W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2001

RE: Content-Type / charset in RFC2518, deltaV and ACL specs

From: Julian Reschke <julian.reschke@gmx.de>
Date: Wed, 3 Oct 2001 19:36:27 +0200
To: "Webdav WG" <w3c-dist-auth@w3c.org>
Message-ID: <JIEGINCHMLABHJBIGKBCCEKGDCAA.julian.reschke@gmx.de>
Geoff,

maybe it's obvious for people who constantly deal with these issues, but
many people blindly copy from examples, not knowing exactly what they do...
Of course properly declaring the charset is a Good Thing (and often
necessary), but then maybe the spec should come with a warning that the
charset declarations in the examples are just that -- examples -- and must
match the *actual* encoding used for the request/response bodies.

For instance:

<?xml version="1.0" encoding="UTF-8"?>
<town>Münster</town>

may look completely innocent -- until it appears within something which
actually is encoded in a different (and incompatible) charset. I guess this
mail will be transferred in some ISO encoding, so the XML fragment wouldn't
parse when given byte-by-byte to an XML parser, right?.

The examples in the specs generally don't need anything outside the ASCII
range, so maybe it would be wiser to leave out the encoding declarations
(and have a separate paragraph in RFC2518++ explain the issue).

(sorry for cross-posting to "both" lists).

Julian

> -----Original Message-----
> From: w3c-dist-auth-request@w3.org
> [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Clemm, Geoff
> Sent: Wednesday, October 03, 2001 6:54 PM
> To: Webdav WG; ietf-dav-versioning@w3.org
> Subject: RE: Content-Type / charset in RFC2518, deltaV and ACL specs
>
>
> Actually, section 3.6 states that a "value" is either a "token" or
> a "quoted-string", so the examples in 2518 are syntactically valid.
> But that doesn't mean that they are acceptable to common implementations,
> so it's a fair question.
>
> There is time to strip off the quotes in the DeltaV spec in the final
> editing pass, but that is likely to be soon, so if anyone feels strongly
> about this, please let me know ASAP.
>
> I don't follow your rationale for why examples in specs should not
> contain encoding information.  An example is supposed to accurately
> reflect what the client should send and expect to receive, and a client
> should send encoding information in a charset parameter in the
> Content-Type header.
>
> Cheers,
> Geoff
>
>
> -----Original Message-----
> From: Julian F. Reschke [mailto:julian.reschke@greenbytes.de]
> Sent: Wednesday, October 03, 2001 11:20 AM
> To: Webdav WG; ietf-dav-versioning@w3.org
> Subject: Content-Type / charset in RFC2518, deltaV and ACL specs
>
>
> Hi,
>
> I just noticed that in their examples, all these specs specify:
>
> 	Content-type: text/xml; charset="utf-8"
>
> However, as far as I understand RFC2616 (section 3.6 and section 3.7), and
> >from experience when setting the encoding in Java servlets, it should be
>
> 	Content-type: text/xml; charset=utf-8
>
> (so the value is not quoted).
>
> In general, I think it really doesn't make sense to specify character sets
> in specs (unless the spec is talking about encodings, of course). The spec
> contains characters after all (not an octet stream). Of course this also
> affects XML declarations in the specs.
>
> Julian
>
>
Received on Wednesday, 3 October 2001 13:35:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:58 GMT