This is an old article written by Sabu Francis that explains the direction that TAD would take. It predicts the usage of the Internet for the spread of this seminal software ...
The Past
Architects uses drawings to represent their design before it is built. Using drawings (iconography) as a form of representation is just one among many methods that exist to represent a design problem. Few realise that point, however. They are so much surrounded by drawings that they do not get a chance to lift their heads and see that drawings as form of representation of architecture has been in use for only the last 250 years or so. That is hardly a time span when we know that the basic subject of Architecture has always existed, at least since the advent of mankind.
Drawings were always mired in confusion. Fortunately, the problems that came about were mutually resolved by social/professional protocols. In fact, much of Architectural practice consists of problem-solving that arise out of distorted communiciation between the specialists. Such techniques sit side by side along with the problem-solving of the design problem at hand. Sometimes they become more important than the real problem itself. It is only because these procedures have been forcefully integrated into our professional life that drawings as a form of representation has been accepted. The education system also constantly tries to ensure that drawings are looked at as a common ground. Often the training fails and the bare truth remains: There is nothing intrinsically right about using drawings as a form of representation of design.
The Present
What started as small misunderstandings while interpreting drawings has now become a large headache indeed. Communication methods has improved to such an extent that many more specialists are called in to solve a problem that used to be handled by fewer people. Today, the practice of architecture is so split up among specialists that there are just too many confusions due to cultural/temporal differences in understanding drawings and therefore the design problem. And, as any mathematician would know, the clear statement of a problem is itself an intrinsic part of the solution. It is therefore imperative that we search for a more efficient representation system.
With the use of computers, it is now possible to use a different method completely to represent architectural design problems. Computer science has now entered a more mature phase. It no longer requires to have direct analogies from the real world to write a compter program. Much of the early part of the computer software development had been spent in directly translating algorithms and procedures found in the real world. Thus, software categories like 'word-processing' or 'CAD' use easy real-life analogies.
There has been many advances in computer software: Object-oriented programming, A.I. studies, hypertext, etc. Many of the new nuggets of understanding do not really have a non-computer analogy. This has resulted in freeing computer science from picking up and clearing the mess from where the problems in the real world had last left off. It is now possible to emulate problems even at conceptual levels, rather than rely on just analogies.
Therefore, word-processing has moved away from being just an enhancement of the way we type and arrange words. Spreadsheets are no longer just quick 'n easy calculations. Now what has happened to CAD?
CAD as a part of the design process itself has not been scrutinized in detail. That is because a picture speaks a million words. There are just too many mathematical issues in the very act of putting a vector into a raster device like a computer. People involved in CAD has not really been able to tear themselves away and look at its role in the complete holistic view of designing.
When I came into this field, I was fortunate that I did not enter the arena as a CAD expert. I was not motivated to extend the work of other researchers. I did not even realise that such an approach was not the 'accepted' method. As one researcher had put it: "I know perfectly well that the representation problem in architecture is not like a "Eureka" discovery but more of a gradual solution (or solutions) that will be based on the work of many bright people working in so many places and for so many years on it."
I came into the field simply because I was directly facing the outcome of bad communication while practising as an architect in Navi Mumbai. The amount of furious construction that is going on in Navi Mumbai makes it almost like a battle field. Small errors are easily magnified. I had no choice but to look deeply into very basic aspects of representation systems, if I had to make any sense of my practice. I even spent time with conventional CAD systems, before discarding them. I therefore came into the field from the other end, rather than from the theoretical side. I needed a solution that grappled at the problem from its very roots, rather than rely on some higher level solution that started by accepting the premise that CAD started from and then tried to work on the problem.
After having got myself involved in the area for the last eleven years, almost involuntarily, I can now state that the field of representation problem in Architecture (and may be in other fields too) would benefit if occasionally someone could hit it on the side of the head. I am posing a very basic question here which may not beg 'gradual solution (or solutions)' but would indeed demand a necessary departure from conventional thinking, at least at the starting point. And the question is:
Can we always use immediate analogies as a form of design representation?
Representation systems imply a falsity and therefore at the very core, it must be abstract. If it is to be abstract, then it runs the risk of being considered eccentric. So how do we choose an appropriate representation system? My hypothesis is to choose...
- a) one that explains many of the end-effects of that part of the real world which is concerned with the field.
- b) one that can be boiled down to other forms of representation systems
- c) one that is all encompassing - which means the framework must be in place even before the thing being represented is discovered
- d) one that is graceful enough to be extended to more detailed representation systems
Atoms were proposed as the basic elements of matter. Such an 'atomic' system to represent matter is a good candidate that satisfied the above four criteria: It explained many end-effects of the real world. It explains why people had cruder representation systems earlier on ('ether', etc). The model was in place before one encountered the real thing. And the model got later extended into dividing an atom into 'protons', 'electrons' and 'neutrons' and so on. There are many other such valid, abstract representation systems in other fields too.
So the first step in computerising architectural representation must come as a pure abstract exercise and it is only later that it should be extended by "many bright people working in so many places and for so many years on it". Computers or mathematics by itself do not have that starting spark: A pen cannot start a novel. Such a hypothesis must start from the mire within the field itself.
I have developed a classification system in architecture that could form the very core of the representation of architecture. It divides architectural elements into 2 classes: 'spaces' and 'solids'. 'Spaces' are further subdivided to what are known as 'atoms', 'connectors' and 'envelopes'. The solids are sub-classed into 'artefacts', 'linked-nodes' and 'leftovers' (I am not explaining the classification system in detail here. A separate paper explains those 6 sub-classes in detail)
The above abstract classification system has fitted the four deciding criteria:
a) It is explains many end-effects found in architecture. For example, one can do climate analysis using this model and derive the energy requirements of a building. (That particular example is being explored in IIT, Bombay by Dr. J.K.Nayak and his team)
b) One can even derive current forms of architectural representation system (i.e. drawings) from it
c) It seems to be all encompassing, as I have yet to encounter an architectural scene that cannot be described using this classification system. It is not dependent on any 'scale' of architecture and it should be able to represent any architectual problem. It can be used as a GIS system used for planning and it can even go right down to explaining furniture design. As a practising architect, I have used it in the scale of architectural designing as found in Navi Mumbai, for over 2 million square foot of constructed work.
d) There is nothing in the model that prevents any extension. I have written a computer software called The Architect's Desktop ( or TAD for short) that is used to develop models that exploits this classification system. The software is capable of developing and extending more sub-classes, specifically as per the project which is being represented. Such extended classes still utilizes the basic classification system that I have developed, and further enhancements can be added to it without any hinderance. TAD took up much of six years to come to this stage. It is now extremely powerful and it even has a built-in object-oriented Prolog based A.I. language that I have named 'Ardela' (Architectural Design Language)
Today, the debate on whether such a classification system should exist is a moot point. It has been exposed to scrutiny for many years now and I have not really encountered any serious objection to it. Many of the buildings that I have designed using this system are now actually in use. The basic system is in place now. It just needs to be fleshed it out completely.
The Future
I can visualise a situation in the future that would consist of design professionals putting down their design as TAD files instead of drawings. Every piece of information concerning the field of architecture would be properly organised. Buildings themselves and the building material information would also be put in a format that integrated easily with TAD, for undistorted access. Instead of relying on standard reference books like Time Saver Standards or Neuferts, features would be added to TAD in order to access reference information immediately.
There are two areas where further work need to be done:
A) The classification system is to be recorded formally, along with all the objections and the answers thereto. I may not require a large team with me to do this. At the most, I may need some guidance from some expert who has formal knowledge in specifying classification treatises.
B) The second is refining the TAD computer software to do the following, among many others (no particular order):
- Work out details of TAD in various scales of architecture. For example; use it for GIS and planning
- Work out specific connections to the various specialities in architecture; like structural designing, plumbing and utilities, HVAC, etc.
- Enhance the geometric modelling capability of TAD
- Integrate TAD into other architectural practice areas like time and personnel management
- Improve production of output documents of various types, (for e.g., it may be based on the XML standards)
- Identify various expert system areas and improve Ardela, the built-in languge used within TAD accordingly
- Solve authentication and other security issues of TAD
- Add features into TAD to take advantage of new technologies like Virtual Reality systems, etc
- Cross-platform development of TAD
This work needs a proper team, with full fledged supporting infrastructure. The background material for this work is complete. In fact, much progress has already been made in many of those areas. All this while, I have been using the funds gathered from my Architectural practice to drive this part of the work. Now, as the work has reached the penultimate stage, I am seeking the formation of a full-fledged research team with access to the latest technology.
This can be done from within an academic institution, under an appropriate doctoral research heading. The disadvantage of a pure academic environment would be that the practice of architecture, which used to be one of the wheels that was driving the research, may be cut off. This may prove to be more of a hinderance than an asset.
Alternatively, I may be able to gather up such a team from the Internet. It allows both formality (as everything would simply have to be written down) and nothing prevents people in the team from the actual practice. This is the path that I'm pursuing to successfully conclude this project. The web-site for TAD at: http://www.teamtad.com would be the starting point for such a team.
View Sabu Francis's profile
