On 20 Aug 01, at 22:06, Morris, Steve wrote:
Maybe. Really I'm just interested in what exactly a theoretical basis for software engineering would look like.
well try to make a general theory of communication which covers all possible communication, speech, evocation, persuasion, poetry, art and expression.
But the latter is being worked upon and indeed shares much of its basis with CS (Shannon and Moore both worked in comms) but whether such a goal can be achieved, at least they are trying. But in CS we seem to be bogged down in the study of techniques for programming rather than the fundamentals of understanding the nature of the information we manipulate. For example using Shannons work I can define the information content of a bit stream, by applying similar principles from Huffman I can give meaning to a character stream and McCabe(?) applied similar concepts(albeit via a bizarre mapping) to software complexity metrics. It would seem to indicate a basic commonality stretching from bitstreams through to program code. If we could start to understand the nature of that commonality we could start to address issues of program quality and correctness based on the impact on the underlying information transforms. Indeed we would also be able to start to specify systems in terms of the underlying information requirements. At present we are a long way away from this. (And no, I do not believe that data driven design ala JSD etc is directly related to this. Anymore than most of the work in AI does.)
software engineering would look like." Engineering doesn't have theories. Science has theories; engineers apply these theories to solve specific problems.
This is true and I probably should be using CS instead of SE. The IEEE papers were headed SE so I guess that steered my wording.
Without specific questions you will never have useful theories. When you narrow your self to specific questions you will likely find plenty of literature that discuss them.
And indeed that's been the direction of CS for the past 50 years or so. But the point the paper's authors were making was that we can no longer go on far in that direction without deriving a better understanding of the nature of information.
The results suggested by these theories, once validated as useful and predictive, can be the basis for the next level of theories.
And that seems to be the nub. We appear to have reached the point that we are not uncovering much that is new, we need a new set of theories to advance the art.
Physics is too general a thing to have one theory.
I'm not sure I agree. All of physics ultimately comes down to studying the interaction of particles and the forces between such. Kind of like layer 1 in my model. At layer 2 things start to diverge into specialisms And by layer 3 we reach the differences between say optics and electronics - based on the same electromagnetic wavre theories but applied in different areas.
Physicists have been hoping for this one for 80 years.
But again they are trying. All I'm asking is who is trying to understand information in the CS community? Thanks to everyone for the comments so far. It's interesting to see the different approaches taken. Alan G