Re: new draft "Framework: Process & Operational Guidelines"

See "IJ:" below.

Lofton Henderson wrote:

> As lead editor for this part, Kirill will be the one to respond to most 
> points.  I see that there are a couple with potential issues -- ckpt 
> priorities, commitment table, etc -- that should be raised for telcon 
> discussion, if anyone disagrees with the suggestions.
> 
> I have just a couple of replies, about editorial bits that I was 
> involved with...
> 
> At 03:42 PM 1/25/2002 +0100, you wrote:
> [...]
> 
>> >    All review comments against the previous version
>> >    have been accepted and implemented, except as indicated in an
>> >    associated [19]disposition of comments document. This version still
>>
>> The disposition of comments actually lists a bunch of "taken"
>> comments, so it's unclear if it is not up-to-date or if the
>> description here is wrong.
> 
> 
> The latter.  I wrote the Status and posted the document first, before 
> D.o.C. was written.  I think this reference should go away in FPWD, in 
> any case.
> 
>> [...]
>> >   1.3 Terminology
>> >
>> >    The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
>> >    "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" will be
>> >    used as defined in RFC 2119 [39][RFC2119].
>>
>> Since the checkpoints define the requirement/normativeness level, and
>> they use priority instead of verbiage, one may wonder why we say we're
>> going to use the RFC2119 model. The WAI guidelines do not refer to it
>> I think.
> 
> 
> I think one of them actually does (in briefer form), but I can't 
> remember which one (and don't have Web access now).  I'd propose that we 
> revisit this in a later draft, whether the rfc2119 reference can be 
> eliminated completely.


IJ: It's still important to use the terms of RFC2119. If you only
use the terms of RFC2119, your conformance granularity is binary:
either you MUST do it, or you don't have to. If a specification
requires additional granularity, then the RFC2119 terms should still
be used in addition to whatever other mechanisms the specification
imposes. In the case of WAI Guidelines, "Priority 1" checkpoints
create a subset of all checkpoints. Within that subset, you still
MUST do some things and SHOULD do other things. In UAAG 1.0,
there are even more subsetting mechanisms. But for any subset,
you still MUST do some things and SHOULD or MAY do others.

In the WAI Guidelines, by the way, we did not capitalize the
key words. That's optional in RFC2119. It may be helpful
to capitalize them (e.g., to distinguish "may:is allowed to"
from "may: is capable of").

Hope this helps,

  - Ian


-- 
Ian Jacobs (ij@w3.org)   http://www.w3.org/People/Jacobs
Tel:                     +1 718 260-9447

Received on Sunday, 27 January 2002 22:16:53 UTC