Currie Colket wrote:

> Another important milestone has occurred in the life of ASIS.  Back in
> 1992, Al Kopp and Jim Bladen, both of TeleSoft, speculated that ASIS
> could be used to reconstruct the source code of an entire Ada partition
> simply by making ASIS queries into the compilation environment. This
> speculation has now turned into reality!!!
> Alfred Strohmeier, Vasiliy Fofanov, Sergey Rybin and Stephane Barbey
> will be presenting a paper at Ada-Europe'98 (see
> titled: "Quality-for-ASIS: A Portable
> Testing Facility for ASIS." The paper describes the development of an
> extensive testing facility for ASIS implementations. Of particular
> interest is the approach taken for the testing of the generic procedure
> Traverse_Element. The strategy is to compile the ACVC test suite, which
> contains all possible Ada 95 syntactic elements, and reconstruct the
> ACVC using only ASIS queries. The paper reported: "The implementation of
> Traverse_Element in ASIS-for-GNAT passed successfully this test."
> This is very important achievement as it is a clear indicator of the
> maturity of ASIS. This test demonstrates that ASIS implementations can
> correctly handle every Ada 95 syntactic element, as evidenced by
> reconstructing the ACVC using ASIS queries.

I'd like to underscore the importance of the above demonstration by
pointing out that it confirms *two* significant milestones:

1) In general, as Currie indicates, it shows the completeness of ASIS
itself, proving that no Ada 95 syntactic element has been overlooked
in the specification.

2) In particular, it shows that at least one specific implementation
(ASIS-for-GNAT) can pass this aspect of the ACVC suite, thus setting a
precedent for other ASIS implementations to follow.

Regarding the latter: It has long been a concern of mine that, while
compilers provide a level of confidence via the ACVC suite, *other* Ada
tools do not (e.g., for code analysis, documentation generation, coding
standards support, metrics, code browsing, reverse engineering, etc);
consequently, the quality of these tools has often been uneven. My hope
is that ASIS implementations (upon which such tools can be built) will
find it necessary to advertise their success with demonstrations like
the one above, thereby raising the level of integrity in Ada tools of
the future.


C. Daniel Cooper ==========v=======================.
Adv Computing Technologist | All opinions are mine |
206-655-3519               | and may not represent |
[log in to unmask]  | those of my employer. |
The question is not "What is the answer?"; rather, |
the question is "What is the question?" --Poincare |