SC22/WG20 N793

 

October 26, 2000

 

TO:                   WG20

 

From:                Ann Bennett

                        Convener, ISO/IEC JTC 1/SC 22/WG 4 - COBOL

 

Subject:            Liaison report from WG4 to WG20

 

 

J4 and WG4 have looked to WG20 for advice on internationalization and character set issues over the years and have appreciated the support provided by WG20 and especially by Keld Simonsen as liaison.  This support has enabled WG4 to include substantial support for large character sets and cultural adaptability in the COBOL revision currently in progress.  WG4 now intends to forward a draft for FCD ballot after the December 3-9, 2000, meeting of J4.  J4 produces the draft standard for WG4. 

 

The COBOL support for large character sets and cultural adaptability depends on the following ISO documents for normative reference:

 

·         ISO/IEC 10646-1:2000 - for the purpose of data interchange

 

·         ISO/IEC 9945-2:1993  (POSIX) - for specification of locale details

 

·         ISO/IEC 14651:2000 - for comparison using an intrinsic function

 

·         ISO/IEC TR 10176 and amendment 1, Annex A - for the specification of extended letters in identifiers

 

·         ISO/IEC DTR 14652 (category LC_CTYPE) - for specification of case folding of extended letters in identifiers, using the "tolower" specification

 

Recent electronic mail has indicated two controversies involving DTR 14652:

 

1)   It is a Type 1 TR and considered by some to be inappropriate as a normative reference.

2)   There are differences between DTR 14652 and the Unicode database.

 

COBOL has been planning to reference 14652 ever since an SC22 plenary, in Finland I believe, where several WG20 members met with working group conveners and I was advised that 14652 could be used as a reference for case folding.  Given that there is no other document available for normative reference, WG4 has the following options:  reference TR 14652, provide its own specification, or remove extended letters.  I consider the second 2 options extreme.  To deal with the Type 1 TR issue, I am requesting that WG20 include the "tolower" specification of TR 14652 in a future amendment of TR 10176.  Because of the timing of the COBOL FCD, the current COBOL revision will need to reference DTR 14652, but another reference will be needed in the event TR 14652 is withdrawn in the future.

 

A large COBOL vendor in J4 is interested in having no inexplicable differences between the Unicode specification and TR 14652.  This will provide an advantage to users by minimizing the differences in case folding between COBOL and other programming languages.  I am requesting that WG20 resolve any differences between the 2 specifications with respect to "tolower" where it is feasible to do so.  At present J4 prefers the "tolower" specification in DTR 14652 for the following two characters:

 

            Latin capital dotted I               0130

            Latin small dotless I               0131

 

It is an acceptable approach to me for COBOL to override 14652 in the COBOL specification if WG20 deems the Unicode specification of these two characters more appropriate for TR 14652.

 

COBOL does not currently have a dependency on WG20's project for Internationalization API's.  J4, on behalf of WG4, reviewed an early draft of the API standard and concluded that some of the functionality would be of interest for a future COBOL standard. 

 

I will be available by telephone on October 31 or November 1 if WG20 wishes to discuss COBOL requirements with me.

 

 

Regards,

Ann Bennett