[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Two technical questions on IEEE Std 754-2008



> And along my point has been that the BID/DPD distinction should ALSO be
> at the execution-environment level, and not at the individual-field level.
> But I understand Chuck's point of view if COBOL does not now have STANDARD
> means to do this kind of execution-environment tagging, yet field tagging
> is under active construction and can thus accomodate the distinction now.

It IS at the execution-enviromment level -- you specify which one the "operands" need to be in as part of arithmetic operations via the ARITHMETIC clause. 
 
It is ALSO at the individual field level -- but not at execution time, rather, as a description of what the datum looks like. 
 
As Bill Klein has pointed out, "tagging" of data isn't now part of COBOL, and my opinion is that providing for that is a HUGE tail that wags a VERY SMALL dog from COBOL's perspective.  Unless there's a major outcry from the COBOL user community, I wouldn't want to add that feature in general, and certainly not on behalf of the DPD/BID controversy. 
 
>There are still differences at the file level (e.g. how records are separated, ...
 
 COBOL "outermost" records are simply separate entities, and historically have been represented externally as of fixed length (that's changing quickly).  As much as COBOL understands about "separation" of individual records is illustratred by "BLOCK CONTAINS <integer> RECORDS" on file declarations in I/O subsystems.  Records aren't separated "by something", they're separate entities.  A record in a file declaration may have multiple potential descriptions even at the outermost level, but they all are considered cospatial and map to the same "memory space", which is the size of the largest single record description subordinate to the file. 
 
> whether there is a difference between "binary" and "text" files (addressed by FTP and TFTP for example),
 
Also, files in COBOL are inherently TEXT (character-based) from the external view.  They may contain non-text data.  That's been true pretty much forever -- a group item was always USAGE DISPLAY prior to 2002, when GROUP-USAGE BIT and GROUP-USAGE NATIONAL were added.  Note that these clauses are restrictive on what data can look like subordinate to the declaration, but there's nothing wrong with an outermost record (USAGE DISPLAY) haveing subordinate records declared GROUP-USAGE BIT and/or GROUP-USAGE NATIONAL.  
 
    -Chuck Stevens

754 | revision | FAQ | references | list archive