Examples of architecture use: Cars

With the launch of their “Vision EfficientDynamics Concept”, BMW provides a good example to explain what architecture means for a company. While car itself will probably never be used as a production model, it does portray the vision that it carries in all its fibers will be embedded in BMW’s future models. Together with architected versions of the more production ready models, they make a pretty good picture of where this car manufacturer is headed. And this in turn makes them a perfect example to explain how architecture should work in the field of IT.

The car resembles the vision or, in IT-architect-speak, the reference architecture. To be absolutely clear: That DOESN’T mean that the concept will be built! However, the various subsystems and quality aspects of the reference car can be incorporated into the various models that the manufacturer does make in to production models.

On the performance part, a key quality aspect of BMW, this car sets the direction in which the powertrain will be developed. This means powerful engines, torque, rear-wheel drive and low fuel efficiency. Even the look of the engine has to abide by the guidelines set in the reference model. The technical subsystems in the car are also similar through the entire product line. They might use higher quality materials on the high-end cars, but the vision behind them is the same.

On the inside of the car, comfort is a more important quality aspect. While there are obvious references to the concept, like similar knobs and materials through the product line, there are also less obvious guidelines. These, for instance, dictate how the car should feel, sound or behave. The feel of the fabrics, the feel of a switch being flicked, the sound of the engine in- and outside the car. The sound of the door closing, et cetera. They all contribute to the image the of the manufacturer. Thus, they should all be the same or, in the worst case, pretty similar to achieve the same experience by the user.

Isn’t that true for all car manufacturers? Yes, Renault and Audi (to pick two) also do a good job of visually portraying their vision for the future and incorporating similar/the same car bits in various models. It’s the various details in looks, feels, sounds, smells and experiences that make a car undeniably BMW, Renault, Audi or any make whatsoever. And that is exactly what makes the reference architectures so important: It creates a family that share the same characteristics and that efficiently make use of the same resources from a vision for the future. And in case of cars and architecture we drive the fruits of that vision every day.

Semantic what? Lets call it Information Recycling

When talking with customers and colleagues, I'm often confronted with the question "What is semantic technology?". As a standard reaction I start explaining that semantic technology can enable businesses to consistently store and reuse its information, improving the quality of the information and increasing the speed of which the information can be found/obtained. By now the listener understands, but why didn't he beforehand?

When we're talking about "semantic technology" we're talking about something, uh, technological. We should be talking about something that has meaning to the people that are going to use and/or adopt it. So, while the technology might be called "semantic", we should use the term "Information Recycling".

Information recycling captures the essence of what semantic technology does and instantly explains it to non-technical listeners. Of course, there are still questions about what to recycle and how we are going to do that. What it does is keeping us talking about solving a business problem without using technological slang.

And I bet we can easily link it to saving the environment too. :-)

Its time for the Social Semantic Web

How many times have you filled in your name on a new Social Networking site? And one of your mail addresses? And your photo? The answer is most likely: Too many times. With the Social Landscape still forming, you're probably in for many times more. What's awkward is that this problem can easily be solved.

Linked Data
The answer has been around for quite some time: Linked Data. It is one part of the Semantic Web technology and aptly called Social Semantic Web when it comes to Social Networking. Linked Data offers a universal way to store data which enables sharing.
For Social Networking, the "Friend-of-a-friend" (foaf) can be used to store one persons names, e-mail addresses, photo's (only profile, not the gallery) and all his/her friends, family and collegues. This data can in turn be used by all Social Networking sites.

The advantages are plentyful:
- With one set of your data, it takes only minutes or seconds to sign in to a new website and you can start working with it
- Updating your linked data via one of the sites means that it can be propagated to all other sites that use your data
- Linked Data can be queries, so finding a cross section of all your contacts (e.g. age less than 40 and have a pet cat)

So, why aren't the Social Networking sites jumping to implement this for you? It could be one of many reasons. Here's two:
- There is no business model for it (yet). How to earn money off just data? The data just generates traffic while other companies display it on web pages with ads. Asking money for it just kills the whole priniple of sharing.
- Sharing data with your competitors is not-done. This is where many companies have a paradigm to overcome. It will become benificial to share data with your competitors.

Lets hope one of the big players (Google, Facebook, Microsoft) sees the light and starts to offer a service that gives you control over your data. Hopefully we can all spend more time collaborating and socializing... and less time just filling in the same data again and again.

Information Integration Explained

While I was working on a presentation on Semantic Wikis I tried to find an image which captures the benifit of the integration of information in supply chains. After searching for a while, I found the image with "the tree and the swing". It effectively captures how (perception of) information can change during interdisciplinary processes. I decided to change the original to show how things should and can be when integrating information through the chain:

... and I'm not quite sure about the billing part yet. :)

more information about the tree and swing pictures can be found here, here.

Infinite Social Media feedback loops

Automated publishing mechanisms, like RSS, Twitter, Facebook and, soon, Google Wave are great innovations. They are used to make people communicate better. But maybe they can also potentially be service killers.

Feeds could be configured into infinite loops: An item is published on a blog-site, which triggers an RSS feed post, which is published on twitter, which is published on the blog-site mentioned earlier, etc... and there you have your loop. And this is a fairly elaborate example. One could think of one to many or many to many publishing mechanisms, causing the feedback loop to grow exponentially in each iteration.

Now take this last idea and create some publish-subscribe patterns using multiple publishers and subscribers whom all feed on eachother. How would this affect processing of the service you are using? Could this be a web2.0 variant of the DDOS attack, one without the need for compromised machines? If one simple loop causes all sites to continuously use up some resources for all services, then maybe a complex loop can bring down the weakest service in the chain.

With the growing number of media feeds people can create and subscribe to, the chances that a feedback loop is created by accident is growing. People are also continually exploring the limits of social media... The only thing we don't know (yet) is who will be the first to create a loop.

Working on Paradigm Change Towards Web3.0

Recently, I've been asked to set up an architecture practice. This requires setting up the processes and information systems needed, expanding the software development process to properly facilitate architecture. Turns out that this does not take most effort. Something else does: Working to create a paradigm change to the Web3.0 way of thinking.
My current employer is very open-minded. In exploring the possibilities for an information system we've concluded that Semantic Web Technology can supply us with all tools to do the job. Some of the advantages beinig: small learning curve, cheap, easy to maintain, accessible through the entire enterprise, et cetera.
After creating the information model and setting up the tooling, I walked into an invisible wall: Paradigm mismatch. The people that have to work with the semantic tool need time to get their head around it. The concepts are new and different from anything they're used to. Same thing with architecture collegues. Some are critic, many are sceptical and only a handful of them understand the potential.
A lot of time goes into explaining what it is, what it does and how it works. And I guess this is a good thing. In order to convince the world (everyone) this is the way to go, some of us have to break the proverbial ice. Some have been doing so for over a decade. And I'm glad I'm contributing. Please feel free to join.

Putting a Semantic Wiki to work

The buzz is gradually getting louder: Terms like Web3.0, Semantic Web/Wiki and Linked Data are picked up more frequently every day. Discussions are shifting from the definition of terms to the application of them in the field. But where do we stand? Can we get this “Semantic Prophecy” into action today? The answer is surprisingly simple: Yes we can. And it’s quite easy too!

Why the need for a wiki?
The architecture department of my client needs an information structure. Some of the requirements are: instantly accessible for all roles in the development process, able to trace system properties back to their roots and able to expand to other areas in the development process (which prevents construction of an ivory tower).

After reviewing a number of architecture tools, none of them seemed to fit the bill: They specialize only in the field of architecture and/or does not have instantly accessible data.

Then I started to look at other options to answer our question: How can we structure our information in a flexible and accessible manner? After a bit of googling, I found that “Semantics” could help us out in form of a Semantic Wiki: Structured, Accessible, Adaptable.

Wikidsmart (zAgile)
There are increasingly more options for a Semantic Wiki. Semantic MediaWiki, Wikidsmart, OntoWiki, with more popping up every week. However, there was not much debate on which one to use. Confluence Wiki and Jira are already being used in the development process here. And since Wikidsmart is the only tool that integrates with both of them. It was not much of a choice.

Install, Use … and the ontology
Now all I had to do was install it on a test machine and prove that it works. And that’s all I did, really! No long story about failing a install, complex manuals, etc... It just works. With no prior knowlegde of Confluence, Wikidsmart, Ontologies, templates, etc... Within a week I had confluence running, with Wikismart with my own ontology on it.

Oh, yeah, the ontology. I almost forgot. It took about three weeks to create the ontology (the meta-model that structures the semantic wiki) with the architecture department. And we’re still working on it: It is the hardest part of it all. Get this wrong and the information you generate has neither head nor tail. It’s quite easy to get it wrong, too. If you get it right, however, you’ll be able to generate templates, pages, queries and views in no-time. Put them to work and your users can generate, modify and read the information they really need.

Nothing new, you say? The big plus here is that the relations between the pieces of information take the information itself (and the possibilities for using it) to a new level: Questions like “Which stakeholders are involved with system property X?” are answered by one single query. Glorious!

Conclusion
After reading this you’d probably say I’ve seen the light. And in a way I have: Semantic Wiki’s are one form in which our use of information will evolve.

But are we there yet? We’re very, very close. We are nearing the breakthrough of Semantic Wiki Tooling. Before year’s end, the first steps are taken in making Semantic Wikis a standard corporate tool. It is able to integrate information from the entire development process into one wiki. And we’re going to love using it. :-)