[lsb-discuss] Perl in LSB 3.2 - RFC

Nick Stoughton nick at usenix.org
Fri Jun 8 09:43:58 PDT 2007


On Fri, 2007-06-08 at 17:48 +0200, Rafael Garcia-Suarez wrote:
> On 08/06/07, Nick Stoughton <nick at usenix.org> wrote:
> > Not well enough ... it should say something like:
> > The implementation shall provide a version of the Perl interpreter at
> > version 5.8.8 or later.
> 
> Indeed.
> 
> > I also don't approve of the specification being off in some other
> > non-normative place
> > (http://search.cpan.org/~nwclark/perl-5.8.8/pod/perl.pod)
> 
> What would be a normative place ?

Well, the LSB itself is the ideal!

The rules under the "ISO process" (in the event that this document were
ever to go down the ISO path) allow for normative references to come
from:

+ ISO standards
+ IEC standards
+ National Standards
+ ITU standards
+ Approved standards development organizations (includes, for example,
IEEE, IETF, W3C)

Other sources have a rigorous set of criteria that must be met to be
considered. Approval of such other sources comes from the subcommittee
responsible for the document. The criteria include
+ clear patent policy in line with the ISO policy
+ IPR, copyrights and trademarks 
+ market acceptance as the definitive reference
+ clear and full conformance requirements that meet or exceed those of
the referencing specification (this includes use of appropriate language
such as "shall")
+ any normative reference within the normatively referenced document
meets the same requirements
+ has been developed in a suitably open manner
+ ongoing maintenance plans in place
+ cannot be changed without the referencing standard's approval
+ Quality: "The SC (subcommittee) shall establish that the proposed RS
(referenced specification) is of adequate quality, considering topics
such as the length of the time the specification has existed, whether
products have been implemented using it, whether conformance
requirements are clear, and whether the specification is readily and
widely available."

Look, for example, at what we did for the zlib compression library,
where the original reference was deemed unsuitable for normative
inclusion.

For those interested in the whole set of rules, please see Annex N
starting on page 167 of the JTC 1 directives at
http://www.open-std.org/jtc1/sc22/open/n4031.pdf
-- 
Nick




More information about the lsb-discuss mailing list