Re: interesting use of NEXT SENTENCE vs. CONTINUE


"Joe Zitzelberger" <joe_zitzelberger@xxxxxxxxxx> wrote in message
news:joe_zitzelberger-1EAD2A.14213911062005@xxxxxxxxxxxxxxxxxxxxxxxxxxx

> If you truly need to do "informed coding", then you should not waste
> your time with an HLL, but hand code it in assembler to really ensure
> performance, or get a better compiler.

Ummm ... explain to me once again how you accomplish coding in assembler
language on a system on which there is no assembler language ...

One could argue that on the Unisys MCP systems the various ALGOL dialects
serve the function of an assembler -- but ALGOL isn't all that well-suited
to business application coding, being an "algorithmic language" rather than
a "business-oriented" one.

ANSI X3.23-1985 (page VI-18" notes "Items and constants in working storage
which bear no hierarchical relationship to one another need not be grouped
into records ... Instead, they are classified and defined as noncontiguous
elementary items. Each of these items is defined in a separate data
description entry which begins with the special level-number, 77."

Why is the idea that access to an item named X and declared
77 X PICTURE S9(11) USAGE BINARY.
might behave differently from access to an item named Y and declared
01 Y PICTURE S9(11) USAGE BINARY PICTURE S9(11).
somehow an assembler-language issue, whereas the idea that "everybody knows"
the contents of an item declared
01 Z PICTURE 9 PACKED-DECIMAL.
could be negative is a matter that all COBOL programmers in any operating
environment should take for granted?

Where I come from, if the data item is DISPLAY or PACKED-DECIMAL, and you
don't declare a sign, there's no space allocated for a sign. Why should it
be otherwise?

There are lots of things about COBOL that relate to particular operating
environments, and it is appropriate for the application programmer to be
aware of what works well in the environment in which he's operating, and
what doesn't. The programmer who says there's *never*a good reason to
declare an independent data item that isn't related hierarchically to any
other data item at Level 77 instead of Level 01 is mistaken. There may be
no performance difference in *some* operating environments, but that does
not mean there is no performance difference in *any* operating environment.

-Chuck Stevens


.

Relevant Pages

  • Re: To cross develop or not - was Re: eBay WTH?
    ... environment that could start up quickly and load the binary I had ... You don't know an assembler IDE by chance? ... code on the Apple II where possible (I like the "no distractions" ... I'll have a look at cc65 which, judging by your efforts, contains ...
    (comp.sys.apple2)
  • Re: Best 8-Bit Assembler?
    ... archives posted for download; when I tried to use it, I ended up with chunks of zeros in the middle of my object code files. ... This was a common environment situation for the era. ... After spending a little bit of time trying to roll my own assembler, ...
    (comp.sys.apple2.programmer)
  • Re: Hutch insane conduct
    ... and start macticating on that hat to keep your promise to the members. ... well as i fully disagrea with Herbert's ones, ... Not considering his confusions in between Assembler and ... Environment, Environment and Environment, Assembler and ...
    (alt.lang.asm)
  • Re: C++ recovery from (an otherwise) system abend
    ... I'm using assembler along with C. ... Perhaps an ESTAE that's set just before the PC and cleared just after? ... you may be able to set up the environment you are looking for. ... Subject: C++ recovery from system abend> To: ...
    (bit.listserv.ibm-main)
  • Re: If you were inventing CoBOL...
    ... > the assembling of the commands, the assembler only translates them. ... Originally computers had to be programmed by punching cards with 20 ... Operating System decided this. ... peripheral frames around the outside. ...
    (comp.lang.cobol)