SWIG-UK Semantic Web workshop

November 30th, 2007 by Inigo

Last Friday I attended and spoke at the SWIG-UK Semantic Web workshop in Bristol, organized jointly by Hewlett-Packard Research and the Oxford Semantic Web Interest Group (OxSWIG).

It was a fun day, with a number of interesting speakers. I’d say that the common theme was that the speakers generally were using Semantic Web technologies to solve real world problems, without being too concerned about the “Semantic Web” in abstract. Alberto Reggiori of Asemantics described their “RDF on the inside, Web 2.0 on the outside” approach that they’re using for work around the BBC Memoryshare site, and this pragmatic approach was typical of the other speakers.

A talks that particularly stood out for me was the first talk, from Ian Davis of Talis, talking about their “Platform as a Service” model: providing a hosted RDF datastore with a REST API. It seems very much like a higher level version of Amazon’s S3 - but rather than providing a simple data-bucket API, they’re providing a more sophisticated SPARQL interface, as well as a full text search.

I was also very interested in what Graham Klyne of Oxford University said about their experience in providing access to scientific data on the web. I’ve been working with Graham recently, so I’m generally familiar with their work, but it was good to hear more background. They are using a “Data Webs” approach - putting a SPARQL endpoint on top of their existing ePrints repository, so the data stays in its original location. The key points from his talk were the importance of the loose coupling enabled by their current approach, the advantages of leaving data in place rather than extracting it to a separate store, and that rapid progress is possible

I talked about applying Semantic Web technologies to data integration and visualization, working with Oxford University to help visualize gene expression in fruit flies; and working to combine customer data from several sources including a CRM system and visualize it. My slides are available. In the same session as me, Daniel Lewis spoke about the intersection between the Social Web (i.e. Facebook, LinkedIn, etc.) and the Semantic Web. Unfortunately, he borrowed my laptop to do so, so I don’t have any detailed notes on his talk.

There were a few discussions on the day about a follow-up meeting. With luck, there might be another meeting, possibly in the Oxford/Milton Keynes area, early in the new year. It also sounds like Talis might be arranging a SWIG meeting for next year.

In any case, the next evening meeting of the Oxford SWIG is on Wednesday 12th December, at the Lamb and Flag in Oxford.

Why Wikipedia is a bad example of knowledge management

September 15th, 2007 by Inigo

Nowadays, most peoples first contact with a wiki is through Wikipedia. Wikipedia is a great resource, but I think it’s a bad example of using a wiki for knowledge management. The limitation of Wikipedia is that it aims to be purely a repository of existing knowledge. Wikipedia policies require that content on the site is written from a neutral point of view, is verifiable and contains no original research: even going so far as to ban synthesis of other sources. Although reasonable in an encyclopaedia, these restrictions give the wrong idea about how wikis can be used for knowledge management. They encourage the idea that wikis are for knowledge capture, rather than knowledge creation, and that their main purpose is to produce a good knowledge store.

This bias extends to commercial wiki vendors, too. Wiki vendor MindTouch describes their wiki as making “knowledge easier to capture, find and consume” (thanks to Zoli’s Blog for the link), and Atlassian describe their wiki Confluence as “lowering the barriers to knowledge capture“.

However, there are other wikis than Wikipedia, and the original Wiki was designed for knowledge creation. The Portland Pattern Repository (also called the WikiWiki, or Ward’s Wiki) was the first Wiki, created by Ward Cunningham in 1995. Although technically simple, the Wiki’s brilliance was in its design: It facilitated asynchronous conversations between distributed contributors, enabling a dialogue across time and space. At its best, these conversations led to new insights, that were then formalised into documents.

The WikiWiki was originally focused around design patterns, but by the time I first encountered it in 1997, it was starting to be focussed around Extreme Programming, and since then has drifted into other topics without a clear focus. Some of the most interesting pages on it are the longest established ones: I particularly like the patterns pages, and the proto patterns.

At 67 Bricks, we’re using Semantic MediaWiki to help us create new knowledge, as well as to store our existing knowledge. I’ll write more about it in a future blog entry.

Early days - big decisions

August 20th, 2007 by Sam

Early days - big decisions

It is exciting to be here now. Inigo and I have been discussing setting up a business for many months. We came up with an idea that utilised our skills, had revenue potential and most importantly we were both very excited about.

“Deriving value from existing information using the Semantic Web”

This ties together Inigo’s strengths in software development, my experience in running high performance teams and our joint interest in how people, information and technology come together. We believe that in certain industry sectors the key success factor is, and will become more so, the ability of a company to derive value from the knowledge its staff has access to. We want to help companies understand how to do this.

We decided that if we were planning to do this for others then we should also do this for ourselves. We are going to practice what we preach. We also plan to document this process as we go so that others can gain as much as possible from our journey.

One area we are looking into developing is a community of interest for colleagues and family and friends who are keen to support our new venture. Communities of practice and communities of interest are solutions we will be discussing with clients to support knowledge sharing.

Setting up a website using WordPress

August 17th, 2007 by Inigo

I’ve set the 67 Bricks website up using WordPress as a content management system. Previously, I’ve used either simple static HTML pages, or a traditional, full-featured CMS. I decided to use WordPress here because it is easy to use, and has an ecology of templates written for it making site design much simpler. It also describes itself as a “Semantic Publishing Platform”, and for a knowledge-management company such as our own, semantic publishing is important.

Setting up WordPress requires very little technical knowledge: my ISP, Heart Internet allows you to do it with a few clicks in their script library, but even without that, it’s still simple. The changes that I’ve made to make it work as a CMS are:

  • Installing the Navigo plugin: which makes it easy to create a menu
  • Installing Search Everything: so pages can be searched as well as posts
  • Installing WP Last Posts: to put the text of the last few posts on the home page

Then, I created the bulk of the site using WordPress “pages”, and the news items as “posts”. I set the front page to be a static page under the Options | Reading menu.

To create a design for the site, I looked through the library of templates on the WordPress site, downloaded four or five that looked interesting, and tried them out locally. Having found a template with a good layout, clean HTML code, and a reasonable license, I then customized it to remove the sections I didn’t like, and to change the graphics - a much simpler process than building a web site design from scratch.

That’s really all there is to it.

So, what did I learn?

Setting up a website with WordPress is very quick. Installation of WordPress took a few minutes, installing plugins took maybe quarter of an hour, choosing and customizing a theme about an hour, and the rest of the time was spent creating content.

The WordPress way of separating content from presentation using PHP functions works well. This is more powerful than the pure CSS approach of CSS Zen Garden, since it allows different templates to display different content. An alternative would be to use XQuery: which would probably be more versatile, but the PHP developer community hasn’t had a wide take-up of XQuery and seems generally sceptical of XML.

(thanks to twenty3design for help with choosing WordPress plugins)

About this blog

August 16th, 2007 by Inigo

Sam and I both feel that a company should be as open as possible - with its employees, with its clients, and with potential clients - which is everyone else! One of the advantages of working with a small company is that we’re able to enter into a proper dialogue with our customers, to benefit us and them. We’d also like to help out any other small companies by sharing our experiences, and there are plenty of friends and former colleagues that are interested in following our progress.

So, we’ll be using this blog to talk about our thoughts on the Semantic Web, knowledge managment and team performance, as well as some of the issues in running a small consultancy, and some technical advice on managing our websites, managing communication, and our software development.

67 Bricks is set up

August 14th, 2007 by Inigo

67 Bricks is a consultancy and software development company that focusses on delivering value to companies from their existing information  using the Semantic Web; which Sam Herbert and myself, Inigo Surguy, have set up.

Sam and I have both had experience as consultants before, so we’re familiar with setting up a company, handling accounts, and so on. We’ve also both worked a lot with the Semantic Web, knowledge management and with helping teams be highly effective. I’m also very experienced with software development. However, setting up in business together and selling software products as well as consultancy is new to both of us, and I’m sure there’ll be plenty to learn!

So, we’re looking forward to having a great time building a successful company. The next few years are going to be very exciting.