On Thu, 20 Oct 2016 13:17:07 -0400, Rick Troth wrote:

>On 10/14/16 18:50, Paul Gilmartin wrote:
>> CMS Pipelines (perhaps other CMS utilities) use 0x25 instead of 0x15.
>> There's some very Bad History behind all this.
>
>So you're saying even Hartmann makes mistakes? Shock!   :-)
> 
Rather, I'd call him a "pedantic stickler[] for doco".

>0x25 is EBCDIC "linefeed".
>Sadly, that's printable if misinterpreted as ASCII. The value of 0x15
>and 0x0A is they're both non-printable in both worlds.
> 
Many ASCII editors (and, I believe, Info-Zip) make decisions on statistics.
Any UTF-8 file is valid ISO8859-1.  Most ISO8859-1 files are not valid
UTF-8.  Files containing no characters >=0x80 are probably ASCII,
etc.

>0x0A is referred to as newline in Unix land, but is officially
>"linefeed", so would map to EBCDIC 0x25. Certain pedantic sticklers for
>doco (against the grain of actual /usage/) ... cough ... IBM ... cough
>... would insist on translating EBCDIC 0x25 to/from ASCII 0x0A. Bad
>History indeed!
> 
The burden of that Bad History impelled IBM to transgress its own
specs when implementing iconv.

Linux iconv insists ... I might less rather call the developers "pedantic
sticklers" than merely naive adherents to the doco.  IBM could alleviate
this by defining a code page, call it 1047x, with LF at 0x15 and NL at
0x25.  But pedantic sticklers of another ilk declare, "Curst be that
moves my [control characters]."

And IBM could have evaded all this and many other conflicts if IBM
had chosen to make OpenEdition ASCII-based instead of EBCDIC.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to