Dmitry wishes to see more blogs in XTM format.

Now, this is far from being a Topic Maps driven blog, or blog in XTM-format, but I figured it would be fun to write a Topic Maps “feed” plugin for Wordpress (inspired by Lars Marius Garshol’s Blog Metadata in Topic Maps).

So from today (or evening, if you wish), this blog delivers an XTM 1.0 feed. The plugin’s source code is not perfect — and kinda ugly, but feel free to look at it and/or use it if you wish to (I wont guarantee anything, of course, not even the accuracy of the outputted topic map ;). (more…)

Recently, I stumbled upon a comment (Norwegian) about the semantics of Web 2.0 technologies, which prompted a response.

Since the response turned out rather lengthy, and a bit off-topic, I decided to post it here. Hopefully, it can be of use to others as well.

Web 2.0

Although the Web 2.0 has been hyped as something big and new for a while now, and things such as tag clouds are a popular part of many Web 2.0 sites, Web 2.0 sites build on old technologies (+user generated content). It’s all just (X)HTML + JavaScript and CSS. No magic.

There is not much semantics to Web 2.0. The semantics of HTML is still limited to layout/presentation of information (as will HTML5’s be). Even though one can express the start/end of a paragraph or heading, one can for example not explicitly state what an article is about (except through limited meta elements), or how different “topics” — or even what kind of topics we’re talking about — are related.

There is no way to explicitly state that “this article is about X and mentions Y”.

The tag clouds of Web 2.0 are in effect just free-floating words, and do as such not bring anything new to the table (an exception is Roy Lachica’s Fuzzzy.com where tags are structured in an ontology using Topic Maps).

The Semantic Web

The last couple of years / decade, people have been working towards a more “intelligent” web (although hardly intelligent); the semantic web. Such a web, often referred to as Web 3.0, adds the possibility of expressing semantics outside the limited domain of (X)HTML (or micro formats).

This is accomplished through adding layers of metadata to the web. The semantic web is therefore an extension (not replacement) of todays web, where different technologies, like Topic Maps and RDF/OWL, allow us to express anything about any subject - in a structured, explicit and unambiguous manner. I.e.: not limited to the element set of a given vocabular.

By doing so, one can (hopefully) bring order to the chaos, and develop better applications. Luckily, the technologies are already there, and by using Topic Maps one can e.g. boost findability.

For more on the semantic web or Topic Maps, see e.g.:

• Wikipedia - Semantic Web
  
• Wikipedia - Web 3.0
• TopicMap.com - Resources

Topic Maps Snippets

Some people at Bouvet, incl. Lars Marius Garshol and Arnar Lundesgaard, started a Topic Maps Snippets blog last year. It’s already a (the?) good source for news on Topic Maps.

TMRAP

Lars Marius Garshol wrote a post last week on how he has set up his tmphoto application to fetch remote information on topics using TMRAP. A good example on how PSIs can be used to exchange data about the same subject across different application domains, and an interesting read.

AToMs 2007 Slides

The presentations from AToMs have been made available online. Lots of intriguing stuff there.

For example, Steve Pepper seems to have been thinking a lot about converting the world to a subject-centric approach of computing lately.

I’m sure that, in a lot of ways, he’s right about how the human mind does not operate on documents, but subject-centric and associatively. At least that’s what my psychology classes told me, and most people experience on a daily basis: thinking of something might lead your thought wandering off to something completely different, but related (e.g. strawberries => milk, or even The Beatles (Strawberry Fields Forever)). When humans learn, we hook the new experiences onto existing knowledge.

Moving to a subject-centric way of computing and information management may be a good way of moving things forward, and, by nature, Topic Maps surely has its role in a subject-centric world. I’m sure we’ll hear more about it at Topic Maps 2008 (there’s also some more slides touching upon the topic).

FYI: Topic Maps 2008 is being held in Oslo, Norway on April 2-4 2008.

Topic Maps 2008 is targeted towards users and potential users of Topic Maps, in particular:

• Information architects, software engineers, and project managers working with portals and web sites;
• Knowledge officers responsible for intranet-based knowledge management;
• Corporate and commercial publishers;
• Public sector providers of information;
• Students and teachers of computer science and library and information science.

Links: Call for Presentations, Keynote Speakers.

This month’s Scientific American (Dec 2007, 297:6) features an 8 pages long article entitled The Semantic Web in Action by Feigenbaum, L. et.al.

The authors basically discuss the Semantic Web vision as laid out by Tim Berners-Lee, et.al. in the famous 2001 The Semantic Web article, and how some steps towards such a Semantic Web have been made.

In it, they define the semantic web as

A set of formats and languages that find and analyze data on the World Wide Web, allowing consumers and businesses to understand all kinds of useful online information.

A tad general perhaps, but anyways… On the technology front, RDF/OWL and some uses and applications of these technologies, like FAOF, is mentioned, while a few use cases are discussed.

Although the use cases are interesting, the authors fail to even mention Topic Maps as a semantic web technology, and limit the technological discussion to RDF/OWL alone, stating that “The data language, called Resource Description Language (RDF)…” (emphasis added). (more…)

During his open space session on a possible Topic Maps driven Web 3.0 (or whatever one may want to call it) at TMRA 2007, Graham Moore rushed past a slide that mentioned ontology driven user interefaces.

To my knowledge, researchers have already touched upon this problem (see for example Furtado et. al., 2001) but I don’t really know whether anyone has been successful in generalizing or implementing such an approach.

Perhaps this is a problem more easily solved using Topic Maps than alternative technologies..? The concepts of a given UI can of course be represented in an ontology (using e.g. ontology annotation as supported by the OKS). Combining UI-ontologies, or fragments thereof, with Topic Maps’ scope could perhaps be a way of realizing “multiple UIs”?

Most web application use a specialized UI, targeted towards a primary user group. The layout of the UI is probably dictated by a single template, or by a collection of “static” template fragments. What if one used scope to support “switching” between various types of “templates”? Certain template fragments for certain topics under a certain scope?

An application could, for example, “learn” from the user’s behavior and present an alternative UI based on the user’s interests … in some way. Accordingly, one - especially public service providers - could, for instance, create alternative, more accessible interfaces optimized for disabled users (and thus avoid Section 508 lawsuits, if located in the US :D). This would bring “Skip to content” to a whole new level. One could of course also offer alternative views based on user type, interests, preferred form of visualization (e.g. text centric vs. image centric), etc.

Hmmm…

This year I was lucky enough to get to participate at the TMRA 2007 conference in Leipzig, Germany.

It was very interesting to hear some of the most prominent topic mappers giving talks on a range of interesting subjects, and promising to see that the community is getting there and that the missing standards work is not too far away.

I did also enjoy the tutorials on TMQL and CTM, and having done some work with XQuery in the past - at least enough to have experienced how powerful and easy to use it is, I really liked TMQL’s XQuery-like notation for simple expressions. For example, given a topic type identified by person, the expression: (more…)