Dear Mr. Karasik,

My name is Truyen T. Tran from Vietnam. I am a relatively newcomer to TRIZ.

I really like what you put in the Anti-TRIZ-Journal. I’ve read many of your criticism of the TRIZ-Journal, and I do agree that it is very misleading and I quitted reading it already for a long time. I feel that TRIZ (as far as I know it from what was written about it in English) today may differ significantly from what it was first proposed to be. People may have interpreted it differently and probably wrongly and then spread TRIZ to the world.

However, I am very interested in the way TRIZ pushes forward. I am using it to advance my thinking in my research of computing. Here I have some thoughts to share with you and to seek your clarification if I misunderstood the matter.

First about TRIZ’s informality, I would appreciate if TRIZ was described in more formal way, i.e. using mathematical formulations to avoid ambiguity. That is how science is based on. But to me until now, a significant part of TRIZ is rather ambiguous and error-prone. Some of them have been pointed out already, for example, Trends of System Evolutions. Also, I don’t know why people keep using the notion of ideality as a function of {harmfulness and usefulness} because it makes no sense of using mathematics representation without any operators applicable to it.

Another issue is contradiction. I do feel good about the pattern of resolving contradictions as a way toward ideality. But on the road from Administrative Contradiction, to Technical Contradiction and finally ending up at Physical Contradiction, practitioner needs to follow their intuition to translate the pair of {improving, worsening} to the extremum {A, NOT A} at the SAME time (delta time = 0}. My intuition is that the former pair can be translated to many different latter pairs, then we may getting the wrong road to solve the real problem. In addition, the translation hardly helps to point out any way to get out of the stuck.

May be it is my own problem because I am working in the field of software, something that is already ideal according to TRIZ (costs nothing, no materials, very fast, etc.). But I am inspired by the way TRIZ was derived (I guess you participated in its emergence, didn’t you?), and its ideas of inventive thinking. That is everything can be grouped in some much smaller subset of patterns. I use the term “pattern” for anything repeated in time or space (including the idea space).

My question then is how was TRIZ created? Of course the obvious answer is by using analysis. But that is not what I am expected. Could you please introduce any methodology used to analyze that huge body of knowledge (some sources say TRIZ creators had dug millions of patents)?

I guess the philosophy behind TRIZ is based on dialectics and idealism (that’s how Marxism existed in the former Soviet Union). I suspect that originally TRIZ was emerged mainly from mechanical systems, and those relating directly to classical physics laws (Newton mechanics, thermal dynamics, electromagnetism and optics). What if quantum mechanics and relativity play the role here?

This may explain why Contradiction Matrix is hardly applicable to computing, since we don’t have such features in computing. On the contrary, many inventive principles are highly applicable (as you have successfully shown), not only in technical systems. The first indicates that we can see and measure directly from a physical systems (with some exceptions) but the second relates to what we can “play” on such systems. The second, I think, closely relates to abstract models of real systems in our mind. Actions like division, merging, extraction, aggregation and nested-doll are manageable in human thought.

I then arrive at a “pattern” of creating TRIZ: that we need to divide the subject into a set of “attributes” (similar to the concept of representation in computing) and a set of “actions” (like operators in computing) to play on the former set. Interestingly, this is exactly all about computation {= data representations + algorithmic operators)!

Since computing is a direct application of mathematical logics & symbolic manipulation, I guess in order to use it in this field, we have to re-build TRIZ from the start. I strongly agree with TRIZ underlying philosophy, as I listed above, contradictions and idealism. As contradiction is an inherent property of the nature and thought, I see no problems in using this principle. However, TRIZ idealism seems not to play a very important role in computing, especially those are not directly related to hardware or physical objects. According to TRIZ, software is what we can expect of ideality. A way around this is to re-define ideality, may be in term of algorithmic complexity, entropy increase for running a program, or so.

One objection to above statement is the execution of CPU, which is the heart of computational process. CPU is a physical system, which is directly constrained by physics laws. Second, in current model of computing, RAM also plays an important role and makes Information Theory a field of study. I am not qualified to comment on these, but I keep my focus on higher abstraction level where software lies.

That is what I am thinking of TRIZ and computing. Currently I am looking widely for a chance to:
(1) apply TRIZ framework directly to computing,

(2) re-form TRIZ from ground up, or
(3) combine those two

If we follow what Altshuller and colleagues had been doing, we need to examine the vast numbers of computing patents and research papers. This is too much work, so I guess we have to find a shorter way.

In (1) we can do the following:

(2) probably leads to a new formation, similar to but essentially not TRIZ. I still don’t know yet. At large sense, TRIZ can be considered as a meta-language, like UML, for specifying, representing, documenting and solving inventive problems. I say meta-language because it is the language of language (like ARIZ).

I just have some immature ideas to share with you. Please feel free to comment on anything I give here.

Best regards,

Truyen T. Tran,
Research Institute of Posts and Telecom,
122 Hoang Quoc Viet, Cau Giay, Hanoi,


Dear Mr. Tran,

You are raising interesting questions.I am unprepared to comment on all of them. Let us address them to our readership at large. May be they will join the discussion.

On my part, I agree with you that the notion of the Ideal Final Result (IFR) is hardly applicable to software systems without a modification. A possible modification could be this: The ideal computer program is such a program which contains no code, occupies no memory, etc.

I also agree with you that for making TRIZ applicable to sofwtare, it has to be rebuilt from scratch. Although some of its elements are applicable (such as logical contradictions), they do not play as important role in software as in other fields of engineering. There are elements which are much more important for software engineering and which were missed in TRIZ.

As for the methodology of creating the TRIZ like theories, I do not know one except for intuitive collecting of useful examples, intuitive grouping of them, etc. The role of the sheer brute force in such a research (say, analysis of millions of patents) was largely exaggerated. It is the same as to say that a chess grand master analyses millions of combinations to find the winning one. He does not do so. The grand master has a talent to see the promising combinations amongst the billions of all other, without applying the brute force analysis. A TRIZ researcher has to have a similar talent.

Best regards,

Yevgeny B. Karasik

P.S. The readers are invited to join the discussion of the issues raised by Truyen T. Tran.