by mwilliams » Fri Mar 29, 2013 6:54 am
Well,
this is interesting. Though, I totally, agree that “assembler has too many quirks to make it automatically translatable to any other language”
I am well aware, that there are probably a number of tools which claim to convert assembler into Cobol. Although, I had not personally used any of these tools, I can’t see the translation from Assembler into Cobol as single iterative process.
As with my experience, of needing new Cobol programs rewritten from existing assembler code, I though about how tedious of the task of doing it by hand. So, I pondered over ways of reducing the manual labor involved. Overtime, after upon recognizing some common patterns of how the code was written, as a result, I had written a utility that reads a particular output file from the assembler which then build partial Cobol programs based upon a common coding style. I had gotten as far as deciphering assembly language defined storage into Cobol’s working storage, and creating linkage parameters (to some degree) based upon the usage of register 1. But in all fairness, upon all the work put into this utility, it only reduced my efforts by a third of the time than if I had written the code totally by hand. Inclusion, I believe, the translation of assembler into Cobol is a multiple iterative process, which involves research, planning, and design.