Re: [docbook] [docbook-apps] Biblioentry markup standards -- identifying the type of entry

[Peter Flynn <peter@silmaril.ie>]

On 12/06/2020 08:21, Richard Hamilton wrote:
> Thanks Peter, Tony, and Norm for your thoughts,
>
> A lot to consider here.

Indeed. It also impinges a lot on other disciplines, especially Library 
and Archive aspects, which have a different take on what 'a record' 
ought to contain. As a former colleague said "I gathered 40 librarians 
and asked them what I should do, and I got 40 mutually incompatible 
answers." That's not to knock librarians, just to highlight that this 
has been rattling on since the 50s, and doesn't look like being resolved 
any time soon :-)

> - What guidance can these initiatives provide in designing markup
> conventions for biblioentry?

Lots, but long-term.

> - How can these initiatives make it easier to process biblioentrys
> into multiple bibliographic formats?

By minutely documenting the algorithms for selection, placement, 
punctuation, and formatting; in a way that they can be implemented in 
any suitable processing language.

Or by providing algorithms for the transformation into a language that 
already knows all about reformatting into different formats.

> My thoughts are that on the markup side, these projects donât 
> provide much guidance that can be applied directly to biblioentry.

I think you're right; but nor to other XML formats for doing the same 
thing (eg TEI, JATS). Those projects did a LOT of research into the 
bibliographic area, but they were looking at it from an entirely 
different angle.

> The main thing they do is provide help with identifying what
> information needs to be captured for a wide variety of bibliographic
> types, but not much guidance on how to structure that information
> consistently in biblioentry. And, given the variation in how
> biblioentry is used in practice, I believe that some conventions are
> needed if we want to make biblioentry useful.

That would be great.

> I think there is also a pretty good argument for being able to 
> convert biblioentry into BibTeX (as Peter does) and for converting
> BibTeX (or some of the other formats) into DocBook. 

JabRef can export BiBTeX as DocBook 4.4.

> There there are thousands of citations available in some of these
> formats, often easily accessible in web databases. Being able to
> select and convert citations into DocBook would give DocBook users
> access to a huge number of citations 

Most desktop bibliographic database systems can read and write RIS, 
BiBTeX, and EndNote (XML).

> (I searched around in one database and found one of the first
> conference papers I ever wrote, in 1979:-)
:-)

> On the processing side, the style information that CSL or BibTeX 
> encode might be used by a program to generate a stylesheet that could
> process biblioentrys into any format that CSL or BibTeX has encoded
> but it doesnât look like an easy job, so the question becomes whether
> it would be used by enough people to justify the effort required to
> write such a program.

[Old] BiBTeX styles (.bst) files are probably not an option as they are 
written in a language used only by the bibtex program.

biblatex, however, is written in LaTeX, so it is at least 
comprehensible, and biblatex already uses some XML in its config and 
runtime directives. But the .lbx files which define a particular style 
(eg APA, MLA, IEEE, etc) are in LaTeX, which is not really parseable 
except by the latex program.

I believe this is a blind alley, but I would be delighted to be proved 
wrong.

> My opinion is that it is worth the effort to create a set of 
> conventions for creating biblioentry elements 

Yes, absolutely. The rules we have created in-house notoriously trample 
all over bits of DocBook in order to make life easier when transforming 
to LaTeX and BiBTeX. Bringing order to chaos would be welcome.

> and for creating programs that would convert between biblioentry and
> one of the common citation standards, 

I'm not clear what this means. None of the common citation standards 
have a file format, only a typographic appearance.

Or did you mean convert biblioentries to other bibliographic formats 
(RIS, EndNote/XML, BiBTeX, etc); for which I suspect lots of code 
already exists.

> but Iâm not convinced itâs worth the trouble to write a generalized
> program to take CLS or BibTeX style definitions and convert them into
> XSL stylesheets to process biblioentries.

No, absolutely a waste of time, IMNSHO.

> I think it would be easier to create modules similar to the ISO690 
> module, hopefully in a consistent manner that could be easily
> modified to support other styles. 

Yes, building on existing work is usually A Good Idea.

> I think you could handle a fairly large number of output formats

I'm still unclear what this means, file formats or common citation 
standards typographic layouts (which would need a file format).

There are only a handful of relevant file formats for bibliographic 
work: those above plus TEI, JATS, a number of libary-based ones 
including MARCxml, and perhaps HTML-with-divs-and-spans-and-classes.

But there is an infinity of bibliographic reference formats: basically 
every book publisher and journal publisher on the planet has seen fit to 
create their own incompatible layout. On my system, there are 360 unique 
biblatex .lbx files and 425 unique old bibtex .bst files.

> before you would come close to the amount of time it would take to
> create a reliable generalized program, especially considering the
> issues that Norm found with CSL. If there were a bunch of people
> clamoring for a whole bunch of different formats, that might be a
> reasonable course, but after 20+ years of DocBook development, that
> doesnât seem to be the case.

Yes, the most common reference formats probably number a dozen or so. 
But they are a nightmare to program: the work on APA and MLA still isn't 
complete in any language. It reminds me of the work to create the STIX 
fonts for mathematics: no sooner than there is a new release, someone 
invents a new notation.

> P.S. Regarding my original question, I like Norm's idea to allow
> pubwork on biblioentry.

I think he had it on citetitle.

On 11/06/2020 09:12, Norman Tovey-Walsh wrote:

> > I think the title in a bibliography entry is more semantically a
> > title citation than a title, so thatâs my preference anyway.
Norm, I didn't understand that. I think the title in a biblioentry is 
expected to be what is printed on the front cover of the book. You can 
cite it as something else, but a title element type seems to me to be 
pretty much intended as THE title. But perhaps ceci n'est pas une pipe.

<biblioentry xml:id="walsh" role="book">
  <title>DocBook 5</title>
  <subtitle>The Definitive Guide</subtitle>
  <citetitle>The DuckBook</citetitle>
  <author>
    <personname>
      <firstname>Norman</firstname>
      <surname>Walsh</surname>
    </personname>
  </author>
  ...
</biblioentry>

> > (In DocBook, a <title> is usually the the title of a thing, and
> > thatâs not what is going on in a bibliography.)

What *is* going on, then? I seem to have missed something, as usual.

Peter