Re: Only Sham SW Engineering Without System Approach and Scientific Method
From: Richard Riehle (adaworks_at_earthlink.net)
Date: 03/21/04
- Next message: H. S. Lahman: "Re: xp and simple design"
- Previous message: Nick Landsberg: "Re: Code Coverage and QC"
- In reply to: Tsolak Petrosian: "Re: Only Sham SW Engineering Without System Approach and Scientific Method"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sun, 21 Mar 2004 18:08:50 GMT
"Tsolak Petrosian" <tsolakp@yahoo.com> wrote in message
news:a052897d.0403041931.11545a6e@posting.google.com...
> Thats called iterative development, and code pretty much covers both
> "specs" and "documentation".
One reason I like to use Ada for important projects is that both the
Specs and the Implementation are representable in code. The Specs
(Ada specifications) are compilable. The corresponding implementation
must conform to the specification because the compiler ensures that
this is so.
Further, one can compile the specifications long before there is any
implementing code and check that all the pieces fit together as they
should. This is a low-level architectural issue, but it is quite handy
for knowing about the integrity of one's architecture prior to creating
implementations.
Development can proceed incrementally. There is no need to implement
any code that does not meet the current priorities, even though some of
it may be specified in the Ada package specification.
Ada provides two approaches to extensible programming, inheritance
and child library units. Inheritance has always been a heavyweight
approach to extensibility. Child library units are more of a lightweight
approach that involves much less overhead in the final executable
code.
I suppose my point is that, when one chooses the right tools for creating
software, there need not be a gigantic disconnect between specification
and implementation. With C++, Java, and many other languages, I see
too many developers succumbing to the temptation to write implementation
code before they have fully validated the integrity of their specifications.
This seems to lead to more errors than it they had done it differently.
Richard Riehle
- Next message: H. S. Lahman: "Re: xp and simple design"
- Previous message: Nick Landsberg: "Re: Code Coverage and QC"
- In reply to: Tsolak Petrosian: "Re: Only Sham SW Engineering Without System Approach and Scientific Method"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
