Cobol - Syntax (another approach/thread)

Author

Syntax (another approach/thread)

William M. Klein

2004-03-29, 8:30 pm

I thought that I would start another thread on the whole issue of "syntax".
Maybe this approach will help clarify MY position (and I think that of some
others).

The underlying problem seems to be that some people would LIKE for the COBOL
Standard to specify when source code is "guaranteed" to compile without any
syntax errors - if compiled by a "conforming compiler" with its "conformance
options" turned on. And/or they would like the Standard to "define" what may
and may not be issued as a "syntax error".

I do NOT think you are ever going to get anyone who has worked closely with the
current (2002) or recent (1985) Standards to agree to this - no matter how you
word it. (and even more unlikely to find any COBOL vendor to agree to this.) If
anyone following these threads can actually find some place in the 2002
Standard (which cleared some of this up) that actually SAYS that, I would be
very interested in knowing where it is.

The current (and recent) Standards *DO* tell you when source code is
"guaranteed" to get an ERROR or an extension flagging message (if that is turned
on) when compiled with a conforming compiler. (Although even this has "slight"
loop-hole in the area of "undocumented features" that are in the compiler but
not officially supported by the vendor. These may compile, get no extension
message and still violate the Standard. However, this only occurs when the
feature is NOT supported by the vendor and not in their documentation as an
extension.)

The current (and recent) Standards simply place ZERO constraints on what
ADDITIONAL "syntax errors" (or other types of errors or messages) the compiler
vendor may place on their CONFORMING implementation. (What the implementor CALL
the "messages" is UP TO THE VENDOR.) As far as I can tell, the implementor may
document that source code must have an even number of lines of source code and
if not a SYNTAX error will occur. The vendor may state that their compiler will
only compile cleanly on Monday, Wednesdays, and Fridays, and issue a compiler
SYNTAX error if you try and compile on other days of the w

. I don't see any
(successful) compiler vendor doing either of these, but there is simply NOTHING
in the Standard that prohibits this.

Now, this may not "match" the common usage of the words "syntax error" and it
certainly may not PLEASE all the readers of C.L.C. - but I simply find nothing
in the current (2002) COBOL Standard to prohibit vendors implementing a COBOL
compiler as "conforming" and placing WHATEVER constraints (called "syntax
errors" when detected at compile-time) on programs they wish and being a
conforming compiler.

--
Bill Klein
wmklein <at> ix.netcom.com

docdwarf@panix.com

2004-04-03, 8:30 am

In article <96AqC-h9flB@jpberlin-l.willms.jpberlin.de>,
Lueko Willms <l.willms@jpberlin.de> wrote:

[snip]

> I beg all participants, especially DocDwarf, to stop nitpicking and
>quarelling about nothing.

Quite right... let's get back to discussing Important Things, like
paragraphs versus sections, when a GO TO is permitted and paragraph
numbers are unnecessary because one will never have to modify a 30-yr-old
check-writing program.

Some say 'the devil is in the details'... others, like van der Rohe, have
it that 'God is in the details'. In either case the battle might be seen
as theological, sure, but taking a side, appropriately, becomes a
moderately curious... detail.

DD

Lueko Willms

2004-04-09, 7:30 pm

Am 30.03.04
schrieb wmklein@nospam.netcom.com (William M. Klein)
auf /COMP/LANG/COBOL
in J73ac.7473$lt2.7325@newsread1.news.pas.earthlink.net
ueber Syntax (another approach/thread)

WMK> The underlying problem seems to be that some people would LIKE for
WMK> the COBOL Standard to specify when source code is "guaranteed" to
WMK> compile without any syntax errors - if compiled by a "conforming
WMK> compiler" with its "conformance options" turned on. And/or they
WMK> would like the Standard to "define" what may and may not be issued
WMK> as a "syntax error".

I think it is simply due to the fact that DocDwarf's COBOL compiler
simply flags an error by writing "SYNTAC ERROR!" instead of explaining
it.

When I was working with COBOL, I had compilers which would simply
differentiate between different levels of severity of the error,
issuing "W" as "Warning", "E" as "Error", and always pointing to what
the compiler identified as the erroneous statement.

The word "Syntax error" was not mentioned by the compiler, but to me
it signified that I has written a syntactically wrong statement, and
had to change my program, if I would like to have an operational
program to be used in the real world.

I beg all participants, especially DocDwarf, to stop nitpicking and
quarelling about nothing.

Lüko Willms http://www.mlwerke.de
/--------- L.WILLMS@jpberlin.de -- Alle Rechte vorbehalten --

"Die Arbeit in weißer Haut kann sich nicht dort emanzipieren, wo sie
in schwarzer Haut gebrandmarkt wird." - Karl Marx 12.11.1866

Sponsored Links