Though we are basically an architects firm, we consider ourselves a knowledge organization in the true sense. Our research work starts from the representation of architecture onwards. We believe that representation theory is critical in any field. After all, if a problem is incorrectly represented in the first place, then any solution that may emerge from such a representation would often be delusional.

The entire "food chain" of knowledge starts with raw data. That changes to become structured information and thence to knowledge and finally to wisdom. The last (Wisdom) is attainable only in the minds of some human beings and therefore cannot really be codified. So is the case with knowledge: Pure knowledge can exist only within the mind of people, albeit a bit more than those who go around with wisdom.

However, part of information can be massaged to what can be called proto-knowledge and this report concentrates on that. I shall be using the term "knowledge" to mean this proto-knowledge to reduce the verbiage of this report.

History of knowledge management in architecture
Architecture is considered to be the mother of all arts. This can be extend that to state that it is in fact the mother of all arts and sciences. After all, everything is couched within space. The context of managed space can never be removed from anyone's lives. Because of that unique aspect, handling knowledge in architecture is quite difficult. A building can be looked upon as a complex assembly of information components. The "What" needs to be put together along with the "how" and the "where". All that needs wisdom. As architecture is one of the oldest profession known to humans, this issue is not new. However, the reason why it was not critical till now needs to be understood from a historical perspective .

The information age is the age of blurred boundaries. The cultural and intellectual framework of people is far more homogeneous today than during the times of our ancestors. Today, it is not considered strange when people from different countries (with supposedly different ancestral background) together discuss critical issues facing mankind. The whole world is now one village. We form our friends via emails, Instant Messengers and websites.

That is not the way it used to be: Long time back, architecture was fairly fragmented and amorphous. The cultural context of India was way different from that in Greece, and so on. The knowledge management problems that each coterie of people faced could then be solved by sheer interaction of people participating in various face-to-face discussions. Whenever people faced complexities, it was assumed that the simple technique of "divide and conquer" would work. Thus as society progressed and started facing more and more complex issues, society responded to knowledge issues by making more and more specializations. The roots of that methodology can be traced to Aristotelean philosophy from Greece.

The last century has seen many momentous events like two world wars, and the birth of the Internet, etc.. that has put humans all together into one place facing more or less the same set of issues. The analytical approach of trying to handle complexities using divide-and-rule approach of the specialists have started becoming invalid. It is now important to see things holistically, because often a mismanaged whole would throw up more complicated problems later on. At the same time, the earlier method of thrashing things out face-to-face is no longer possible because of sheer distances involved and issues dealing with concurrency. (Somebody may be forced to take a decision based on someone else's work which was published earlier, etc. i.e. the participants are not interacting at the same slot of time. E.g. Use of emails.) This means, society sorely needs theories and tools that will aid the holistic assessment of knowledge. Whereas earlier we could have lived with flaws in the representation of problems (because they would get all smoothened out by sheer human power), today that is not so. Representation theory has become quite critical for knowledge management.

Knowledge for proactive actions
Initial attempts at knowledge management was nothing much more than sophisticated document management. People now have realized that not only do they need to move out of document management but they need to quickly connect to the context that they are caught up in. Much of the context consists of "verbs" (or actionable issues) rather than "nouns". This can be understood by an example:

When an architect visits a construction site and finds that certain finishing material has not reached on time, he should ideally take out his WiFi enabled palm computer and connect to the Internet and invoke a program that will assess the gravity of the situation, and not only suggest alternative but also allow him to act (that is where the "verb" comes in) on the decision by help him order a new material. Because of the high costs that are involved, an architect may not have enough time on his hand to read a document (which is only a "noun")

And even if he does, he may not have the other decision making capabilities in his hand to ensure that the "noun" can be used for a valid "verb" (i.e. action)

KM Issues faced in architecture can now be seen in other fields too. Any complex field (e.g. Medicine) that has an underlying bed of highly dynamic context needs to be handled the same way that we've described for architecture. In such fields, there were attempts by AI (Artificial Intelligence) gurus to develop expert systems that laid out the actions that people were to do in quite some detail.

However, Mr. Sabu Francis experience with expert systems was that they were quite flawed because they rarely dealt with the underlying changing context. Our assessment is that KM need not go all the way to dictating all the verbs as done by the earlier kind of expert systems. At the same time, some decision making capabilities to undertake actions can be quite effective. The catch would be to figure out where to draw the line.

Technology for knowledge management
It is no longer efficient for standard database based software designs to handle KM problems. Databases were invented in the early days of the computer revolution and it had issues similar to the "divide-and-conquer" problems that I spoke about. For example; the designing of a database table with its set of fields is an act of "divide-and-conquer". Once you classify data into the fields of a database, then to re-classify it using some other strategy usually becomes a thankless exercise once the data is large.

KM processes are today being aided by processing meta-data rather than just data. The concept of XML was born only due to such a need. Several technologies such as web services, SOAP, RSS support this point of view.

Nobody used Bayesian networks to perform data mining. The Ranganathan facet based classification was known only as an librarian's esoteric theory. Today, all of them are being actively used to develop new tools to help KM activities that I've mentioned earlier.

Thankfully, some old AI languages such as Prolog and LISP which had envisaged the need to handle meta-data is now back in the picture. In fact, Prolog has an excellent connection with XML and I have worked in that area for many years now.

Conclusion: Research work at SFA
In todays information world, consisting of blurring of boundaries between specializations; it is important to have the right tools that handle the development of knowledge rather than the handling of mere documents. Meta-data processing using XML and/or AI languages seems to be the right direction. All that kind of research work is simultaneously being undertaken by SFA along with regular architectural practice and other services that is enumerated on this website.

All the software products that SFA has now made available are bye-products of this research work

• Add new comment