Semantic HTML for artwork – can you help?

(Kicking this blog back into life to pursue a solution to a problem)

The Dream

Imagine a Google Image Search where a potential customer could search for contemporary artwork that is available for sale in a specific price range which is no larger than a specific size (the right size to fit on the wall in their new living room) – and have your artwork come up in the search results.

Imagine a Google Image Search where an art student or academic could search specifically for woodcuts created between 1550 and 1700 matching certain keywords – and get accurate results.

The Problem

I’ve been thinking recently about the semantic web and artwork, or more specifically: “what’s the best way to markup artwork in HTML so that third parties such as search engines can best understand what my web page is about.”

Personally I don’t care whether the semantic markup is achieved through HTML5 microdata or RDFa Lite, so long as it doesn’t involve me having to doing anything other than output HTML with specific properties and values.

Schema.org

I’ve been looking at schema.org, which is supported by the major search engines, and they have http://schema.org/CreativeWork which has three relevant sub-types: painting, photograph, and sculpture.

They all seem a bit basic, though – for example, for a painting there is no way to semantically markup the dimensions or painting type (oil, acrylic, watercolour, etc) – both of which are fairly important and commonly recorded properties of any painting.

And printmaking and drawing (amongst other media) are not represented at all.

If Not Schema.org, then what?

So I’ve looked around for some other vocabularies, and there are some around but they don’t really fit my criteria.

There are a variety of metadata formats for the arts, but many rely on producing separate XML files of semantic data, rather than just marking up your web pages semantically. As a web developer this rings alarm bells because it breaks the principle of DRY (Don’t Repeat Yourself!), and it also doubles the amount of work you need to do – not a good thing for a busy artist with limited time.

So, I went back to first principles, and made a short list of what my ideal semantic markup would be:

• It must be open and free to use
• It must be easy to implement as either HTML5 microdata or RDFa Lite.
• It must be something that major search engines are likely to adopt (I’m not doing this for fun or from some utopian dream of the semantic web!)
• It must have a rich vocabulary that allows fine classification of artwork
• It must be valuable for artists, galleries, museums, and aggregators (e.g. search engines)

OK, back to Schema.org

All of which took right back to Schema.org. It’s free to use, it can be implemented in either HTML5 microdata or RDFa Lite, it’s already supported by the search engines – that’s the first three criteria met.

However, as stated above, it doesn’t have a rich vocabulary that allows the fine classification of artwork (and it misses out printmaking and drawing completely), which reduces its value.

You can extend Schema.org but it would seem pointless doing this in isolation because Google and Bing aren’t going to suddenly adopt my extensions because I’ve invented them – I’m guessing it would need a certain critical mass of artists, galleries, museums, and universities adopting a set of schema.org extensions to make it worthwhile for search engines to support them.

A call for proposals, suggestions, or existing solutions

So I’m putting this out there to see if anyone else is struggling with the same issue and wants to collaborate on either a small number of extensions to schema.org for various Creative Works to make them really useful for semantically marking up artwork in HTML (and likely to be adopted by artists, galleries, etc.), or alternatively if someone knows of an existing solution that meets my short list of criteria above.

Your comments, as always, are most appreciated.