Re: Dialects of Cobol ?

You may be able to get substance in COBOL dialects.


COBOL is remarkable for how standard it is generally. But even before
the advent of PCs there were some differences offered as 'extensions'
by vendors, and folks haplessly coded a lot of these in production
code, and there may atleast be brief examples at vendor sites.


Some mainframe vendors offer meaningful example suites, and with the
correct inquiry you may qualify for something off an older tape or
download that has elements deprecated or even nolonger supported in
new compiler releases.


With the advent of the PCs a number of special capabilities were
evolved by compiler writers to get things first to the CONSOLE (the
stdin/stdout) and then along the line to windows APIs. It should be
possible to get examples of PC oriented code that just will not
compiler on a mainframe, effectively a dialect.


More recently, the COBOL standards committe evolved an object oriented
definition for the language. This took time, and some vendors went to
market with early variations involving features that did not make it
to the final standard. There could be some code available here that
constitutes a dialect distinct from the processable syntax that
becomes the 2002 standard. (An esoteric exable is the CORBA object
attributes: automatic set/get methods, for example).


Also the entire object oriented feature set can be a major dialectical
distinction if you want to look at it that way. There should be a
growing body of different code then in the object oritented form
(these are major differences).


Also of note is the transition from the oldest COBOL that did not
allow _imbedded_ _programs_ to the COBOL that does, circa 1985.


At about the time of the 1985 change all manner of explicit scope
terminators arrived in the language. The before and after of this
transition is perhaps ponderable as a dialectical distinction.


Of possible interest to you is the drop of the support for REPORT
WRITER at about the 1985 transition. Major sections of code nolonger
compiled after this change.


Much more specialized is the deep issues in the syntax on associative
and commutative properties of the NOT operator in the lexical stuff
called COMBINED ABBREVIATED CONDITIONAL RELATIONS in COBOL. Same code
compiles differently before and after the standards change (just a few
gotchas). In the same way, the OCCURS DEPENDING ON clause had
different semantics before and after the 1985 transition (for some
statements).


In general, there is a lot of COBOL code available!


See, also the test suite at


http://www.itl.nist.gov/div897/ctg/cobol_form.htm


There may be both older material available there from prior version of
test suites, and it appears to me that they may need to evolve some
newer test suites for the object oriented syntax.


Best Wishes,
Bob Rayhawk
RKRayhawk@aol.com

Return to the comp.compilers page.
Search the comp.compilers archives again.