W3C home > Mailing lists > Public > www-html-editor@w3.org > October to December 1997

Re: WD-html40-971024/references.html [was: 24 Oct Release of HTML 4.0 spec]

From: Ian B. Jacobs <ij@w3.org>
Date: Sun, 26 Oct 1997 17:46:32 -0500
Message-ID: <3453C848.68DA2BDA@w3.org>
To: Dan Connolly <connolly@w3.org>
CC: www-html-editor@w3.org
Dan Connolly wrote:
> 
> I'm reviewing the list of normative references, i.e. other
> specs upon which the definition of HTML 4 rests, and I
> am surprised to find such things as postscript, CSS1, PICS,
> etc. (though perhaps we import some units or something from
> postscript and CSS1).
> 
> So I did an audit to find each reference to a normative
> citation:

Great idea and great work! This was *very* helpful.


> Suggested changes are marked with ***
> 
> *** I suggest you audit each occurence of "rel=biblioentry" in
> the HTML4 spec and change the normative references
> to "rel=norm-ref" (is it possible to get CSS to do styles
> based on link relationships? if not, perhaps class=normref
> would help for a visual audit of the spec as well.)

I have added this to the TODO list.


> And here are the results for each normative reference:
> 
> ===== ADOBE90
> references.html:33:<DT><STRONG><A
> name="ref-ADOBE90">[ADOBE90]</A></STRONG></DT>
> 
> Evidently there are *no* references to [ADOBE90]
> *** delete. this citation,
>         or move it to the list of informative references

I see no reason to keep the reference. I will remove
it unless I hear otherwise.

> ===== CHARSETS
> charset.html:106:href="./references.src#ref-CHARSETS">[CHARSETS]</a> for
> a complete
> references.html:39:<DT><STRONG><A
> name="ref-CHARSETS">[CHARSETS]</A></STRONG></DT>
> 
> This is only referenced in an informative section about
> the use of the term "character set" in this spec.
> 
> It must be referenced from, for example, by the specification
> of the charset attribute:
> http://www.w3.org/MarkUp/Group/9710/WD-html40-971024/struct/links.html#adef-charset
> 
> *** I suggest s/charset = cdata/charset = character encoding scheme/
> and a new type "character encoding scheme" in the types section:
> http://www.w3.org/MarkUp/Group/9710/WD-html40-971024/types.html

Good idea. I have added a section on "Character encoding shemes"
to types.html that contains a (normative) link to [CHARSETS]
as well as a link to 
http://www.w3.org/MarkUp/Group/9710/WD-html40-971024/charsets.html#encodings.

> 
> ===== CSS1
> convent.html:104:rel="biblioentry"
> href="../references.html#ref-CSS1">[CSS1]</a>).
> 
> That's an informative reference (in the discussion of "deprecated").
> 
> *** s/see [CSS1]/see, for example, [CSS1]/
> 
> issues.html:696:      (cf. CSS1 'width' property)?</TD>
> 
> references.html:46:<DT><STRONG><A
> name="ref-CSS1">[CSS1]</A></STRONG></DT>
> 
> references.html:50:Available at <A
> href="http://www.w3.org/TR/REC-CSS1-961217.html">
> references.html:51:http://www.w3.org/TR/REC-CSS1-961217.html</A> </DD>
> 
> types.html:177:href="./references.html#ref-CSS1">[CSS1]</a>.
> types.html:226:rel="biblioentry"
> href="./references.html#ref-CSS1">[CSS1]</a> style
> present/styles.html:93:href="../references.html#ref-CSS1">[CSS1]</a>,
> but other style sheet
> present/styles.html:165:href="../references.html#ref-CSS1">[CSS1]</a>),
> abbreviated CSS, for
> struct/objects.html:1400:href="../references.html#ref-CSS1">[CSS1]</a>.
> struct/tables.html:1622:"../references.html#ref-CSS1">[CSS1]</A> did not
> offer mechanisms
> 
> OK; css is referenced normatively in types.html (for the
> definition of a Pixel)



> ===== DATETIME
> references.html:53:<dt><strong><a
> name="ref-DATETIME">[DATETIME]</a></strong>
> struct/text.html:1024:href="../references.src#ref-DATETIME">[DATETIME]</a>
> for its
> struct/text.html:1118:<a rel="biblioentry"
> href="../references.src#ref-DATETIME">[DATETIME]</a>
> 
> OK, this is a normative reference. But I think
> it introduces another type in HTML.
> 
> *** s/datetime = cdata/datetime = datetime/ and add a new
>         item in "7 Types"

Good idea. I also added a parameter entity to the dtd.

> ===== ERCS
> references.html:59:<DT><STRONG><A
> name="ref-ERCS">[ERCS]</A></STRONG></DT>
> 
> [ERCS] is cited but not referenced anywhere in the document
> 
> *** delete the [ERCS] citation or move it to the informative list

I'll delete it unless I hear otherwise.

> ===== ISO639
> references.html:65:<DT><STRONG><A
> name="ref-ISO639">[ISO639]</A></STRONG></DT>
> struct/dirlang.html:145:href="../references.html#ref-ISO639">[ISO639]</a>
> language

This is a normative reference in
http://www.w3.org/MarkUp/Group/html4-src/struct/dirlang.html#h-9.1.1
Therefore, I think this should stay.

> ===== ISO646
> references.html:74:<DT class="incomplete"><STRONG><A
> name="ref-ISO646">[ISO646]</A>

I'll remove unless I hear otherwise.


> ===== ISO1000
> references.html:80:<DT><STRONG><A
> name="ref-ISO1000">[ISO1000]</A></STRONG></DT>

I'll remove unless I hear otherwise.


> ===== ISO3166
> references.html:85:<DT><STRONG><A
> name="ref-ISO3166">[ISO3166]</A></STRONG></DT>
> struct/dirlang.html:152:href="../references.html#ref-ISO3166">[ISO3166]</a>
> country code.
> 
> OK.
> 
> *** suggest: add a "for more info, see..." link from
>         "7.6 language codes"
>         to 9.1 Specifying the language of content: the lang attribute

Ok. Good idea.


> ===== ISO4217
> references.html:89:<DT><STRONG><A
> name="ref-ISO4217">[ISO4217]</A></STRONG></DT>

I'll remove it unless I hear otherwise.

> 
> ===== ISO8601
> references.html:94:<DT><STRONG><A
> name="ref-ISO8601">[ISO8601]</A></STRONG></DT>
> struct/global.html:802:rel="biblioentry"
> href="../references.html#ref-ISO8601">[ISO8601]</a>
> 
> that's informative
> 
> struct/text.html:939:<A rel="biblioentry"
> href="../references.html#ref-ISO8601">
> struct/text.html:940:[ISO8601]</A> and limited by the profile defined in
> the section
> struct/text.html:1020:<P><A rel="biblioentry"
> href="../references.html#ref-ISO8601">
> struct/text.html:1021:[ISO8601]</A> allows many options and variations
> in the representation
> struct/text.html:1069:"../references.html#ref-ISO8601">[ISO8601]</A></P>
> 
> OK; those are normative. Acutally, we cite [DATETIME] normatively,
> but since it cites 8601 normatively, and the "cites normatively"
> relation is transitive, it's correct to cite 8601 normatively
> (though not strictly necessary).
> 
> *** s/datetime = cdata/datetime = datetime/ and add datetime
>         to "7 types". Move "10.4.1 Date and time format"
>         and replace it by little more than a reference
>         to [DATETIME] and 8601

Ok.

> 
> ===== ISO8879
> cover.html:50:rel="biblioentry"
> href="./references.html#ref-ISO8879">[ISO8879]</a>).
> references.html:100:<DT><STRONG><A
> name="ref-ISO8879">[ISO8879]</A></STRONG></DT>
> intro/sgmltut.html:48:href="../references.html#ref-ISO8879">[ISO8879]</a>).
> SGML provides a
> intro/sgmltut.html:262:href="../references.html#ref-ISO8879">[ISO8879]</a>
> are expected to
> sgml/entities.html:55:href="../references.html#ref-ISO8879">[ISO8879]</a>).
> 
> *** (strictly editorial): s/(defined in [ISO8879])/[ISO8879]/

I think you mean the opposite:
 s/([ISO8879])/(defined in [ISO8879])/

Ok.


> ===== ISO10646
> charset.html:51:href="./references.html#ref-ISO10646">[ISO10646]</a>.
> This set is
> charset.html:136:href="../references.html#ref-ISO10646">[ISO10646]</a>
> (registered by
> charset.html:250:href="./references.html#ref-ISO10646">[ISO10646]</a>,
> characters
> references.html:107:<DT><STRONG><A
> name="ref-ISO10646">[ISO10646]</A></STRONG></DT>
> intro/intro.html:121:href="../references.html#ref-ISO10646">[ISO10646]</a>)
> as the document
> sgml/entities.html:65:href="../references.html#ref-ISO10646">[ISO10646]</a>
> names are written
> sgml/entities.html:313:href="../references.html#ref-ISO10646">[ISO10646]</a>
> or use
> sgml/entities.html:316:href="../references.html#ref-ISO10646">[ISO10646]</a>
> characters
> sgml/entities.html:603:href="../references.html#ref-ISO10646">[ISO10646]</a>
> or use
> sgml/entities.html:606:href="../references.html#ref-ISO10646">[ISO10646]</a>
> characters
> struct/text.html:78:href="../references.html#ref-ISO10646">[ISO10646]</a>
> to unambiguously
> 
> OK.

Ok.

> ===== ISO88591
> references.html:115:<DT><STRONG><A
> name="ref-ISO88591">[ISO88591]</A></STRONG></DT>
> sgml/entities.html:72:href="../references.html#ref-ISO88591">[ISO88591]</a>-->
> sgml/entities.html:85:href="../references.html#ref-ISO88591">[ISO88591]</a>.
> struct/text.html:809:rel="biblioentry"
> href="../references.html#ref-ISO88591">[ISO88591]
> 
> *** s/The horizontal tab character (encoded in [UNICODE],
>       US ASCII, and [ISO88591] as decimal 9)/at code position 9
>       in [ISO10646], and [ISO88591]/
> 
> rationale: "encoded" suggest you're talking about character
>         enconding schemes, but you're not: you're talking
>         about coded character sets

Fine.

> US-ASCII is a charset, i.e. a character encoding scheme.

Fine.

> We cite 10646, not Unicode, for character code positions. Unicode
>         is only cited for things like the BIDI algorithm.

This I true, and I have just verified our usage of [ISO10646]
and [UNICODE] througout the spec. We refer to [ISO10646]
for characters and [UNICODE] for bidi. However, in the sgml
subdirectory, both are used in all the files. I would ask
the other editors to verify their usage here.

> ===== MIMETYPES
> references.html:121:<DT><STRONG><A
> name="ref-MIMETYPES">[MIMETYPES]</A></STRONG></DT>
> types.html:218:href="./references.html#ref-MIMETYPES">[MIMETYPES].</a>
> struct/links.html:350:href="../references.html#ref-MIMETYPES">[MIMETYPES]</a>.
> 
> *** s/specifies the nature of a linked resource/specifies
>         the nature of a piece of content, for example,
>         the result of dereferencing a resource/
> 
> The wording as is conflicts with
> the web architecture: http://www.w3.org/DesignIssues/Formats.html
> The relationship of resources-to-formats is 1-many,
> and you suggest it's 1-1. The resource doesn't (in general)
> have a media type; what you get back from dereferencing a resource does.

I think I understand this. I have looked for other occurrences
of of media-type attributes in the spec, and the ones for
PARAM and OBJECT may need review. The other elements (SCRIPT, STYLE,
and FORM) seem ok.

> "The text/html media type is defined by this specification. "
> Really? Where? media type definitions have a very specific
> form, ala:
> http://www.w3.org/MarkUp/html-spec/html-spec_4.html#SEC4.1
> 
> *** hmmm... this is a large edit, with non-trivial issues

See other email thread...


> ===== PICS
> references.html:129:<DT><STRONG><A
> name="ref-pics">[PICS]</A></STRONG></DT>
> references.html:131:<DD>Platform for Internet Content (PICS). For more
> information see
> references.html:132:<A
> href="http://www.w3.org/PICS/">http://www.w3.org/PICS/</A> </DD>
> struct/global.html:708:<h5><a href="global.html#edef-META"><samp
> class="einst">META</samp></a> and PICS</h5>
> struct/global.html:711:href="../references.html#ref-pics">[PICS]</a> is
> an infrastructure
> struct/global.html:720:class="einst">META</samp></a> declaration to
> include a PICS 1.1 label:
> struct/global.html:724: &lt;META http-equiv="PICS-Label" content='
> struct/global.html:725: (PICS-1.1 "http://www.gcf.org/v2.5"
> struct/global.html:728:      for "http://w3.org/PICS/Overview.html"
> 
> Those are all informative
> 
> *** move the PICS reference to the list of informative references

Ok.

> ===== RFC1123
> references.html:134:<DT><STRONG><A
> name="ref-RFC1123">[RFC1123]</A></STRONG></DT>
> 
> What the heck do we cite it for?

Gone.

> ===== RFC1468
> references.html:141:<DT><STRONG><A
> name="ref-RFC1468">[RFC1468]</A></STRONG></DT>
> 
> We don't reference it, but it's relevant enough to cite.
> 
> *** move to informative

Ok, I moved it.

> ===== RFC1556
> references.html:155:<DT><STRONG><A
> name="ref-RFC1556">[RFC1556]</A></STRONG></DT>
> struct/dirlang.html:500:href="../references.html#ref-RFC1556">[RFC1556]</a>)
> favors visual
> struct/dirlang.html:556:rel="biblioentry"
> href="../references.html#ref-RFC1556">[RFC1556]</a>,
> struct/dirlang.html:582:rel="biblioentry"
> href="../references.html#ref-RFC1556">[RFC1556]</a>,
> 
> I can't tell if this one is normative or not.
> I guess we can leave it be.

Ok.

> ===== RFC1590
> references.html:162:<DT><STRONG><A
> name="ref-RFC1590">[RFC1590]</A></STRONG></DT>
> references.html:218:obsoletes RFC1521, RFC1522, and RFC1590. </DD>
> references.html:226:obsoletes RFC1521, RFC1522, and RFC1590. </DD>
> 
> We should reference this in the definition of "text/html"
> which is missing.
> 
> *** see [MIMETYPES] above

After hunting around a little, I propose the following:

(1) Delete rfc1590, which is obsolete
(2) Insert rfc2048 as a normative reference. This rfc
    describes the registration procedure.
(3) Put a link from the note on the cover to rfc2048 saying that
    we intend to register text/html.

> ===== RFC1630
> htmlweb.html:54:rel="biblioentry"
> href="./references.html#ref-RFC1630">[RFC1630]</a>
> references.html:168:<DT><STRONG><A
> name="ref-RFC1630">[RFC1630]</A></STRONG></DT>
> 
> hmm... how do you describe it in "types"?
> ... by reference to "5.2 Uniform Resource Locators (URLs)"
> ... which cites 1738 normatively, and 1630 informatively.
> 
> *** move this one to the informative list.
> *** s/(defined in [RFC1738])/(defined in [RFC1738],[RFC1808])/
Ok.

> ===== RFC1738
> htmlweb.html:64:<a rel="biblioentry"
> href="./references.html#ref-RFC1738">[RFC1738]</a>).
> htmlweb.html:93:href="./references.html#ref-RFC1738">[RFC1738]</a> and
> <a
> htmlweb.html:105:href="./references.html#ref-RFC1738">[RFC1738]</a> or
> <a
> htmlweb.html:256:href="./references.html#ref-RFC1738">[RFC1738]</a>)
> issues.html:208:      See also <A
> HREF="ftp://ds.internic.net/rfc/rfc1738.txt">RFC1738</A></TD>
> references.html:176:<DT><STRONG><A
> name="ref-RFC1738">[RFC1738]</A></STRONG></DT>
> interact/forms.html:2040:href="../references.html#ref-RFC1738">[RFC1738]</a>:
> Non-alphanumeric
> intro/sgmltut.html:698:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> sgml/dtd.html:104:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> sgml/loosedtd.html:108:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> 
> OK.

Ok.

> ===== RFC1766
> issues.html:227:    <TD>No, see RFC1766</TD>
> references.html:183:<DT><STRONG><A
> name="ref-RFC1766">[RFC1766]</A></STRONG></DT>
> types.html:236:href="../references.html#ref-RFC1766">[RFC1766]</a>.
> Please consult
> sgml/dtd.html:95:    -- a language code, as per <a rel="biblioentry"
> href="../references.html#ref-RFC1766">[RFC1766]</a>
> sgml/loosedtd.html:99:    -- a language code, as per <a
> rel="biblioentry" href="../references.html#ref-RFC1766">[RFC1766]</a>
> struct/dirlang.html:121:href="../references.html#ref-RFC1766">[RFC1766]</a>
> defines and
> 
> OK.

Ok.

> ===== RFC1808
> htmlweb.html:152:rel="biblioentry"
> href="./references.html#ref-RFC1808">[RFC1808]</a>,
> htmlweb.html:164:href="./references.html#ref-RFC1808">[RFC1808]</a>)
> doesn't contain any
> htmlweb.html:171:rel="biblioentry"
> href="./references.html#ref-RFC1808">[RFC1808]</a>
> htmlweb.html:229:href="./references.html#ref-RFC1808">[RFC1808]</a>. The
> following is a
> htmlweb.html:231:href="./references.html#ref-RFC1808">[RFC1808]</a>
> applies to HTML.
> references.html:190:<DT><STRONG><A
> name="ref-RFC1808">[RFC1808]</A></STRONG></DT>
> intro/sgmltut.html:698:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> sgml/dtd.html:104:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> sgml/loosedtd.html:108:       see <a rel="biblioentry"
> href="../references.html#ref-RFC1808">[RFC1808]</a> and <a
> rel="biblioentry" href="../references.html#ref-RFC1738">[RFC1738]</a>
> 
> OK, but see edit in the 1630 entry above.

Ok.

> 
> ===== RFC1867
> references.html:197:<DT><STRONG><A
> name="ref-RFC1867">[RFC1867]</A></STRONG></DT>
> appendix/changes.html:189:href="../references.html#ref-RFC1867">[RFC1867]</a>)
> may be used to
> interact/forms.html:591:href="../references.html#ref-RFC1867">[RFC1867]</a>.
> interact/forms.html:611:href="../references.html#ref-RFC1867">[RFC1867]</a>
> for
> interact/forms.html:2083:href="../references.html#ref-RFC1867">[RFC1867]</a>,
> section 7.</li>
> 
> OK.

Ok.

> ===== RFC2044
> htmlweb.html:98:href="./references.html#ref-RFC2044">[RFC2044]</a>) as
> one or more bytes
> references.html:204:<DT><STRONG><A
> name="ref-RFC2044">[RFC2044]</A></STRONG></DT>
> 
> Shit! We modified the definition of a URL in the HTML spec!
> How did I let that happen?
> 
> Hmm... and we cite the URN syntax draft as well.
> 
> I'll have to follow up during the Proposed Recommendation
> phase, based on the outcome the IETF UR* discussions
> I recently launched.

Offhand, I don't remember where the text in htmlweb.html came from.

> ===== RFC2045
> charset.html:91:href="./references.src#ref-RFC2045">[RFC2045]</a>). This
> conversion
> 
> Hmm... "6.2 Character encodings" is a little fuzzy. It
> talks a lot about what Authors should do, but it doesn't
> formally make the connection between a text/html MIME body
> part and an SGML document entity the way it shoud. c.f.
> 
> http://www.w3.org/MarkUp/html-spec/html-spec_4.html#SEC4.2
> 
> *** add: "A message entity with a content type of `text/html' represents
> an
> HTML document, consisting of a single text entity. The `charset'
> parameter (whether implicit or explicit) identifies a character encoding
> scheme. The text entity consists of the characters determined by this
> character encoding scheme and the octets of the body of the message
> entity. " to 6.2. It could be edited for readability if you're
> certain you won't change the precise meaning.

I don't yet understand the relationship you are trying to establish
(due to my own ignorance). As is, I don't know how to make
this text fit into charset.html. For now, I will insert the text, 
but I would like to discuss this further so that I understand it.

> *** add all of
> http://www.w3.org/MarkUp/html-spec/html-spec_4.html#SEC4.1
> (this is the text/html media type definition that I mentioned
> above a couple times.) It really should be in our spec, but
> it needs careful review, since things like "Level" aren't
> necessarily still something we want to specify.

I think that later emails obsolete this request; for
now the cover will refer to our intention to register text/html,
but the spec will not yet include the (modified) text of 
http://www.w3.org/MarkUp/html-spec/html-spec_4.html#SEC4.1

> references.html:211:<DT><STRONG><A
> name="ref-RFC2045">[RFC2045]</A></STRONG></DT>
> types.html:207:href="references.html#ref-RFC2045">[RFC2045]</a> and <a
> interact/forms.html:147:  rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>.
> interact/forms.html:551:href="../references.html#ref-RFC2045">[RFC2045]</a>).
> sgml/dtd.html:87:    -- media type, as per <a rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>
> sgml/dtd.html:91:    -- an Internet character encoding, as per <a
> rel="biblioentry" href="../references.html#ref-RFC2045">[RFC2045]</a>
> sgml/dtd.html:99:    -- a space separated list of Internet character
> encodings, as per <a rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>
> sgml/loosedtd.html:91:    -- media type, as per <a rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>
> sgml/loosedtd.html:95:    -- an Internet character encoding, as per <a
> rel="biblioentry" href="../references.html#ref-RFC2045">[RFC2045]</a>
> 
> *** s/Internet character encoding/character encoding scheme/

I have doubts about this edit. It seems to me we used to
use the term "character encoding scheme" and we were
told subsequently to change it to "character encoding", which
is where it stands today. However, I will drop the "Internet"
part.

> 
> sgml/loosedtd.html:103:    -- a space separated list of Internet
> character encodings, as per <a rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>
> struct/dirlang.html:498:rel="biblioentry"
> href="../references.html#ref-RFC2045">[RFC2045]</a>,
> struct/links.html:380:href="../references.html#ref-RFC2045">[RFC2045]</A>
> (e.g., "euc-jp").
> 
> *** s/charset = cdata/charset = charset/ and add a new charset
>         type to section 7. (did I say that already?)

Yes, and it's been done.

> 
> ===== RFC2046
> references.html:220:<DT><STRONG><A
> name="ref-RFC2046">[RFC2046]</A></STRONG></DT>
> types.html:208:rel="biblioentry"
> href="references.html#ref-RFC2046">[RFC2046]</a>)
> 
> OK.

Ok.

> ===== RFC2068
> charset.html:76:rel="biblioentry"
> href="./references.src#ref-RFC2068">[RFC2068]</a>,
> charset.html:157:rel="biblioentry"
> href="./references.src#ref-RFC2068">[RFC2068]</a>)
> charset.html:175:href="./references.src#ref-RFC2068">[RFC2068]</a>)
> mentions ISO-8859-1
> 
> Those are informative.
> 
> htmlweb.html:81:rel="biblioentry"
> href="./references.html#ref-RFC2068">[RFC2068]</a>) to
> htmlweb.html:238:rel="biblioentry"
> href="./references.html#ref-RFC2068">[RFC2068]</a>).
> 
> "The base URL is given by an HTTP header (see [RFC2068]). "
> That's normative, but it shouldn't be.
> 
> *** s/an HTTP header/metadata discovered during a protocol
>         interaction, such as an HTTP header/
> 
> references.html:228:<DT><STRONG><A
> name="ref-RFC2068">[RFC2068]</A></STRONG></DT>
> appendix/changes.html:187:href="../references.html#ref-RFC2068">[RFC2068]</a>).
> This attribute
> present/styles.html:786:href="../references.html#ref-RFC2068">[RFC2068]</a>
> has the same
> struct/global.html:642:(<a rel="biblioentry"
> href="../references.html#ref-RFC2068">[RFC2068]</a>)
> struct/global.html:800:href="../references.html#ref-RFC2068">[RFC2068]</a>.
> As these formats
> 
> Yes. the http-equiv reference to 2068 is normative.
> 
> So the citation is fine in the list of normative references.

Ok.

> ===== RFC2070
> references.html:235:<DT><STRONG><A
> name="ref-RFC2070">[RFC2070]</A></STRONG></DT>
> intro/intro.html:116:href="../references.html#ref-RFC2070">[RFC2070]</a>,
> which deals with
> 
> we don't cite it; we should, in the changes section.
> In any case
> 
> *** move it to the list of informative references.

Ok. I also added a note to changes.src saying that HTML 4.0
takes into account the recommendations of rfc2070 for
the internationalization of HMTL.


> ===== RFC2119
> convent.html:43:href="./references.html#ref-RFC2119">[RFC2119]</a>. At
> times, the
> references.html:242:<DT><STRONG><A
> name="ref-RFC2119">[RFC2119]</A></STRONG></DT>
> 
> "Words such as "must", "should", "can", and "may" are used in accordance
> with
>      [RFC2119]."
> 
> !!! no they aren't: that spec says they MUST be in ALL CAPS.
> 
> ... not sure what edit to suggest.

There are two things in question here:
(1) Does the spec's use of the words correspond to the meaning of
    rfc2119?
(2) Do we have to use all caps for the key words?

We have been quite attentive to (1), but we need to review
it before releasing the spec as a recommendation.

We did not put the key words in all caps for stylistic reasons.
Either we put a note in convent.html explaining this, or
we put the words in all caps. If we choose the latter, we
can take advantage of the (big) edit to verify our usage. I think
this would be a good idea.

> ===== RFC2141
> htmlweb.html:95:href="./references.html#ref-RFC2141">[RFC2141]</a>. Any
> other
> htmlweb.html:106:rel="biblioentry"
> href="./references.html#ref-RFC2141">[RFC2141]</a>
> references.html:249:<DT><STRONG><A
> name="ref-RFC2141">[RFC2141]</A></STRONG></DT>

The reference in htmlweb seems normative to me.

> ===== SGMLOPEN
> references.html:255:<DT><STRONG><A
> name="ref-SGMLOPEN">[SGMLOPEN]</A></STRONG></DT>
> struct/global.html:133:href="../references.html#ref-SGMLOPEn">[SGMLOPEN]</a>).

In struct/global.html, the reference seems informative. I have
moved the reference to the informative section unless I hear
otherwise.

> ===== SRGB
> references.html:260:<DT><STRONG><A
> name="ref-SRGB">[SRGB]</A></STRONG></DT>
> types.html:107:href="./references.html#ref-SRGB">[SRGB]</a>. A color
> value may either
> types.html:159:href="./references.html#ref-SRGB">[SRGB]</a> color model
> together with

I think this is normative.


> ===== UNICODE
> charset.html:53:href="./references.html#ref-UNICODE">[UNICODE]</a>).
> Both of these
> charset.html:125:href="../references.html#ref-UNICODE">[UNICODE]</a>).
> references.html:268:<DT><STRONG><A
> name="ref-UNICODE">[UNICODE]</A></STRONG></DT>
> struct/dirlang.html:241:rel="biblioentry"
> href="../references.html#ref-UNICODE">[UNICODE]</a>)
> struct/dirlang.html:259:href="../references.html#ref-UNICODE">[UNICODE]</a>
> specification
> struct/dirlang.html:264:href="../references.html#ref-UNICODE">[UNICODE]</a>
> bidirectional
> struct/dirlang.html:386:href="../references.html#ref-UNICODE">[UNICODE]</a>
> bidirectional
> struct/dirlang.html:443:href="../references.html#ref-UNICODE">[UNICODE]</a>
> formatting
> struct/dirlang.html:449:href="../references.html#ref-UNICODE">[UNICODE]</a>
> characters. If both
> struct/dirlang.html:510:href="../references.html#ref-UNICODE">[UNICODE]</a>
> bidirectional
> struct/dirlang.html:548:href="../references.html#ref-UNICODE">[UNICODE]</a>
> formatting
> struct/dirlang.html:590:href="../references.html#ref-UNICODE">[UNICODE]</a>
> specification
> struct/dirlang.html:629:href="../references.html#ref-UNICODE">[UNICODE]</a>
> for more details.
> struct/dirlang.html:636:href="../references.html#ref-UNICODE">[UNICODE]</a>,
> table 4-7). In
> struct/global.html:1166:href="../references.html#ref-UNICODE">[UNICODE]</a>
> bidirectional text
> struct/text.html:661:same way as a <A rel="biblioentry"
> href="../references.html#ref-UNICODE">
> struct/text.html:662:[UNICODE]</A> LINE SEPARATOR character.</P>
> struct/text.html:808:="../references.html#ref-UNICODE">[UNICODE]</A>, US
> ASCII, and <A

This is very normative!

I await remaining comments about the informative references. 

Ian


-- 
Ian Jacobs / 401 Second Ave. #19G / New York, NY 10010 USA
Tel/Fax: (212) 684-1814
Email: ibjacobs@panix.com
Received on Sunday, 26 October 1997 17:47:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:43 GMT