Re: [Axiom-developer] Re: LatexWiki

[C Y]

--- Bob McElrath <address@hidden> wrote:

> C Y address@hidden wrote:
> > Bob, have you had a chance to look at the axiom bibliography stuff
> > I put up?  Would it be at least compatible with your web based 
> > ideas?
> 
> I noticed it but didn't have a chance to look closely.  Looking now,
> I don't really understand what you're trying to do.  You've done some
> transformations on axiom.bib?

Yes.  In essence, I want to be able to view the entire contents of the
bib file in the context of mathematical coverage - i.e. in that
bibliography, what are the papers that pertain to a given topic?  It's
a measure of whether Axiom has considered a particular topic before,
and a good way to check on the "depth" of the coverage we have.

> Why can't you just use latex and bibtex as usual on a literate
> document?

I would prefer to keep axiom.bib as a valid bibliography file, rather
than a literate document, for simplicity in processing.  Bibtex is
after all a type of TeX, and I don't see any particular advantage to
the literate programming structure for a bibliography - the annotation
field can contain our comments, and other than that the only content in
the document is the bibliography entries themselves.  This way, all we
have to do is put new, more or less standard entries into axiom.bib,
and annotate them if we have anything to say about them.

> All it should require is to run noweb, latex, bibtex, latex in that
> order.  Am I being dense?

For pamphlet files, yes.  I want to go one step beyond that - I want
the bibliography itself to become an organized volume.  So axiom.bib
will work as normal for pamphlets - the one difference being the
axiom.bst file (which, with the help of eprint.sty can produce pdfs
linking to arXiv, for example.)  But I don't want the bibliography to
just show up in individual pamphlets, although that is its primary
function - I want it to be organized as a whole.

> From my perspective, the axiom.bib is not of any interest by itself,
> except as it can cause citations[1] to show up in documents.  As
> such, axiom.bib should ideally contain every document ever written.

Agreed that is its primary function.  However, to me the axiom.bib file
is of interest in and of itself.  Perhaps that's just me?

> Bibtex ignores unused bibliography entries.  Why do you want to
> process axiom.bib?

Because it is the best way I am aware of for someone to evaluate the
quality of the background work put into various subjects in Axiom - as
we turn more and more pamphlets into real literate programs we should
build up a bibliography file containing all the significant works in
the particular sub-disiciplines.  Someone could look at the
Bibliography volume in the MSC subject [foo], and point out relevant or
important work we haven't yet considered.  That's a bit difficult just
by looking at individual pamphlet files - there could be a lot of them
covering individual aspects of a subject.  I sort of expect many
subjects will have a "core" file with the basics and then subsequent
pamphlets with more advanced/newer research as time goes by.

>  I think the processing you're interested in could be
> accomplished with a .bst file.  If one wants to store annotations
> about a paper, that's an interesting thing to do.  But I don't think 
> a bibliography is the place to do it.  (It makes sense though if I
> envision the .bib file as a generalized database that will be
> processed in different ways)

Exactly.  That's how I think of the bib file.  If we use something like
bibtool we can do a lot of transformations on the (very large) main bib
file to get useful answers - say, for example, we want to know how many
papers by author X have been used in Axiom.  That's a one-liner with
bibtool, if I understand the technology correctly, and the output is
itself a valid bib file.  The annote field defined in the bst file to
me is the preferable way to go - that way axiom.bib is always a valid
bibtex file and can be easily cited in other work. Personally I'd
rather generate the latex framework to house the bibliography according
to msc2000 and pacs and sort the bibliography into it, than stick
everything into a pamphlet.  If we do a pamphlet file and still want to
use msc2000, then we have to put things in the correct places by hand. 
I'd rather just append new entries to axiom.bib for immediate use and
let the computer sort it out when the time comes, but that' s just me. 


Cheers,
CY



__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com