15 Oct
2019
15 Oct
'19
8:07 a.m.
From: Chuck Guzis
One could argue that it's just as similar to
FORTRAN (cf. computed GOTO
and logical IF statements).
It probably worth pointing out that I never used COBOL, and have little
knowledge of it. So when one reads "it is vaguely reminiscent of COBOL, as it
has a 'Data Division' and a 'Procedure Division'", I must have copied
that
all from some source I found, because I don't know what the 'Data Division'
and a 'Procedure Division' are (although I can guess from the names).
Where it differs mainly from FORTRAN of the times is a
facility for
record layout
I seem to recall that COBOL was the first language with support for
structures? If DIBOL has support for them too, which would be another
similarity between the two.
Noel
15 Oct
15 Oct
9:27 a.m.
On 10/15/19 8:07 AM, Noel Chiappa via cctalk wrote:
I seem to recall that COBOL was the first language
with support for
structures? If DIBOL has support for them too, which would be
another similarity between the two.
Both FLOW-MATIC (1953) and COMTRAN (1957) had support for data
structures, though the data structure notation was not part of the
program procedure. FLOW-MATIC placed the structure and naming
information on data tapes (pretty novel for the period) and COMTRAN put
the specification of data structure on separate fixed-field section the
program input.
The general idea (as explained in Grace Hopper's FLOW-MATIC description)
was to make the data record structuring an independent task. She
mentions being able to work out the procedure section before designing
the data layout. I suspect that her record layout is one of the
earliest DDLs.
COBOL built on this. Looking at the procedure section of COMTRAN or
FLOW-MATIC programs, there's no mistaking them as COBOL predecessors.
DIBOL, not so much.
COBOL, by the late 60s was a very complex language, so much so that the
initial S/360 DOS releases were incomplete; they lacked support for ISAM
files. Somewhere, I have an IBM document that describes the
machine-language subroutines that were provided to fill the gap
temporarily via the "ENTER LINKAGE" statement.
I'm not at all certain that FLOW-MATIC had the first structured data
description language, but it was pretty close.
--Chuck
1895
days inactive
1895
days old
1 comments
2 participants
participants (2)
-
cclist＠sydex.com -
jnc＠mercury.lcs.mit.edu