Re: zotero (or mendeley) integration with org

Re: zotero (or mendeley) integration with org

[Rustom Mody]

Stephen Eglen wrote:
> I've not yet switched to a pdf manager (they're all stuffed into a folder, with a few subfolders, and the only meta-data is
> in the filename!), so I'd appreciate hearing what others to do to look after their pdfs.

Maybe look at tracker [assuming linux] http://projects.gnome.org/tracker/ ?
More generally http://www.wikinfo.org/index.php/Comparison_of_desktop_search_software

Re: Re: zotero (or mendeley) integration with org

[Suvayu Ali]

On Sun, 27 Mar 2011 08:42:19 +0530
Rustom Mody <rustompmody@gmail.com> wrote:

> Stephen Eglen wrote:
> > I've not yet switched to a pdf manager (they're all stuffed into a
> > folder, with a few subfolders, and the only meta-data is in the
> > filename!), so I'd appreciate hearing what others to do to look
> > after their pdfs.
> 
> Maybe look at tracker [assuming linux]
> http://projects.gnome.org/tracker/ ? More generally
> http://www.wikinfo.org/index.php/Comparison_of_desktop_search_software
> 

or Recoll (on GNU/Linux)

-- 
Suvayu

Open source is the future. It sets us free.

zotero (or mendeley) integration with org

[Stephen Eglen]

There was a mail-thread lastyear about zotero and integration with org.
Now that there is an alpha release of 'org-standalone' 
  http://www.zotero.org/blog/2011/02/

has anyone looked at whether this helps integrate org and zotero?

I've not yet switched to a pdf manager (they're all stuffed into a
folder, with a few subfolders, and the only meta-data is in the
filename!), so I'd appreciate hearing what others to do to look after
their pdfs.  Mendeley is a possibility too (although syncing between
machines is a must, and Mendeley doesn't offer that yet.)

Stephen

Re: zotero (or mendeley) integration with org

[Cian]

I use Mendeley and just set it up to export bibtex files. Then you can
just reference those in org-mode using reftex. I think you can do
something similar using zotero.

It works okay. I don't annotate PDFs using Mendeley (I just write my
notes in org-mode), so I'm not sure if you can access those somehow?
But as a system for managing PDFs it's adequate.

On Sat, Mar 26, 2011 at 3:47 PM, Stephen Eglen
<S.J.Eglen@damtp.cam.ac.uk> wrote:
> There was a mail-thread lastyear about zotero and integration with org.
> Now that there is an alpha release of 'org-standalone'
>  http://www.zotero.org/blog/2011/02/
>
> has anyone looked at whether this helps integrate org and zotero?
>
> I've not yet switched to a pdf manager (they're all stuffed into a
> folder, with a few subfolders, and the only meta-data is in the
> filename!), so I'd appreciate hearing what others to do to look after
> their pdfs.  Mendeley is a possibility too (although syncing between
> machines is a must, and Mendeley doesn't offer that yet.)
>
> Stephen
>
>

Re: zotero (or mendeley) integration with org

[Ramon Diaz-Uriarte]

Dear Stephen,

[My ---long--- comments refer only to Mendeley, not org, so maybe this
should be off-list].

On Sat, Mar 26, 2011 at 4:47 PM, Stephen Eglen
<S.J.Eglen@damtp.cam.ac.uk> wrote:
> There was a mail-thread lastyear about zotero and integration with org.
> Now that there is an alpha release of 'org-standalone'
>  http://www.zotero.org/blog/2011/02/
>
> has anyone looked at whether this helps integrate org and zotero?
>
> I've not yet switched to a pdf manager (they're all stuffed into a
> folder, with a few subfolders, and the only meta-data is in the
> filename!), so I'd appreciate hearing what others to do to look after
> their pdfs.  Mendeley is a possibility too (although syncing between
> machines is a must, and Mendeley doesn't offer that yet.)
>

I've been using Mendeley for about 9 months now, after spending a few
weeks examining how Mendeley, Zotero, and some other options, fitted
into my habits. Here are some comments about your questions:

1. Syncing: the "orthodox" Mendeley way allows for two options: a) you
sync the database of the refs (not pdfs) via their servers; b) you
also sync the pdfs, which will most likely require you to pay for
storage in their servers if you have a decent number of pdfs.

However, I do the syncing myself. I've used rsync, then Dropbox, and
for the last four months Wuala. I sync the directory where the dbs are
(~/.local/share/data/Mendeley Ltd./Mendeley Desktop) and the directory
where I store my pdfs. I have had this set up for the 9 months, shared
between four machines, and it works fine (I try not to have Mendeley
open in more than one machine at the same time, to prevent problems,
but have never run into any).


2. All pdfs into a folder with a few subfolders. In Mendeley you can
have each reference under its own subdirectory with some limited
flexibility for the naming of the subdirectories (author, date, title,
etc). Its not ideal (e.g., I dislike spaces in directory names), but I
like it better than Zotero's (names of subdirectories are a random
string). Keeping each ref in its own directory allows me to store
other stuff (e.g., code, suppl. mat, etc) in the same place as the
pdf.


Now, that said, I am not all that happy with Mendeley.

To begin with, Mendeley is not free software. Zotero is, but the
naming of directories and the lack of a built-in pdf editor were a
no-go for me.

Mendeley's pdf editor allows me to underline and add notes to pdfs. At
least in Linux, adding comments and underlining pdfs is not yet well
solved: evince currently allows comments in the devel. version, but no
underlining. Okular allows comments and underlining, but stores them
in ~/.kde/share/apps/okular, which I dislike (it seems very fragile),
and in okular you cannot underline a single column in papers with
multi-column setup (another no-go for me).

However, with Mendeley's pdf editor/viewer you can only display one
pdf at a time, which is a pain if you want to compare two or more
papers side by side. In fact, Mendely's pdf viewer is rather
under-powered compared to, say, okular (e.g., lack of keyboard
shortcuts for viewing to page or width size, difference between
scrolling and moving to next page, fast searching, etc). If I only
need read-only access to the pdf, even from within Mendeley, I use
another pdf viewer.

I thought about going back to JabRef, and using recoll for searching
over all of my pdf collection (I like recoll's search much more than
Mendeley's one). But that leaves unsolved the pdf commenting issue.
Also, I like the ease of adding papers with Mendeley (e.g., when
searching in the web, or its generally successful extraction of
metadata from a paper's pdf); in fact, I find adding papers is even
easier with Zotero (and friendlier also if you add things like web
pages, etc). I also looked into paperpile, but I find it to be much
more beta than Mendeley, and so far less flexible in almost
everything.

So I'd also like to know how others are dealing with pdfs. Anyway, and
in spite of my complaints, since about 9 months ago I have not printed
a single paper. It is great to have all those pdfs synced among my
machines, be able to search them quickly, and not carry around
kilograms of dead trees.


Best,

R.




> Stephen
>
>



-- 
Ramon Diaz-Uriarte
Structural Biology and Biocomputing Programme
Spanish National Cancer Centre (CNIO)
http://ligarto.org/rdiaz
Phone: +34-91-732-8000 ext. 3019
Fax: +-34-91-224-6972

Re: zotero (or mendeley) integration with org

[Stephen Eglen]

Dear all,

Thanks for the informative replies.  I gave zotero a short-test
yesterday,and in general liked what I found, although as it uses sql
databases, it is moves away from my preference for plain text files to
see everything in.   I appreciate that sql may scale better, but I don't
have huge databases.  I think I'll continue to watch how the zotero
standalone develops.  

In the meantime, it made me think a bit more about looking after pdf
collections.  I'll try out using org mode for annotating my pdfs, to see
how that works.


I had not seen recoll before, that certainly looks nice.

Will feedback to the list if I have anything useful to add.

Stephen

Re: zotero (or mendeley) integration with org

[brian powell]

It uses SQLite--Stephen: I'd consider myself a
plaintext-whenever-possible sort of dude too; but,
SQLite (used in ZOTERO) is a simple/short C program and its (last time
I checked) extremely simple--for example there is only "left outer
join".

SQLite "databases" are very easy to work with and are often 1 simple,
small, very portable file.

I was reading some article yesterday: SQLite was ranked #1 in the top
ten best/most useful software of all time.

I understand your concern; and, often an SQL database is overkill and
"NOSQL" seems in vogue right now--but SQLite is something you might
really like--you can manipulate SQLite databases easily, like
plaintext.

I just tried ZOTERO too--very impressive.  I'm a lot more used to
using BibTeX though--which is purely plaintext; but, BibTeX can be a
typing chore!

On Mon, Mar 28, 2011 at 5:14 AM, Stephen Eglen
<S.J.Eglen@damtp.cam.ac.uk> wrote:
> Dear all,
>
> Thanks for the informative replies.  I gave zotero a short-test
> yesterday,and in general liked what I found, although as it uses sql
> databases, it is moves away from my preference for plain text files to
> see everything in.   I appreciate that sql may scale better, but I don't
> have huge databases.  I think I'll continue to watch how the zotero
> standalone develops.

Re: zotero (or mendeley) integration with org

[Erik Hetzner]

[-- Attachment #1: Type: text/plain, Size: 1294 bytes --]

At Sat, 26 Mar 2011 15:47:44 +0000,
Stephen Eglen wrote:
> 
> There was a mail-thread lastyear about zotero and integration with org.
> Now that there is an alpha release of 'org-standalone' 
>   http://www.zotero.org/blog/2011/02/
> 
> has anyone looked at whether this helps integrate org and zotero?
> 
> I've not yet switched to a pdf manager (they're all stuffed into a
> folder, with a few subfolders, and the only meta-data is in the
> filename!), so I'd appreciate hearing what others to do to look after
> their pdfs.  Mendeley is a possibility too (although syncing between
> machines is a must, and Mendeley doesn't offer that yet.)

Hi Stephen,

Re. zotero-plain [1] & integration with Zotero standalone, it would
depend on what external interface Zotero standalone is presenting for
integration with external programs (e.g. the Chrome or Safari
extensions), which is not at all clear to me. zotero-plain currently
depends on mozrepl, which I do not think would distributed with Zotero
standalone.

However, with a few modifications it should be possible to get
zotero-plain to work with the zotero.org web API, which would provide
most of the benefits of Zotero standalone, although you would be
relying on a 3rd party service.

best, Erik

1. http://e6h.org/~egh/hg/zotero-plain/

[-- Attachment #2: Type: text/plain, Size: 53 bytes --]

Sent from my free software system <http://fsf.org/>.

Re: zotero (or mendeley) integration with org

[Matt Lundin]

Stephen Eglen <S.J.Eglen@damtp.cam.ac.uk> writes:

> There was a mail-thread lastyear about zotero and integration with org.
> Now that there is an alpha release of 'org-standalone' 
>   http://www.zotero.org/blog/2011/02/
>
> has anyone looked at whether this helps integrate org and zotero?
>
> I've not yet switched to a pdf manager (they're all stuffed into a
> folder, with a few subfolders, and the only meta-data is in the
> filename!), so I'd appreciate hearing what others to do to look after
> their pdfs.  Mendeley is a possibility too (although syncing between
> machines is a must, and Mendeley doesn't offer that yet.)

One option is to manage metadata in org-mode itself, relying on
org-attach to store and preserve links to the pdf files. Bibtex source
blocks can used to store bibliographical data for each pdf.

I find the combination of emacs-w3m, google scholar, and org-mode to be
an easier and more transparent way to manage bibtex data than an
indirect route via Zotero or Mendeley. But I also prefer to edit all my
bibtex data by hand. :)

Recoll is great for indexing. I have a mess of spaghetti code I use to
pull recoll results into a temporary org outline. I can then open the
relevant files using org links. I'd be happy to share it if anyone is
interested.

Best,
Matt

Re: zotero (or mendeley) integration with org

[William Gardella]

Matt Lundin <mdl@imapmail.org> writes:

> Stephen Eglen <S.J.Eglen@damtp.cam.ac.uk> writes:
>
>> There was a mail-thread lastyear about zotero and integration with org.
>> Now that there is an alpha release of 'org-standalone' 
>>   http://www.zotero.org/blog/2011/02/
>>
>> has anyone looked at whether this helps integrate org and zotero?
>>
>> I've not yet switched to a pdf manager (they're all stuffed into a
>> folder, with a few subfolders, and the only meta-data is in the
>> filename!), so I'd appreciate hearing what others to do to look after
>> their pdfs.  Mendeley is a possibility too (although syncing between
>> machines is a must, and Mendeley doesn't offer that yet.)
>
> One option is to manage metadata in org-mode itself, relying on
> org-attach to store and preserve links to the pdf files. Bibtex source
> blocks can used to store bibliographical data for each pdf.
>
> I find the combination of emacs-w3m, google scholar, and org-mode to be
> an easier and more transparent way to manage bibtex data than an
> indirect route via Zotero or Mendeley. But I also prefer to edit all my
> bibtex data by hand. :)
>
> Recoll is great for indexing. I have a mess of spaghetti code I use to
> pull recoll results into a temporary org outline. I can then open the
> relevant files using org links. I'd be happy to share it if anyone is
> interested.
>
> Best,
> Matt
>
>

I use something like the setup Matt describes as well (except I haven't
played with Recoll).  I use org-attach to keep documents
organized; usually they're attached to a :noexport: top-level heading
creatively called "Documents," which contains the citation info and
keeps my research/reading notes distinct from the paper I'm writing.
Since my org files are themselves indexed and easily searchable, I've
not felt the need for any collection management software beyond that.
(For legal research it's particularly good, because I can get LEXIS to
email me most legal texts as plaintext, which I then just add inline as
subheadings to the Documents heading.)

For bibliography generation, I use RefTeX to do my BibTeXing for me, and
I use a pretty crude one bibliography database to one paper kind of
system.

Best,
Will

-- 
William Gardella
J.D. Candidate
Class of 2011, University of Pittsburgh School of Law

Re: zotero (or mendeley) integration with org

[Rasmus]

> [Matt and William's setup]

I have looked for a good way to keep track of academic papers (pdfs) and
Bibtex for a long time. I'd love to see a worg page on this topic.

Meanwhile, I have found some sweet Bibtex-search interfaces for
Emacs. These will query a academic search engine and can copy Bibtex
entries directly to a .bib file. I found bibsnarfl[fn:1] being the most
interesting, but a similar code is available for PubMed[fn:2].
Unfortunately, being limited to certain fields, I am personally not able
to adopt either. It would be great to have an interface to a general
academic search engine (Google Scholar, ugh?).

Imagine the combination of a Emacs-powered interface to some search
engine, a university network and some magic snip that would download a 
pdf, add it to a .bib-file (removing annoying entries and adding a
sensible key), and making a nice, easy-to-browse Org-file. 

One day, maybe...

–Rasmus

Footnotes:

[fn:1] http://www.emacswiki.org/emacs/bibsnarf.el

[fn:2] http://www.bioinformatics.org/texmed/

Re: zotero (or mendeley) integration with org

[William Gardella]

Rasmus <rasmus.pank@gmail.com> writes:

>> [Matt and William's setup]
>
> I have looked for a good way to keep track of academic papers (pdfs) and
> Bibtex for a long time. I'd love to see a worg page on this topic.
>
> Meanwhile, I have found some sweet Bibtex-search interfaces for
> Emacs. These will query a academic search engine and can copy Bibtex
> entries directly to a .bib file. I found bibsnarfl[fn:1] being the most
> interesting, but a similar code is available for PubMed[fn:2].
> Unfortunately, being limited to certain fields, I am personally not able
> to adopt either. It would be great to have an interface to a general
> academic search engine (Google Scholar, ugh?).
>
> Imagine the combination of a Emacs-powered interface to some search
> engine, a university network and some magic snip that would download a 
> pdf, add it to a .bib-file (removing annoying entries and adding a
> sensible key), and making a nice, easy-to-browse Org-file. 
>
> One day, maybe...
>
> –Rasmus
>
> Footnotes:
>
> [fn:1] http://www.emacswiki.org/emacs/bibsnarf.el
>
> [fn:2] http://www.bioinformatics.org/texmed/
>
>
>
>

That'd be a glorious way to do research.  I can see it happening if a
few of these academic database search engines and library websites
decide to use some kind of free software infrastructure, or at least a
relatively open and consistent API...alas, I don't know if library
science is really evolving in that direction yet.

-- 
William Gardella
J.D. Candidate
Class of 2011, University of Pittsburgh School of Law

Re: zotero (or mendeley) integration with org

[Matt Lundin]

Rasmus <rasmus.pank@gmail.com> writes:

>> [Matt and William's setup]
>
> I have looked for a good way to keep track of academic papers (pdfs) and
> Bibtex for a long time. I'd love to see a worg page on this topic.
>
> Meanwhile, I have found some sweet Bibtex-search interfaces for
> Emacs. These will query a academic search engine and can copy Bibtex
> entries directly to a .bib file. I found bibsnarfl[fn:1] being the most
> interesting, but a similar code is available for PubMed[fn:2].
> Unfortunately, being limited to certain fields, I am personally not able
> to adopt either. It would be great to have an interface to a general
> academic search engine (Google Scholar, ugh?).

Agreed. Google Scholar citations need very close proofreading, as they
can be erroneous or poorly formatted. I would submit that one should
never use a Google Scholar citation without checking it carefully
against the article or book to which it refers. An advantage of Google
Scholar, however, is that it offers skeletal bibtex entries for books
and articles in a wide variety of fields, whereas many of the databases
accessed by bibsnarf are limited to math and sciences. Since I use
biblatex together with the Chicago Manual of Style, any bibtex entry I
clip has to be edited and tweaked substantially. (Indeed, manual editing
is unavoidable when using biblatex.)

FWIW, here's my workflow:

1. Clip a bibtex citation or unformatted bibliographical data using
   org-capture and conkeror or emacs. (This generates a timestamp and a
   link to where I first found the reference --- very useful data for
   reviewing one's own research habits and sources.)
2. Download the pdf, if possible (or make a todo to get/read it later).
3. While in the capture buffer, use org-attach to move the pdf quickly
   from ~/Downloads to the attachments directory (this is much faster
   than it sounds).
4. Create or correct a bibtex source block within the org entry.
5. Make a TODO to read the pdf. :)

Later, when I read the document, I proofread the bibtex entry once again
and call a function that moves it to a centralized bibtex file, leaving
a link in its place. As a rule, any citation that goes into my official
bibtex file *must* be correct and complete. Reftex, another of Carsten's
ingenious creations, is very handy for creating links to the citation
while taking notes.

For my own purposes, there would be little point in automating this
process, as any pdf I download needs to be inspected manually and any
bibtex entry I clip needs to be proofread and tweaked. The process
itself forces me to check every citation for accuracy.

Best,
Matt

Re: zotero (or mendeley) integration with org

[Stephen Eglen]

> Agreed. Google Scholar citations need very close proofreading, as they
> can be erroneous or poorly formatted. 

Thanks Matt - I'd agree with this, having seen oddities from google
scholar.  I emailed them ages ago about one problem (formatting of
initials in author names), but never heard back... it is a pity that
there is no mechanism for tidying up their references, as it seems to be
the best thing out there that covers all the fields.

Having said that, if google scholar can save me some typing, I'll
happilyuse it as a starting point for a bibtex entry.  I've just started
using pdfmeat -- this is nice, as given a pdf, it outputs the
corresponding bibtex entry from google scholar.  Probably works similar
to the way zotero does it, but can be used straight from the command
line:

  http://code.google.com/p/pdfmeat/

(Warning: I couldn't get one of the python dependencies, unidecode, to
work on mac, but it does work on ubuntu for me.)

> accessed by bibsnarf are limited to math and sciences. Since I use
> biblatex together with the Chicago Manual of Style, any bibtex entry I
> clip has to be edited and tweaked substantially. (Indeed, manual editing
> is unavoidable when using biblatex.)

If its not too tangential, why do you use biblatex -- is it the future
for bibtex?

Thanks for summarising your workflow, very helpful.

Stephen

Re: zotero (or mendeley) integration with org

[Matt Lundin]

Stephen Eglen <S.J.Eglen@damtp.cam.ac.uk> writes:

>> Agreed. Google Scholar citations need very close proofreading, as they
>> can be erroneous or poorly formatted. 
>
> Thanks Matt - I'd agree with this, having seen oddities from google
> scholar.  I emailed them ages ago about one problem (formatting of
> initials in author names), but never heard back... it is a pity that
> there is no mechanism for tidying up their references, as it seems to be
> the best thing out there that covers all the fields.
>
> Having said that, if google scholar can save me some typing, I'll
> happilyuse it as a starting point for a bibtex entry.  I've just started
> using pdfmeat -- this is nice, as given a pdf, it outputs the
> corresponding bibtex entry from google scholar.  Probably works similar
> to the way zotero does it, but can be used straight from the command
> line:
>
>   http://code.google.com/p/pdfmeat/
>

Thanks for the link! That looks like a useful tool.

>> accessed by bibsnarf are limited to math and sciences. Since I use
>> biblatex together with the Chicago Manual of Style, any bibtex entry I
>> clip has to be edited and tweaked substantially. (Indeed, manual editing
>> is unavoidable when using biblatex.)
>
> If its not too tangential, why do you use biblatex -- is it the future
> for bibtex?

I use biblatex because I use citation styles in the humanities
(especially the Chicago Manual of Style). Biblatex and the chicago-notes
package (both now part of TeXLive) handle Chicago Style footnotes and
bibliographies beautifully, with an astounding number of options and
flawless formatting -- but the bibtex entries are a bit fussier than
standard bibtex.

Best,
Matt

Re: Re: zotero (or mendeley) integration with org

[Alan E. Davis]

[-- Attachment #1: Type: text/plain, Size: 2158 bytes --]

Is it possible to use org babel to extract bibtex entries from file of notes
to a *.bib file?

The stumbling point for me in saving bibtex sources is I don't see a way to
use the file as a bibtex *.bib file so as to use that as the direct source
for the publication.  Perhaps this could be automated with babel?

Alan

On Fri, Apr 1, 2011 at 6:13 AM, Matt Lundin <mdl@imapmail.org> wrote:

> Stephen Eglen <S.J.Eglen@damtp.cam.ac.uk> writes:
>
> >> Agreed. Google Scholar citations need very close proofreading, as they
> >> can be erroneous or poorly formatted.
> >
> > Thanks Matt - I'd agree with this, having seen oddities from google
> > scholar.  I emailed them ages ago about one problem (formatting of
> > initials in author names), but never heard back... it is a pity that
> > there is no mechanism for tidying up their references, as it seems to be
> > the best thing out there that covers all the fields.
> >
> > Having said that, if google scholar can save me some typing, I'll
> > happilyuse it as a starting point for a bibtex entry.  I've just started
> > using pdfmeat -- this is nice, as given a pdf, it outputs the
> > corresponding bibtex entry from google scholar.  Probably works similar
> > to the way zotero does it, but can be used straight from the command
> > line:
> >
> >   http://code.google.com/p/pdfmeat/
> >
>
> Thanks for the link! That looks like a useful tool.
>
> >> accessed by bibsnarf are limited to math and sciences. Since I use
> >> biblatex together with the Chicago Manual of Style, any bibtex entry I
> >> clip has to be edited and tweaked substantially. (Indeed, manual editing
> >> is unavoidable when using biblatex.)
> >
> > If its not too tangential, why do you use biblatex -- is it the future
> > for bibtex?
>
> I use biblatex because I use citation styles in the humanities
> (especially the Chicago Manual of Style). Biblatex and the chicago-notes
> package (both now part of TeXLive) handle Chicago Style footnotes and
> bibliographies beautifully, with an astounding number of options and
> flawless formatting -- but the bibtex entries are a bit fussier than
> standard bibtex.
>
> Best,
> Matt
>
>

[-- Attachment #2: Type: text/html, Size: 2882 bytes --]

Re: zotero (or mendeley) integration with org

[Matt Lundin]

"Alan E. Davis" <lngndvs@gmail.com> writes:

> Is it possible to use org babel to extract bibtex entries from file of
> notes to a *.bib file? 
>

Yes, you can use babel's tangling facilities to extract bibtex entries
enclosed in source blocks.

#+begin_src bibtex :tangle history.bib
  @InCollection{levi2001_microhistory,
    author =       {Levi, Giovanni},
    title =        {On Microhistory},
    editor =       {Peter Burke},
    booktitle =    {New Perspectives on Historical Writing},
    address =      {University Park, PA},
    publisher =    {Penn State Press},
    year =         2001,
  }
#+end_src

Calling org-babel-tangle-file will put this entry in the file
history.bib.

> The stumbling point for me in saving bibtex sources is I don't see a
> way to use the file as a bibtex *.bib file so as to use that as the
> direct source for the publication.  Perhaps this could be automated
> with babel?

There are a few other routes. 

1. Bibtex discards anything outside of an entry so you could symlink
   your org file to something with a bib extension (e.g., notes.org ->
   notes.bib) and simply point latex/bibtex to that file.

2. You could instruct emacs to edit bib files with orgmode and use babel
   and source blocks to enter items
   (add-to-list 'auto-mode-alist 
	     '("\\.\\(org\\|bib\\)$" . org-mode))

3. AFAICT, reftex is blissfully indifferent to non-bibtex data, so you
   could use reftex to query your org files for citation keys. Then, you
   could use reftex to generate a proper bib file with all entries cited
   in your paper.

Best,
Matt