Hi Felipe,
On Jan 2, 2006, at 2:46 AM, Felipe Csaszar wrote:
I only have one disagreement regarding the design of the system. I
believe your system would be much more powerful, simpler, and useful
if it sticks to the Bibtex file format rather than implementing a new
one.
Thanks for your comments, but I'm afraid you're not going to convince
me on this one. I think BibTeX was fine 15 years ago, but we can do a
lot better today.
The benefits of this are many:
- The work is already done. This is a working format, thoroughly
tested, with tens or hundreds of thousand users. In fact it is the de
facto standard for serious academic writers.
That's *highly* debatable; all of it. Talk to users who do
multi-lingual work, or who work in the humanities of law, and who have
to live in a world where interoperability with MS Word is crucial.
Likewise, talk to developers who have to implement parsers for BibTeX
(try Taco Hoekwater for one, who authored the bibliographic module for
ConTeXt).
- Processing Bibtex format it is faster than XML.
And highly bug-prone, not to mention the BibTeX data model is
poorly-designed, and OOo is based on an XML file format. So why would
it make sense to use a totally foreign system?
- It will also be faster to program. No need to call any library or
develop a new one. You can even copy & paste code from other open
source projects that manipulate Bibtex.
I'm afraid it's not so simple. OOo is coded in C++. You can't just
bolt completely different code into it.
And these are the two most important reasons:
- You would simplify the life of people that start writing a paper in
OpenOffice and when the complexity is high move to Latex. No need to
translate the bibliography. In fact, most people would probably have
just one file with all their citations for all their papers. It is a
great simplification. There are users that need to move back and forth
between Latex and a word processor.
And what about the hundreds of thousands of students and scholars who
won't use BibTeX? In my field, I have never met anyone who uses LaTeX
(except, occasionally, me ;-)).
- There are many systems and websites developed that as part of their
work translate bibliographies to Bibtex. Many of the online reference
services (like citeseer) or even libraries, already output Bibtex. For
all these third parties to move to the new format it would take
forever.
Finally, more philosophical arguments:
- It is against the spirit of open source to fork. Especially when the
previous solution is good.
As I said: it's not "good." I literally find it unusable for my needs.
And our goal is to also accommodate users who are coming from
commercial applications like Endnote and Procite, which work
differently than BibTeX (namely, they have richer data models).
- This is the opportunity to strengthen the position of Bibtex.
To be honest, my wish is that BibTeX would go away. Look at the number
of projects in the TeX world that have attempted to replace it. That's
a hint that all is not well with BibTeX.
You would contribute to living in a simpler world. No need to have
several
backends. Just one, plain text and terse, very in the spirit of Unix.
I'm actually not opposed to the notion of terse text formats for bib
metadata. I just don't think BibTeX is a very good one.
- Even if the Bibliographic project develops good import and export
filters (which will mean extra work). The chore of importing,
exporting, and synchronizing, will be costly for the end users.
The work is already done for us:
http://www.scripps.edu/~cdputnam/software/bibutils/bibutils.html
And again, I realize this is difficult to see if you're coming from
certain science fields, but BibTeX is actually not that widely used,
and there are a whole host of other important formats out there (RIS,
pubmed, MARC, MODS, DC, etc.).
Sorry for the long posting. But I think this is the right moment to
reconsider the formats decision.
I appreciate the post, even if we have different perspectives :-)
I congratulate you again for the project and wish you luck whatever
you decide.
Thanks!
Bruce
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
