n8243274@gonzo.cc.wwu.edu (steven l. odegard) (06/20/91)
I am researching the feasibility of generating a modula-3 compiler for 8088 native code. The main reason for this is to gain programming control over my own personal XT turbo. This will be a summer project which I would release for a nominal fee on the Shareware concept. However, on researching the feasibility of this, I came across the book describing program verification as a development methodology. (Sorry I don't have the title to the book yet). Essentially, the specification is machine-checked for inconsistencies, and the code is deveoloped to the specification and mechanically verified to prove consistency. Realizing this technique, I have temporairly abandoned the original project to explore the possibilities of incorperating these techniques in the specification to modula-3 that I will employ. I love the pragma syntax! I believe extended pragmas will work nicely to implement a version of the stanford verifier. I am also interested in incorporating all of this into a CASE (Computer Assisted Software Engineering) implementation. I believe that a well-developed machine-assisted programming methodology can be implemented without a graphical user interface, and result in a four to eightfold productivity increase. (My personal experience was a threefold increase of productivity after learning and regularly employing the module interfacing technique of Modula-2.) If anyone wishes to provide suggestions or direct me to current research on these issues, I would be most greatful. Please post mail to n8243274@henson.cc.wwu.edu. --Steven Lee Odegard PS Is there an effort to post the march91, april91, and may91 news summaries to comp.lang.modula3 in gatekeeper.dec.com?