Tong-Gao Tang:
Tong-Gao Tang:
(photo 2006),
(News Report 2013),
(photo 2017)
(1) industrial mathematical modeling application (two years since 1967) -- great achievement.
(2) real time control software + electronics + computer hardware design & manufacture (eleven years since 1969) -- normal level.
(3) logics + theoretical computer science + software & hardware verification + Software Symbolic Compiler (twelve years since 1980) -- great achievement.
(4) artificial intelligence (three years: 1989,1993,1994) -- higher level.
(5) distributed business application software + software engineering technology (since 1995). great achievement.
He has never discarded but made unremitting efforts to solve any encountered difficult problems in his projects,
until to reach the world best results even some projects had token him eight years more without financial support.
- He
graduated from a mathematics department, published his thesis and to be a teacher
taught a “Complex Analysis” course in the same mathematics department
in 1963.
- He
has been a programmer in last thirty nine years since
1967.
- He
solved a (little long time) difficult problem in a “computerized ship-body-Line
lofting project” in 1969. He had used his independently created method
“Least Squire Plus Right and Right Calculation”. A similar famous method
“... and Right Leaning” was called a Machine Learning in Artificial
Intelligence around 1966-1967.
- He
received inspiration from the powerful computer role in above “computerized
ship-body lofting project”. He wrote two propsals to mathematics department leaders for setting up a “Computer
Manufactory”. Later, he was allowed to build a computer manufactory in 1968-. He was the leader and a designer of
the new computer manufactory of Fudan University (design, manufacture, sale, application all in one) in ten years since 1969.
-
He with his colleagues set up a “Computer Structure” major in Fudan University in
1970. He was the director of the "computer structure"
teaching and research group over five years.
- He
became a logician and theoretical computer scientist after leaving
his computer factory in 1980. He published about “Formal Semantics of Program”
papers since 1983. He taught "Program Semantics" and
"Computational Theory" in a computer science department.
- As
a logician, he was invited to be a mathematics reviewer of
Mathematical Reviews for theoretical computer science issue in 1985.
- He
was a visiting scientist working in “Software Verification”,
“Hardware Verification” and “Artificial Intelligence”
-- three projects in CMU in 1987-1989.
- He
was as a logician working in Australian universities for “Theorem
Proof”, “Knowledge Representation and Reasoning”, and
“Machine Learning” in 1989-1995. He with Prof. Norman Foo
established the second step algorithm of a famous two steps well-known “Machine Learning” theory for ‘getting a set’
wrong in 1994.
- He
was a compiler expert especially for revised programming languages which may
input algebraic formula as well input digital data since 1987. He created a
a “Program Induction Method” added in a compiler
, and add some Temporal Logic formulas as new statements in programming
languages. So that, a compiled executable program may verify programs simply and
efficient, This verification technique can be used by normal programmers themselves. He thinks that it is the best way to verify programs in practice.
- He
created a new compiler technology for PROLOG program as follows:
His project needed a compiler. So, he planed to buy a Compiler for his PROLOG programs
from an university, but the seller said: “you not only pay this 'Compiler' fee
but later any your made products, which was compiled by this Compiler, should also pay
fee”.
He thought it is unfair. He did not buy it. He spent four months to create a special auto-generated PROLOG compiler that may compile his prolog programs to be C++ or machine codes. The compiled codes can be “executed very quick”.
Around 1996, He made a demonstration of this new PROLOG compiler technology for Prof. Edmund Clarke in CMU. Edmund had asked three questions to him (1) “Has your compiler used the Warren Machine ?”, he answered: “Needn’t”; (2) “Has your compiler used Garbage Collection algorithm ?”, he answered: “Needn’t”;
(3) “What is you mathematical Unification Algorithm in your compiler ?”, he answered: “No mathematical Unification Algorithm needed, I just use a simple ‘one date points to another data?to instead of the complex Unification Algorithm”. Suddenly, Edmund stood up and said: “Yes, you are right”.
The above "special auto-generated" idea has been developed as his new pattern meta logic and software 'model-to-code' technolgy .
- He left computer science research and started moving his theoretical knowledge and methods into Software
Engineering field as he had studied Unix distributing technology and Microsoft window technology. He gave a proposal
“Program C++ Code Auto-Generator” for application software in Unix Tuxedo distribute system to his company's leaders when he was working in FDC in 1997.
-
He thought "it was time now to launch
a software engineering revolution based on his knowledge". Since 2001,
he started to design and develop his Visual D++ language and SDDA system in
ten years (from 2001 to 2010), day and night, in order to realize his above proposal.
Also, before opening to public on network, he had arranged one-year
demonstrations to many people.
-
Current, he is the founder of the new small business "Software Construction Tech." since 2004.