I enjoyed reading the various essays from A List Apart regarding design and usability, but the piece for the week which most engaged me was the article by Elings and Weibel regarding shared metadata standards for museums, archives, and libraries.
My job at the historic house and what I am now doing with CHNM both came down to assigning keywords, metadata, to historic documents and (at the house) objects. One of my roles at the House was to propose, evaluate, and define new keywords for our relational database. As a result, I’m aware of the benefit of a controlled vocabulary, as well as the challenges which accompany it.
As I think about it, the challenges fit well with the readings about design and architecture of websites. Both situations force the builder or implementer to look at the audience, or audiences, they plan to serve, and how the audience(s) will interact with the data they provide.
When you create a controlled vocabulary, you want the terms to be specific, descriptive, and usable. The problem is, usable and descriptive for what audience? An internal objects database for a museum could use terms like linen press or SOMETHING and be fairly certain that the users – the curators – would know what was meant. Once you start making collections information available to populations not in your specialty, the very detailed language you deploy to convey a lot of information in a few words can suddenly be more confusing than helpful. Recently, somewhere I was party to a conversation about the Library of Congress subject classification system; one of the people said they could never determine ahead of time how to phrase a subject search to return the desired results.
This is where design and information architecture comes to my mind. Elings and Weibel describe the difference in the way a date is displayed to a user – “late 14th century” – and how it is made machine readable – a field defined as 1375/1399.1
On the one hand, it makes me think about how much of the curatorial-generated metadata needs to be displayed on the site; whether the entry for an object is narrative with embedded metadata or more like a catalog listing with the fields out in the open.
On the other, I reevaluate how much information I need to put on any web page or site. I now know how much more information there is about every single museum object than ever makes it to the exhibit floor or even to the digital object catalog. Writing exhibit label text is an exercise in concise, carefully pointed and considered writing.2 Does the web page or site section you’re building need to be like an essay in an exhibit catalog, exhibit label text, or a basic field display catalog entry? Once you know that, you can decide how much of your controlled vocabulary needs to be deployed at the visible level, and what only needs to be visible to you and those who read your source code.
The need to consider audience and just how much you need to say makes me think of the great rule of woodworkers, seamstresses and tailors everywhere: “Measure twice, cut once.” Which is to say, the more you plan and consider ahead of time, the more efficiently and elegantly your digital project will (hopefully!) work in the end.
1 Mary W. Elings and Günter Waibel, “Metadata for all: Descriptive standards and metadata sharing across libraries, archives and museums” First Monday [Online], Volume 12 Number 3 (5 March 2007).
2 see Suzanne Fischer’s recent post about writing label text.