wtwolfe@hubcap.clemson.edu (Bill Wolfe) (11/12/89)
> Date: Fri, 10 Nov 89 22:17:03 MST > From: orville@lu.eas.asu.edu (Dr. Orville Weyrich) > Dear Bill: > Would you please be kind enough to post the following to usenet for me? > Thanks in advance, > Orville R. Weyrich, Jr. Subject: COBOL Poisons and Antidotes/Control Flow Nasties I have the dubious fortune of having to come up to speed quickly in COBOL for the purpose of producing software tools for COBOL, and have several questions to ask of the net. 1) A number of years ago there appeared an article by Brian T. Smith entitled "FORTRAN Poisons and Antidotes", which appeared in a Lecture Notes in Computer Science volume entitled "Portability of Numerical Software". This article/book discussed nasty things that could be done when coding FORTRAN and how to avoid/deal with them. CAN ANYONE RECOMMEND BOOKS/ ARTICLES OF A SIMILAR NATURE PERTAINING TO COBOL? 2) I need to analyze the control flow of COBOL programs (especially old unstructured code). This analysis is made difficult by COBOL's weak definition of the boundaries of a procedure which allow things like branching out of a procedure before the end specified in the PERFORM THRU, thus leaving an active return pointer to be blundered into during some other execution of the code segment. Also many old COBOL programs seem to havrecursive calls in them (which work due to the ingenuity of long forgotten debuggers in kludging fixes). CAN ANYONE RECOMMEND BOOKS/ARTICLES PERTAINING TO DETECTING/HANDLING/RESTRUCTURING THESE SORTS OF ANOMALIES? Good graph-theoretical treatments of the subject are most welcome. PLEASE EITHER POST RESPONSES TO COMP.LANG.MISC OR E-MAIL TO ONE OF THE FOLLOWING ADDRESSES: weyrich@csun1.cs.uga.edu orville@lu.eas.asu.edu I WILL SUMMARIZE TO COMP.LANG.MISC. PLEASE DO NOT LET THIS DEGENERATE INTO A DISCUSSION OF WHY COBOL IS GOOD/BAD OR WHY I SHOULD USE Ada OR C OR ....
nam2254@dsacg2.UUCP (Tom Ohmer) (11/17/89)
In article <7011@hubcap.clemson.edu>, wtwolfe@hubcap.clemson.edu (Bill Wolfe) writes: > > > From: orville@lu.eas.asu.edu (Dr. Orville Weyrich) > > Orville R. Weyrich, Jr. > > Subject: COBOL Poisons and Antidotes/Control Flow Nasties > > 2) I need to analyze the control flow of COBOL programs (especially old > unstructured code). This analysis is made difficult by COBOL's weak > definition of the boundaries of a procedure which allow things like > branching out of a procedure before the end specified in the > PERFORM THRU, thus leaving an active return pointer to be blundered into > during some other execution of the code segment. Also many old COBOL > programs seem to havrecursive calls in them (which work due to the > ingenuity of long forgotten debuggers in kludging fixes). CAN ANYONE > RECOMMEND BOOKS/ARTICLES PERTAINING TO DETECTING/HANDLING/RESTRUCTURING > THESE SORTS OF ANOMALIES? Good graph-theoretical treatments of the subject > are most welcome. SIZE 3X Disclaimer in use. I am aware of a product called RECODER that gives the following reports on a COBOL program: 1. Warnings Report 2. Genealogy of Names Report 3. Dead Code Report 4. Structure Chart 5. Modularity Report RECODER does Input Program Preparation, Structuring Process, Option Dependent Code Customization, and Structured Output Generation. RECODER is by Language Technology, Inc 27 Congress Street Salem, MA 01970 For the record, this is a rumor. Off the record, you might want to check it out. -- Tom Ohmer @ Defense Logistics Agency Systems Automation Center, DSAC-AMB, Bldg. 27-6, P.O. Box 1605, Columbus, OH 43216-5002 UUCP: ...{seismo!osu-cis}!dsacg2!tohmer INTERNET: tohmer@dsacg2.dla.mil Phone: (614) 238-9210 AUTOVON: 850-9210 Disclaimer claimed