--On Saturday, May 07, 2011 10:46 -0700 Dave CROCKER
<[email protected]> wrote:
>
>
> On 5/7/2011 12:12 AM, Julian Reschke wrote:
>> If this is used a lot, we should consider adding it. Is it?
>
>
> Interesting. I've never seen that convention in regular
> publications.
At some level, that is irrelevant. It is in use in RFCs and
has, IIR, rather a long history. Also see the STD and BCP
comments below.
And, of course, we may have different definitions of "regular
publications", different selections of what we read and see, or
even different ideas about what is included in "that
convention". In particular, as I said in an earlier note, I
find trying to use <seriesInfo> as a way to sneak unstructured
text into a reference problematic from the standpoint of generic
markup and possible structuring and external indexing; others
may not care and may, as a consequence, consider this
category/convention much smaller than I do.
While I'm not a big fan of the Chicago Manual, the current RFC
Editor Production Center staff likes it, so let's start with it
as a norm. The current edition has, IIR, shifted much more
strongly to the assumption that materials will be written with
citations anchors ("note reference numbers") pointing to notes
that point to bibliographic entries rather than using the RFC
convention (and their earlier one) of citations pointing
directly to bibliographic entries ("references") and overloading
the latter. But the assumption isn't new. I could quote other
style guides, or give you examples from particular publications,
but the style guide is general and the Chicago one is what the
RFC Editor is trying to use, so... (All citations below refer
to the current (16th) edition. )
(1) "...more than one note reference should never occur at a
single location (such as ⁵,⁶)..." (in the event that you
can't render the last three characters before the right
parenthesis, they are superscript-5, comma, and superscript-6,
i.e., what we would show in RFCs as "[5,6]" or maybe "[5],[6]").
The manual does cite the AMA referencing rules as an exception
to this principle. (Section 14.23)
Note that, if one isn't going to have multiple citation
anchors ("note references") at a single point, and is going to
combine notes and (bibliographic) references as the RFC Series
generally does, there is no alternative other than making those
references compound.
Also, our occasional efforts to say things equivalent to
"[Section 1.1 of RFC9999]" --which is also not possible with
xml2rfc and results in a collection of ad hoc solutions-- is
ultimately a way to get around the convention that specific
section and page number references appear in notes and not in
bibliographies.
(2) "...The number of note references in a sentence or paragraph
can sometimes be reduced by grouping several citations in a
single note..." followed by a rather long example of multiple
references separated by semicolons. (Section 14.52)
(3) Book chapters and the like are covered in Section 14.112,
including specifically the 'Author1, Author2, "Chapter Name". In
_Title_, BookAuthor1 and BookAuthor2...' form. Section 14.113
explicitly discusses listing the book once and then
cross-referencing it from the bibliographic entries for various
chapters.
(4) Section 14.119 gives several examples of compound references
for the purposes of properly identifying citations of reprint
editions and multiple-source materials.
> As for using an STD citation, the STD is a single citation.
> The fact that it's "purpose" is as an indirect pointer to a
> set of documents probably does not put it in a class of its
> own: is true for /most/ citations. Most of what we cite
> contain points to other references.
>
> That an STD is 'pure' in the sense of having no content other
> than those multiple citations does make it look interesting,
> but there are other such documents that primarily serve to
> aggregate references.
I don't immediately see how you would turn this into practice,
and your "other such documents ... serve to aggregate
references" comment identifies the problem I'm having. If there
actually were such a thing as an STD "document", even if all it
did was to aggregate references, then the bibliographic entry
could look like:
<reference anchor="STD9999">
<front>
<title>Warping the Woof</title>
<author><organization> IESG </organization></author>
<date year="2011" month="February" day="30"/>
</front>
<seriesInfo name="STD" value="9999"/>
</reference>
But we don't have a title, we don't have an agreed-upon author
or organization, and we don't have a date. The list in
ftp://ftp.rfc-editor.org/in-notes/std-index.txt seems to assume
that the author and title of the numerically-first RFC in an STD
package is the name of the STD, but that is more than a little
irrational from a bibliographic standpoint and certainly is not
a convention to which the community (or the IAB, or any RFC
Editor-related group in my memory) has ever agreed. The
closest we could get in xml2rfc, even after a change or two to
the slightly-mythical Style Guide, would be:
<reference anchor="STD0010">
<front/>
<seriesInfo name="STD" value="10"/>
<seriesInfo name="RFC" value="821"/>
<seriesInfo name="STD" value="1869"/>
<seriesInfo name="STD" value="974"/>
</reference>
I'm fairly sure that wouldn't compile and more sure that it
would produce something that it would take an expert to
interpret if it did compile. So, if you take the position that
an STD should be a single citation (with which I agree), how
would you propose constructing the reference for it without some
change, either in the RFC formatting rules, to xml2rfc, or both?
best,
john
_______________________________________________
yam mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/yam
