Re: Normative vs Informative

From: Mark Davis (
Date: Mon Nov 06 2000 - 04:57:31 EST

I was looking for a different message from Michael, and ran across this one.
This issue was not followed up on this list, so I thought I would report the
status so that people are not confused. As we found out later, Michael and
Ken's first take on this issue did not actually match the ISO directives:
the ISO definition of "normative" does not deal with process and level of
control, it is specifically conformance.


----- Original Message -----
From: "Michael Everson" <>
To: "Unicode List" <>
Sent: Thursday, October 26, 2000 03:15
Subject: Re: Normative vs Informative

Ar 00:04 -0700 2000-10-26, scríobh Mark Davis:

>I am leary of using normative your way unless we find strong evidence of

Well, that's just wrong, Mark. (Sorry, it's beat-up Mark day I guess.)

Ken explained Normative and Informative as they are used in ISO, that is,
in the standardization world. Unicode is also a Standard, ergo, Unicode
should use the words and concepts in the same way.

I imagine that somewhere in the ISO directives the words normative and
informative are specifically defined, and you could ferret those out if you
wanted, but I assure you that what Ken has said (in his rather good
summary) is what we do in ISO.

Ken's summary of the issues looks to me like something that could result in
text which could be added to the Introduction or to Chapter 3.

>> So, to summarize so far, what I see are four distinct issues:
>> 1. Normative vs. Informative
>> This bespeaks process and level of control. Who is in charge of
>> it, and what process must be followed to add to or change it?
>> 2. Scope of applicability to conformance
>> Because the standard is not an all or nothing prospect, what
>> normative parts are required for conformance in what implementation,
>> and what informative parts are relevant?
>> 3. Overridability
>> What normative (or informative) parts of the standard are
>> subject to overriding by higher-level protocols, and to
>> what degree, and by what kind of protocol?
>> 4. Immutability
>> What normative parts of the standard are forever fixed, and
>> which are subject to change (besides addition of new characters),
>> and what degree of lability is predictable for the mutable parts?
>> This also applies to informative parts, although there is little
>> expectation that an informative part will actually be immutable.
>> I don't want to see us fumble this by "hacking" at the problem,
>> introducing yet another partly thought-through patch on
>> normative versus informative as applied to some character properties.

Michael Everson ** Everson Gunn Teoranta **
15 Port Chaeimhghein Íochtarach; Baile Átha Cliath 2; Éire/Ireland
Vox +353 1 478 2597 ** Fax +353 1 478 2597 ** Mob +353 86 807 9169
27 Páirc an Fhéithlinn; Baile an Bhóthair; Co. Átha Cliath; Éire

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:15 EDT