At 11:43 AM 10/8/99 -0500, Ron Finkbine wrote:
>I think that changing languages is futile. We must change to some form of
>case tool and auto-code generation. Change and raise the level of the
>representation.
Isn't this what formal methods are intended to do - to specify a system so
that code can be automatically generated? Anyone have information on the
learning curve of formal methods versus conventional languages? I can't
help suspecting that going down this route actually restricts the
capabilities of the programmer. For instance: if we specify a system for
a code generator, we must be certain that we can a) specify exactly what we
want with no possibility of misinterpretation and b) that we will get
perfect code according to what we put into the generator. Neither of these
seems especially likely, particularly given the potential for differences
in understanding of the problem between developers and clients. Compare
this to a software development process along more conventional lines, where
the conceptual gap between the implementation and the design seems to be
shorter than if we go along the high-level path. Making changes to a
well-designed piece of software, and reusing code would both seem to be
easier in the conventional approach. Are formal methods used in highly
interactive software, or primarily in the control of well-defined processes?
Pete
------------------------------------------------------------------------
Peter Hornsby,
Department of Computer Science
Loughborough University,
Loughborough,
Leicestershire,UK EMAIL:[EMAIL PROTECTED]
LE11 3TU. Tel: +44 (0)1509 222799
Fax: +44 (0)1509 211586
------------------------------------------------------------------------
