W3C home > Mailing lists > Public > uri@w3.org > September 2003

Re: definition of "character"

From: Dan Connolly <connolly@w3.org>
Date: Sat, 27 Sep 2003 08:01:09 -0500
To: Mike Brown <mike@skew.org>
Cc: uri@w3.org
Message-Id: <1064667669.2698.341.camel@dirk.dm93.org>

On Sat, 2003-09-27 at 00:58, Mike Brown wrote:
> Dan Connolly wrote:
> > On Fri, 2003-09-26 at 08:52, Mike Brown wrote:
> > > Yet nowhere is "character" actually defined.
> > 
> > Nor is 'point' defined in geometry.
> > 
> > The URI spec is much more concrete than geometry
> > in that the characters are enumerated explicitly.
> > http://gbiv.com/protocols/uri/rev-2002/rfc2396bis.html#collected-abnf
> > 
> > i.e. you want to know what a character is? 'a' is
> > a character, as is 'b' and so on, and there aren't
> > any others.
> > 
> > [...]
> >
> > While I don't think this is necessary, I don't suppose
> > it would hurt. But suggesting that the editor "should at
> > least make an attempt to define..." is, as far as I can tell, less
> > effective than offering suggested text. I would guess
> > that Roy has attempted to draft text that makes everybody
> > happy some hundreds of times (hmm... thousands, if you
> > count the HTTP spec too); I wouldn't be surprise
> > if he scans messages for suggested text and deletes
> > those that don't have any summarily.
> 
> I seem to have struck a nerve.

Did it seem that way? Maybe I have grown a nerve around
"you should...", not to mention requests for definitions
of primitive terms like 'character' and 'resource'.

Sorry about that.

>  I thought I was humbly deferring to the experts
> to choose the ideal wording to address certain issues, and to decide whether I
> am overestimating the importance of the issues I am raising -- that is, if
> anyone thinks that I even know what I'm talking about at all.

If I didn't think you had something to contribute, I probably
wouldn't have answered at all. ;-)


> I feel that I can invest time in explaining the nature of an issue and why I
> feel it warrants an edit, and I can invest time in making explicit suggestions
> of edits (essentially making the edits myself). As a very new and temporary
> party to the discussion, with no credentials and no formal role other than
> interested observer (unless RFC 2396bis is more of a community effort than I
> am led to believe),

It is most certainly a community effort.

"1. What Is the IETF?
The Internet Engineering Task Force is a loosely self-organized group of
people who contribute to the engineering and evolution of Internet
technologies.

There is no membership in the IETF. [...] The closest thing there is to
being an IETF member is being on the IETF or Working Group mailing lists
(see Section 1.3)."

 -- http://www.ietf.cnri.reston.va.us/tao.html

>  it seems reasonable to concentrate more on the former
> investment than the latter, until directed otherwise. If there are posting
> guidelines that I have missed, please send the URL my way.

I thought it would be easy to find a message from Roy saying
"the best way to contribute is to send suggested text"
but a few minutes of searching turned up nothing like that.
Maybe I'm imagining it.

> In the meantime, I hereby revoke my suggestion that "character" be explicitly
> defined. Your message made me realize that it would probably be inappropriate
> to make such a definition, since URIs are canonically defined in ABNF terms,
> and therefore inherit ABNF's implicit "definition" (declaration, rather) of
> character.

Glad you agree.

> I do still feel that opportunities to correct widely-held misperceptions about
> character & encoding-related concepts & terminology should not be squandered,
> if retention of said misconceptions makes the spec more difficult to
> understand or detracts from its usefulness and implementability. But whatever.
> Never mind.

Yes, well, I think lots of people would like to clear that sort
of thing up. But finding the right words isn't easy.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Saturday, 27 September 2003 09:01:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:06 UTC