XFML template

I wrote template to publish an XFML feed. It still could do with some additional MT wizardry, but it works.

Step 1. Create a new index file in your MT admin screen. Set your output file to xfml.xml Copy the text of this text file into the template.

Step 2. Rebuild and check if the generated file displays as valid XML.

Step 3. Add an XFML button and a link tag to your main template.

Step 4. Save your XFML file on your computer, go to Facetmap and import it. Behold.

Another teaser image of the

Another teaser image of the metadata authoring tool I’m playing with. Interface-wise, I’m going for ugly but efficient. I’ll worry about looks later. When playing around and building a metadata map, I realised I needed a way to easily manipulate topics in all facets really quickly – the following interface allows for that, without resorting to fancy DHTML that I wouldn’t be able to program anyway. The stars indicate parentage – just add a star to make a child topic. You can type “delete” after a topic and it gets deleted. Type “facet [22]” after a topic and it gets moved to another facet. Ugly but fast.

This is exactly the sort

This is exactly the sort of thing an upcoming metadata authoring tool I’ve been working on will make a lot easier: indexing other people’s content. Once you have good tools for indexing, and a format for sharing the data, a lot of things become possible. I’ve been writing a taxonomy and indexing some information architecture sites with the beta of the tool, and I can’t wait until it’s finished, because that will mean other people will index too. And I’ll be able to tap in to these other indexing efforts. Boy!

I’m trying to find different

I’m trying to find different methapors and ways of explaining faceted classification. “It lets you triangulate when searching” seems to work well. The “It’s like a diamond, you can look at things through different facets” seems less convincing.

Seb’s Open Research: “faceted classification consists in choosing several attributes (“facets”) and classifying things under particular combinations of values for these attributes. The appeal of this scheme is its combinatorial power that lets it automagically make room for things that don’t exist yet.”

Mo’ Faceted Classification

Faceted classification is quietly taking off. Here’s a quick overview of some implementations.

The original browsing tool for faceted classification was Flamenco, a research project at Berkeley university, which happens to also be the home of much of the more interesting research on how humans categorize the world (think Lakoff). But anyway.

Further research is being done in universities, often focussing on trying to implement a system using faceted classification for access to information. Facet project at University of Glamorgan for example. Or FATMKS at UCL.

Commercial implementations exist as well: bpallen (responsive engineers) and Endeca (hidden behind an almost inpenetrable wall of marketeese) both provide systems for browsing information with FC. They don’t seem much better than the Flamenco system, interface-wise, though.

In a category of its own we have Facetmap: it’s interface is excellent and evolves fast. Facetmap is the work of one man.

Faceted classification (FC) is also being taught. This homework paper (PDF) is an excellent tutorial for FC in itself – check it out. Really. If you’re still confused, here’s an excellent Berkeley slideshow onclassification systems. More papers are yours if you keep digging: Experiences with a Faceted Classification Scheme in a Large Reusable Software Library (in short: faceted classification alone isn’t enough). But we were talking about implementations.

I believe implementations will quickly reach an acceptable level of maturity. The interesting bit comes when real life experiences start coming back, and when easy availability of tools allows more experimenting. I think faceted classification theory will have to become more rich, and that richness will come from the experiences with interfaces. And not just browsing interfaces, creating faceted metadata (or any kind of metadata really) is a huge challenge, and tools can really help there as well. And then distributed metadata. Anyway, my point is: tool availability is coming, and that’s good because that will allow us to experiment and then refine the theory. I’m gonna shut up now.

I’m re-reading “Women, Fire and

I’m re-reading “Women, Fire and Dangerous Things”, and came across this discussion on Re: Basic Level Categories. Linguistics is one of the most interesting fields around – I’m also trying to follow a linguistics course my girlfriend is taking, and it already changed many of my preconceptions. (Such as that of a dialect not being as valuable as the official language somehow.)
As for basic level categories, I’m sure the insights gained from the research around categories can be useful for information architects.
Yum. Slideshow on categorisation at Berkerly. Berkeley teacher papers. Finding the Flow in Web Search. (pdf)

The metaphors of interviewing

gladwell on the metaphors of interviewing: “Not surprisingly, interview specialists have found it extraordinarily difficult to persuade most employers to adopt the structured interview. It just doesn’t feel right. For most of us, hiring someone is essentially a romantic process, in which the job interview functions as a desexualized version of a date. We are looking for someone with whom we have a certain chemistry, even if the coupling that results ends in tears and the pursuer and the pursued turn out to have nothing in common. We want the unlimited promise of a love affair. The structured interview, by contrast, seems to offer only the dry logic and practicality of an arranged marriage. “

The difficulty of categorization: “I

The difficulty of categorization: “I also wonder whether classification experts simply cultivate the perception that classification is extremely difficult. Even manual classification can be done quickly, if my experience with professional indexers is any indicator. It’s not unusual for a professional indexer to generate a comprehensive, high-quality back-of-the-book index for a new title in less than three weeks.”