Re: m-bib, xml, etc.

Re: m-bib, xml, etc.

[Taco Hoekwater]

Hi all,

See the questions by Bruce included below. Before I get to actually
doing something with it, I think it is best to outline my plans with
the bibliography module.

1. I intend to switch to using XML as input format, using the 
bibteXML DTD (loose format) as definition file. 

See http://bibtexml.sourceforge.net for details of this project. 

A big advantage of using this is that we can run piggyback on
their bibtex/marc conversion tool :)

I'm not sure whether to keep the current syntax. Advantage:
easy backward compatibily. Downside: more complex code and
possible confusion. Personally, I prefer to drop the support 
for the old syntax altogether (perhaps providing a conversion
script?)

2. Add support for citations with tagged data as requested
by Bruce (and previously by other people). 

3. At the same time, drop support for latex \cite syntax.
(At that time it is no longer needed), and go for a more
standard context syntax using 

	\cite[<id>][<type> or <option=value>]

4. create a bibl-XXX definition wizard (same functionality
   as makebst)

On Thu, 18 Jul 2002 22:56:12 -0400 (EDT), Bruce wrote:

> 
> Hi Taco,
> 
> I've been searching for information about m-bib as I'm trying to get it 
> working, and came across this very interesting post of yours regarding 
> your plans to replace bibtex with an xml-based system.  First, though, 
> two simple questions:
> 
> 1) Is it now possible to have a citation in the form (see Smith, 1990, 
> page 34)?
> 2) I was led to believe I can use any bst I would use with LaTeX with 
> ConTeXt.  But reading the m-bib doc suggests this isn't the case. Can 
> you please explain?  One of the nice things of the BibTeX world is 
> makebst, and I'd like to be able to use bsts generated with it.
> 
> Now to the bigger question regarding the below.  Any update on the 
> status of this shift to xml?
> 
> I ask because I currently use Word and Endnote.  Because I am in the 
> social sciences, no publisher will accept docs in TeX, so I want to 
> find a simple way that I can use bib data and documents to produce pdf 
> via LaTeX and/or ConTeXt and something that Word can understand (html 
> and/or rtf; but am told Word can read xml, but don't know what this 
> means in practice).  I still can't figure out how to do this though.  
> Any thoughts/updates/ideas would be much appreciated!
> 
> Bruce
> 
> 
> From: Taco Hoekwater <taco@elvenkind.com> 
> Subject: Re: m-bib: passing extra information to \cite 
> Date: Fri, 07 Sep 2001 14:48:06 +0200 
> 
> ------------------------------------------------------------------------
> --------
> 
> Hans Hagen wrote:
> > 
> > \definecomplexorsimple\cite
> > 
> > now \cite[bla] and \cite{bla} both work (but i favor the [] only 
> approach).
> 
> Roughly speaking, this is what happens now. But soon we will probably 
> have
> calling conventions like \cite[...][...][..=..] or \cite 
> {..}[..][..]{..}
> (more or less the same as for \in), and supporting the latex syntax is
> getting problematic. Besides that, there is usually other stuff wrong
> with bibtex (latex) databases, like commands that are defined in the
> preamble using \newcommand; \text... commands in the text; etc.
> 
> Eventually, I plan to phase out bibtex alltogether, and replace it with
> an XML input filter (with a formalized DTD).
> 
> -- 
> groeten,
> 
> Taco
> 
> 

-- 
groeten,

Taco

Re: m-bib, xml, etc.

[Taco Hoekwater]

On Tue, 6 Aug 2002 13:37:55 -0400 (EDT), Bruce wrote:

> In general, any new xml-based system should not inherit the limitations 
> of previous approaches.  What I am about to say may indicate some 
> ignorance of BibTeX, with which I am not intimately familiar, so take 
> it for it's worth (and correct me if I'm wrong!):

True but also applicable to RIS. Desperately missing from both
bibtex and endnote/reference manager file format are character sets.

> Finally, my biggest issue is this:
> 
> Exactly how do I write documents such that I can easily produce not 
> just beautiful pdf output via pdftex, but also some format that word 
> processors (Word) and dtp apps can import?  It doesn't seem tagged pdf 
> is coming anytime soon.  In theory, a ConTeXt to html script would 
> work, but I'm not aware of this coming soon either.  Finally, given 
> support for xml in ConTeXt and your interest in moving m-bib in this 
> direction, I suppose I could just write in xml.

I don't believe I can answer the general question. For bibligraphic
data the solution is rather straightforward: I'll just write some perl 
code to go from XML to RIS/bibtex & vv.

> OK, but adding RIS (or refer) import to m-bib would really help me out, 
> as BibTeX export from an app like Endnote is less-than-ideal!

We ware working on this issue off the list. If someone wants to be in on
the discussion, please e-mail me.

greetings, Taco

-- 
groeten,

Taco

Re: m-bib, xml, etc.

[Bruce D'Arcus]

Taco Hoekwater wrote:

>On Mon, 5 Aug 2002 11:05:04 -0400 (EDT), Bruce wrote:
>> 
>> Sounds great!  I can't comment on any technical issues because I don't 
>> know much about xml.  I do, however, know what I need in a bib system, 
>> and so will be happy to give input going forward.  It sounds like I 
>> ought to wait a bit before moving my fairly complex bib database off 
of 
>> Endnote though, and go with xml rather than bibtex.  
>
>You can assume that the XML bibtex format is roughly
>the current bibtex functionality + explicit tagging for
>the parts of an author's name, separated keywords, and
>support for subtitles. Just think about what features you
>miss in bibtex and write those down, that will be quite
>helpful.

In general, any new xml-based system should not inherit the limitations 
of previous approaches.  What I am about to say may indicate some 
ignorance of BibTeX, with which I am not intimately familiar, so take 
it for it's worth (and correct me if I'm wrong!):

I am a social scientist whose work is basically historial.  So I often 
have record types that are not explcitly supported in BibTeX.  Archival 
records are easy enough to deal with I suppose (I can just use fields 
like "where published" to indicate archives, collections, box numbers, 
etc.), but what about an archival document that is a memo.  In my 
dissertation, the bib entry might look like this:

Smith J. to N. Scott (1973) memo.  Smith files, box 23, folder 2.

I have no idea how to do this in BibTeX.  I imagine I could cite this 
differently though and get around the problem.

The other issue is citing in footnotes (or also common, endnotes), in 
which the first citation lists the entire reference, and subsequent 
ones shortened versions.

Finally, my biggest issue is this:

Exactly how do I write documents such that I can easily produce not 
just beautiful pdf output via pdftex, but also some format that word 
processors (Word) and dtp apps can import?  It doesn't seem tagged pdf 
is coming anytime soon.  In theory, a ConTeXt to html script would 
work, but I'm not aware of this coming soon either.  Finally, given 
support for xml in ConTeXt and your interest in moving m-bib in this 
direction, I suppose I could just write in xml.

Any insight on this issue--which is absolutely essential for me--would 
be much appreciated...

>> Any sense of timeframe here Taco?
>
>Couple of weeks I think, it has taken long enough by now ;)

Wow; great!

>> Also, has anyone taken a look at refDB?  
>> 
>> see http://refdb.sourceforge.net
>> 
>> It's a bib system built on MySQL that processes both Docbook and LaTeX 
>> documents.
>
>The XML files that define the formatting stylesheets are quite 
>interesting. I think I will adopt these instead of the current
>TeX syntax. 
>
>Their input format for bibliographic data is not XML though,
>but RIS. I'm not overly please with that, and using docbook or 
>tei markup for bibliographic databases is probably a mistake, 
>so I vote for sticking with the bibtex-DTD.

OK, but adding RIS (or refer) import to m-bib would really help me out, 
as BibTeX export from an app like Endnote is less-than-ideal!

Re: m-bib, xml, etc.

[Taco Hoekwater]

On Mon, 5 Aug 2002 11:05:04 -0400 (EDT), Bruce wrote:
> 
> Sounds great!  I can't comment on any technical issues because I don't 
> know much about xml.  I do, however, know what I need in a bib system, 
> and so will be happy to give input going forward.  It sounds like I 
> ought to wait a bit before moving my fairly complex bib database off of 
> Endnote though, and go with xml rather than bibtex.  

You can assume that the XML bibtex format is roughly
the current bibtex functionality + explicit tagging for
the parts of an author's name, separated keywords, and
support for subtitles. Just think about what features you
miss in bibtex and write those down, that will be quite
helpful.

> Any sense of timeframe here Taco?

Couple of weeks I think, it has taken long enough by now ;)

> Also, has anyone taken a look at refDB?  
> 
> see http://refdb.sourceforge.net
> 
> It's a bib system built on MySQL that processes both Docbook and LaTeX 
> documents.

The XML files that define the formatting stylesheets are quite 
interesting. I think I will adopt these instead of the current
TeX syntax. 

Their input format for bibliographic data is not XML though,
but RIS. I'm not overly please with that, and using docbook or 
tei markup for bibliographic databases is probably a mistake, 
so I vote for sticking with the bibtex-DTD.

-- 
groeten,

Taco

Re: m-bib, xml, etc.

[Bruce D'Arcus]

Sounds great!  I can't comment on any technical issues because I don't 
know much about xml.  I do, however, know what I need in a bib system, 
and so will be happy to give input going forward.  It sounds like I 
ought to wait a bit before moving my fairly complex bib database off of 
Endnote though, and go with xml rather than bibtex.  

Any sense of timeframe here Taco?

Also, has anyone taken a look at refDB?  

see http://refdb.sourceforge.net

It's a bib system built on MySQL that processes both Docbook and LaTeX 
documents.

Bruce

Taco Hoekwater wrote:

>Hi all,
>
>See the questions by Bruce included below. Before I get to actually
>doing something with it, I think it is best to outline my plans with
>the bibliography module.
>
>1. I intend to switch to using XML as input format, using the 
>bibteXML DTD (loose format) as definition file. 
>
>See http://bibtexml.sourceforge.net for details of this project. 
>
>A big advantage of using this is that we can run piggyback on
>their bibtex/marc conversion tool :)
>
>I'm not sure whether to keep the current syntax. Advantage:
>easy backward compatibily. Downside: more complex code and
>possible confusion. Personally, I prefer to drop the support 
>for the old syntax altogether (perhaps providing a conversion
>script?)
>
>2. Add support for citations with tagged data as requested
>by Bruce (and previously by other people). 
>
>3. At the same time, drop support for latex \cite syntax.
>(At that time it is no longer needed), and go for a more
>standard context syntax using 
>	
>	\cite[<id>][<type> or <option=value>]
>
>4. create a bibl-XXX definition wizard (same functionality
>   as makebst)
>
>
>On Thu, 18 Jul 2002 22:56:12 -0400 (EDT), Bruce wrote:
>
>> 
>> Hi Taco,
>> 
>> I've been searching for information about m-bib as I'm trying to get 
it 
>> working, and came across this very interesting post of yours regarding 
>> your plans to replace bibtex with an xml-based system.  First, though, 
>> two simple questions:
>> 
>> 1) Is it now possible to have a citation in the form (see Smith, 1990, 
>> page 34)?
>> 2) I was led to believe I can use any bst I would use with LaTeX with 
>> ConTeXt.  But reading the m-bib doc suggests this isn't the case. Can 
>> you please explain?  One of the nice things of the BibTeX world is 
>> makebst, and I'd like to be able to use bsts generated with it.
>> 
>> Now to the bigger question regarding the below.  Any update on the 
>> status of this shift to xml?
>> 
>> I ask because I currently use Word and Endnote.  Because I am in the 
>> social sciences, no publisher will accept docs in TeX, so I want to 
>> find a simple way that I can use bib data and documents to produce pdf 
>> via LaTeX and/or ConTeXt and something that Word can understand (html 
>> and/or rtf; but am told Word can read xml, but don't know what this 
>> means in practice).  I still can't figure out how to do this though.  
>> Any thoughts/updates/ideas would be much appreciated!
>> 
>> Bruce
>> 
>> 
>> From: Taco Hoekwater <taco@elvenkind.com> 
>> Subject: Re: m-bib: passing extra information to \cite 
>> Date: Fri, 07 Sep 2001 14:48:06 +0200 
>> 
>> 
------------------------------------------------------------------------
>> --------
>> 
>> Hans Hagen wrote:
>> > 
>> > \definecomplexorsimple\cite
>> > 
>> > now \cite[bla] and \cite{bla} both work (but i favor the [] only 
>> approach).
>> 
>> Roughly speaking, this is what happens now. But soon we will probably 
>> have
>> calling conventions like \cite[...][...][..=..] or \cite 
>> {..}[..][..]{..}
>> (more or less the same as for \in), and supporting the latex syntax is
>> getting problematic. Besides that, there is usually other stuff wrong
>> with bibtex (latex) databases, like commands that are defined in the
>> preamble using \newcommand; \text... commands in the text; etc.
>> 
>> Eventually, I plan to phase out bibtex alltogether, and replace it with
>> an XML input filter (with a formalized DTD).
>> 
>> -- 
>> groeten,
>> 
>> Taco
>> 
>> 
>
>
>-- 
>groeten,
>
>Taco
>