ANN: refdb-0.9.8-dupcheck-pre1 available for testing

Hi,

here's the first prerelease from the dupcheck branch of refdb, built
from repository revision 191:

http://refdb.sourceforge.net/pre/refdb-0.9.8-dupcheck-pre1.tar.gz

This prerelease is identical with 0.9.8-pre4, except for the new
checkref command which I'll explain shortly. I'd appreciate bug
reports and success stories with the dupcheck branch, as I'd like to
merge this code back to the trunk as soon as it turns out to be
functional and stable.

You don't need to install this prerelease in order to test it if you
have 0.9.8-pre4 installed. The new prerelease works with the same
databases and does not alter them when running checkref. All you need
to do is to download the tarball, configure and build it as usual, and
then run the server (./refdbd -s <other options>) and the client
(./refdbc <options>) from the src directory. Again, "make install"
should not be required, and I wouldn't recommend it at this time. If
you find some time, please export your references from your existing
database and add them to a new database using the new prerelease, then
query, update, and delete to your heart's content to check that
everything is working ok.

Now what is that new refdbc: checkref command? It is a simple way to
check your reference data for duplicates and faulty data before adding
them to your database permanently. The data will be added to temporary
tables in your database, using the same functions as when adding data
permanently. If your input data are somehow screwed up, you'll notice
this without altering your database. Once the tables are filled,
refdbd tries to find references in the permanent tables that are
likely to be duplicates of the ones in the temporary tables. Currently
the titles and the location (periodical, volume, issue, start page,
publication year) are compared which should work ok for books and
journal articles. Currently only identical matches are detected, but
more refined queries may be added later. refdbd also checks the
citation keys. If a citation key is already taken, you may choose to
specify a different one instead of letting refdbd create keys like
Miller1999c and so on.

In addition refdbd tries to find pairs of author names with full and
abbreviated first names which may point to the same person. If you get
a match, you may replace the abbreviated version with the full name,
as refdbd can always generate the initials from the full name, but
not vice versa.

The syntax of the checkref command is identical with the addref and
updateref commands. Currently there is only screen-readable output,
but other backends may be added later. I'd also like to get input from
people fiddling with RefDB frontends about how to return the data best
for graphical interfaces.

Finally, here is an example:

refdbc: checkref checkref.ris
406:1:Russwurm2002
408:1
406:2:Cerwinka2002
408:2
AUTHOR ABBREV: temp_id:1 temp_author:Mullershausen,F. author:Mullershausen,Flori
an
AUTHOR ABBREV: temp_id:1 temp_author:Russwurm,M. author:Russwurm,Michael
AUTHOR ABBREV: temp_id:1 temp_author:Koesling,D. author:Koesling,Doris
CITEKEY IDENT: temp_id:1 id:10
LOCATION IDENT: temp_id:1 id:10
TITLE IDENT: temp_id:1 id:10
AUTHOR ABBREV: temp_id:2 temp_author:Feelisch,M. author:Feelisch,Martin
CITEKEY IDENT: temp_id:2 id:21
TITLE IDENT: temp_id:2 id:21
999:2 checked:0 skipped:0 failed

This output indicates the following:

- three authors of the first reference may already be known to the
  database with their full firstnames. If you can verify this, you
  might want to edit that reference before permanently adding it

- however, the (auto-generated) citation key and the title are
  identical with the reference ID=10 in the database. You should
  verify whether these datasets are the same and not add the new one
  if this turns out to be the case.

- the same holds true for one author of the second reference, as well
  as for the citation key and the title. This is most likely a
  duplicate of reference ID=21.

regards,
Markus



-- 
Markus Hoenicka
[email protected]
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV