sure right now) from Open Graph or DC.date form Dublin Core, or both. At that point, I noticed that partials/header.html is ued both in the blog post content and the archive, likely any context where page.html gets used as the template. Which isn't a big deal, I just conditionally add the date if it's available, but now my archive page (erroneously) gets classified as a blog post too. Which I could fix by spinning out a separate header template for the archive page but... This just adds a lot of fluff and unnecessary complexity, especially since header.html is one of the partials that changes frequently, and every time it changes in Tabi, I have to go and merge the change into my custom version too, so... not exactly ergonomic.

I propose the following options:

• Have a separate header partial for blog posts and archives in Tabi itself, and set all the metadata accordingly there (or possibly conditionally import extra snippets into the header partial depending on whether we're using this as an archive or a blog post).
• Give a user the option to set the type of the page manually, in the [extra] section; ideally this should be doable on a section level, but that, again, leads to complexity as /blog very likely should have a different type than /blog/some_post.html
• Offer an option to set sustom metadata, be it as a string, or as an array of strings, which get included into the template. This has the advantage that it gives the user more freedom to include their own metadata, allowing Tabi to only provide minimal metadata for its functionality, while letting them leave Tabi's templates alone. However, this would likely lead to a lot of duplication in the [extra] metadata (though a clever person could probably write themselves an appropriate macro to reduce boilerplate).

I don't think there's any one "correct" approach to this and some or all of these approaches could get mixed and matched.

It's also not super important--but I'd be curious to see which of the approaches, if any, you choose. ^^ If you decide that this shouldn't be something Tabi provides and leave that particular responsibility squarely with the theme users, that's also fair.

[welpo]

Tabi would (at least) have to add as well as a importing the necessary Dublin Core namespaces and vocabularies

Is the <link> a lot of code? Does it provide any benefits, beyond Zotero imports?

Also, I think this will interest you: #528 :)