19 AprSqueeeee!

Patrick Lambe of Green Chameleon linked on his blog to my post about the whole Popline thing. ZOMG! If I were the ZOMG-run-around-the-house-fanning-my-face type, I’d totally be doing that. In fact, perhaps I will go do that for a bit. Then I suppose I should read his book since I got it for my birthday & it sounds like he addresses some stuff in it that will support what I’ve been saying for a couple years & I’ll probably get to learn even more which is a super bonus!

ZOMG! < runs around the house fanning my my face >< settles down momentarily >

Of course, he points out what I left palely implied rather than said directly, which is that this is why taxonomy governance is very, very important & the implications of structure & restructure can have far reaching affects & need to be considered in the process of re/structuring & developing taxonomies. I tend to push for more than one taxonomist to work on a project, just because we do bring our own biases to the table, but no one ever wants to pay for that. Meh to them. I will continue to push for it because I think it’s a really important point.

I saw a taxonomy yesterday that illustrates why these things need to be thought out… it was for a Christian organization & had a bunch of top level terms, among them “Programs”, “Products’, “Human Rights” & “Human Wrongs”, etc. Trying to recall this correctly, they had things like genocide under “human rights” & genital mutilation under “human wrongs”… Wha? I don’t necessarily consider a morality-based taxonomy necessarily wrong — taxonomy by it’s nature reflects a world/domain view. I do consider wrong arbitrary categorizations that are so murky that one would find genocide under human rights & genital mutilation under human wrongs. That’s just… it’s stupid. Someone thought they were being cute & clever & they’re so not. Then again, this is the same organization whose adopt-a-child uses the shopping basket metaphor. Pick a kid, put it in the basket, get some accessories. It’s just like a doll. Or, y’know, human slavery, speaking of human wrongs. I do believe we said, “Uh, not cool metaphor there” but they’re resisting, is the story coming back to me. Not my project though so I only get bits & pieces of it. The taxonomy containing human wrongs is an internal taxonomy used for tagging then directing content to a specific location, so it’s not displayed externally anywhere, but still. One of the things we think about both in “taxonomy” and information architecture/user experience is the story of the perspective/world view the organization of information is telling. I really think these people should take a good look at the stories they’re telling. Not just but especially because of their morality-based views.

In other work-related news, I think I’ve settled on a personal definition for ontology. There’s the formal definition of these things (which no one can agree on) & then there’s how everyone uses them. I’ve seen taxonomy used for a huge range of things & I’ve seen ontology used just as broadly. The general public doesn’t get as much exposure to the concept of thesaurus except as “that book with synonyms & antonyms that sits next to the dictionary” so I just don’t hear it used very often, except correctly. So ontology. I’ve been sitting in on ontology tool reviews for work in the last couple weeks. What I’m gathering from what I’m going to consider formal ontology tools (TopQuadrant’s TopBraid, Ontology Works, Protégé, Swoop, & the like) is that in addition to just the semantic relationships with named edges that a complex thesaurus can support, there’s a whole ‘nother layer of logic & inference that can be applied on top of the ontology that isn’t generally available in a thesaurus tool. So, I can call out in a complex thesaurus an “Bob’s yer uncle” associative relationship (Robert is your male or female parent’s brother), but I have to manually make that relationship, just like I’d have to make associative relationships to call out cousin relationships, etc. Very time consuming. In an ontology, I can write a rule or query that says, “[find * where] parent’s male sibling = uncle” and gather those inferences rather than having to explicitly call out those relationships. In addition, in a thesaurus, there’s a preferred term & then equivalent terms, authority control lists are good in that they can help diverse domains know what a preferred name (think legal or other authoritative/formal name) is, which works to varying degrees of success. In an ontology, there’s one concept which can be any number of data types, which can have associated term names associated with the concept, allowing more flexibility for multiple domains/systems to work together — as long as the concept is agreed on, they can pick the name they want to use, but it still ties into that central concept.

I know — so what. So, where this all comes into being very handy is for research, say, pharmaceutical research. Make a rule that goes across a series of drugs to find possible contraindications for use, or a query that finds possible new uses for drugs. The flexibility for complex querying of knowledge bases is an order of magnitude greater in ontologies than it is in thesauri & taxonomies. I don’t necessarily see the clients my company has using true ontology, but I have no doubt that some of them will come to us asking for “an ontology”, just like they come to us asking for “taxonomy” when they need a content audit. But it is good to be clear in one’s own head what each is, even if we smile & provide them with a complex thesaurus that has semantic relations/named edges.

I’m rambling — I’m not entirely clear on all this yet, but I think I’ve made great strides in the last week or so. Damn, I love what I do for a living. Taxonomy/Thesauri/Ontology is just about the best job ever. Anyway, I’m still sorting it out a little in my head & I logged on to write about the baking I did today… perhaps the baking is another post now though. :)

Comments are closed.