[SCC_Active_Members] FW: Doug Ross - Obit

[Van Snyder]

On Sat, 2007-02-10 at 16:58 -0800, Ike Nassi wrote:
> While studying and later teaching at MIT in the 1950s, he directed
> development of APT, the computer language used for automatic
> programming of numerically controlled machine tools. He came up with
> the term computer-aided design, led MIT's CAD project through the
> 1960s, and helped develop the software engineering language AED. Mr.
> Ross also developed the Structured Analysis and Design Technique and
> founded SofTech to market products based on the system. 

He also wrote a paper I believe to be significant, and under-
appreciated: "Uniform referents: An essential property for a software
engineering language," Software Engineering 1 (1969) pp 91-101 (J. T.
Tou, Ed.).

David Parnas ("On the principles for dividing programs into modules")
advised to hide the details of implementation of data structures inside
of procedures, so that accesses all looked the same (procedure
references), and therefore one could change the representation of a data
structure without having to change all the references.

Ross, and later Geschke and Mitchell ("On the problem of uniform
references to data structures") advised to hide the representations in
plain sight, by designing the language so that different data objects
(and procedures) have the same syntax of reference.  I.e., using
Person.Name is a bad idea, while Name(Person) is a good idea, especially
if the language provides updaters, and Array[10] is a bad idea while
Array(10) is a good idea, again especially if the language provides
updaters.  If ones ambitions for the program's features increases, one
can replace some or all of the fields of the structure Person, e.g.
Name, by updaters without changing the syntax of reference -- which
(again) is what Parnas advocated in a different (and less efficient)
way.

Because we followed the advice of Parnas instead of that of Ross,
Geschke and Mitchell, we have enormous programs composed of numerous
tiny trivial procedures.  A compiler that can inline such is a rare
bird, so now we need gigabytes in our processors.

-- 
Van Snyder                    |  What fraction of Americans believe 
Van.Snyder at jpl.nasa.gov       |  Wrestling is real and NASA is fake?
Any alleged opinions are my own and have not been approved or
disapproved by JPL, CalTech, NASA, the President, or anybody else.