Real Old School Programming (was: Re: Where to buy a Selectric?)
cclist at sydex.com
Sat Dec 31 04:55:29 CST 2005
On 12/31/2005 at 8:45 AM Nico de Jong wrote:
>vertically divided. For Cobol, you could see e.g. that the first 3 columns
>were blank, the next 3 numbered from 010 to 300 or so, then a divider
>between between 15 and 16 (IIRC), and again "around" col. 73 (marking for
IIRC, the first six columns were "programmer sequence numbers"; ( i.e., the
compiler didn't really care about them, but would issue a warning if they
were out of order) divided into two groups of three (1-3 were the page
number, 4-6 were the line number). Most folks just left them blank.
Column 7 was the continuation; an asterisk in this column signified a
comment card; any other non-blank signified a continuation. Area "A"
started in columns 8-11 (Used for divison and section headers and
paragraph names and 01 and FD level items, etc.); Area B (used for
everything else) started in column 12. And, as in the case of most other
card-oriented languages, 73-80 were reserved for sequence numbers or other
identification, should you drop the box containing your source code.
The continuation scheme in COBOL was different from FORTRAN. Since COBOL
is sentence-based (approximately), statements could continue from one card
to the next with no special consideration, much like 'C'. You needed the
continuation column only if you were writing a long character literal. I
don't believe that COBOL allowed for continued numeric literals.
In spite of its starting out as a language using more-or-less English
sentence syntax, COBOL is a tough language to learn because it's quite
large, particularly if you consider all of the variations of data types
and statements (e.g., how many variations of the INSPECT statement can you
More information about the cctalk