[Jabref-users] JabRef silently removes entries

[Stefan Björk]

I have tried JabRef for a while, but there is a major problem for me. My
BibTeX database conforms to the apacite package, which - among other
things - means that I have some new entry types, such as 'newspaper',
defined.

When JabRef loads such a file, all unknown entries are silently removed.
I first have to configure JabRef to understand them, then I can load my
BibTeX file.

The fact that JabRef does not give me a warning or an error message,
makes me feel that I can't really trust JabRef. What if I add a new
entry by hand, with a new type, loads in into JabRef, makes some
changes, saves - and suddenly, without me noticing, my manually edited
entry is lost!

Is there a possibility that JabRef at least could warn the user if it
finds unknown entries, and give the possibility to either continu (and
remove the entry) or cancel?


Stefan

[Morten Omholt Alver]

Post by Stefan Björk
I have tried JabRef for a while, but there is a major problem for me. My
BibTeX database conforms to the apacite package, which - among other
things - means that I have some new entry types, such as 'newspaper',
defined.

Perhaps "newspaper" should be added as a standard type in JabRef.

Post by Stefan Björk
When JabRef loads such a file, all unknown entries are silently removed.
I first have to configure JabRef to understand them, then I can load my
BibTeX file.
The fact that JabRef does not give me a warning or an error message,
makes me feel that I can't really trust JabRef. What if I add a new
entry by hand, with a new type, loads in into JabRef, makes some
changes, saves - and suddenly, without me noticing, my manually edited
entry is lost!

This is a valid complaint, but it has already been addressed in the
latest version (1.7.1) - you are supposed to get a warning of which
unknown entry types were encountered, and the entries should be
converted to the "other" type, not lost.

If you are using version 1.7.1 and still losing entries, I would like
to look at an example file where this happens.


Morten

[Stefan Björk]

Post by Morten Omholt Alver

Post by Stefan Björk
I have tried JabRef for a while, but there is a major problem for me. My
BibTeX database conforms to the apacite package, which - among other
things - means that I have some new entry types, such as 'newspaper',
defined.

Perhaps "newspaper" should be added as a standard type in JabRef.

Maybe - but there are also other significant differences between how
apacite handles the other standard entry types. See page 15 in the
apacite manual (of July 1, 2004).

Post by Morten Omholt Alver
This is a valid complaint, but it has already been addressed in the
latest version (1.7.1)

I must admit that I have not tried the latest version of JabRef, which I
of course should have done. Please excuse me.

Stefan

[John Relph]

Post by Stefan Björk
I have tried JabRef for a while, but there is a major problem for me. My
BibTeX database conforms to the apacite package, which - among other
things - means that I have some new entry types, such as 'newspaper',
defined.
When JabRef loads such a file, all unknown entries are silently removed.
I first have to configure JabRef to understand them, then I can load my
BibTeX file.
The fact that JabRef does not give me a warning or an error message,
makes me feel that I can't really trust JabRef. What if I add a new
entry by hand, with a new type, loads in into JabRef, makes some
changes, saves - and suddenly, without me noticing, my manually edited
entry is lost!
Is there a possibility that JabRef at least could warn the user if it
finds unknown entries, and give the possibility to either continu (and
remove the entry) or cancel?
This is a valid complaint, but it has already been addressed in the
latest version (1.7.1) - you are supposed to get a warning of which
unknown entry types were encountered, and the entries should be
converted to the "other" type, not lost.
If you are using version 1.7.1 and still losing entries, I would like
to look at an example file where this happens.

Perhaps this is a design flaw in JabRef? What if JabRef were changed
to have per-database entry types rather than global entry types? As
it stands, the database files contain comments describing the entry
types, their names, optional and required fields.

This is an issue for me because I am trying to roll out JabRef for use
on a number of librarians' desktops. As part of our projects, we use
a number of non-standard entry types. In order to make sure the entry
type definitions are the same across all of the JabRef installations,
I have to modify the Windows Registry on every system running JabRef
so that the customizations to entry types are the same. What that
also means is that if we need to change an entry type, it has to be
changed separately on every system. It would be so much easier if the
database file could describe the custom entry types used in that
database! When the database changed, the other systems would read the
changed entry type definition from the database. Voila!

-- John

[Morten Omholt Alver]

Post by John Relph

Post by Morten Omholt Alver
This is a valid complaint, but it has already been addressed in the
latest version (1.7.1) - you are supposed to get a warning of which
unknown entry types were encountered, and the entries should be
converted to the "other" type, not lost.
If you are using version 1.7.1 and still losing entries, I would like
to look at an example file where this happens.

Perhaps this is a design flaw in JabRef? What if JabRef were changed
to have per-database entry types rather than global entry types? As
it stands, the database files contain comments describing the entry
types, their names, optional and required fields.

The need for global entry types becomes apparent when you consider the
creation of new databases - there must be some place to get the custom
types, to avoid having to create them for each new database.

Post by John Relph
This is an issue for me because I am trying to roll out JabRef for use
on a number of librarians' desktops. As part of our projects, we use
a number of non-standard entry types. In order to make sure the entry
type definitions are the same across all of the JabRef installations,
I have to modify the Windows Registry on every system running JabRef
so that the customizations to entry types are the same. What that
also means is that if we need to change an entry type, it has to be
changed separately on every system. It would be so much easier if the
database file could describe the custom entry types used in that
database! When the database changed, the other systems would read the
changed entry type definition from the database. Voila!

It's very interesting to hear that you are trying to use JabRef in a
setting like that.

I think what you want to accomplish is possible with just a minor
change to JabRef. The entry type information is stored in the bib
file, so the only problem is that it is ignored because the other
JabRefs already have a custom type of the same name.

If you had the option of updating the local definition with the one
from the bib file, the changes could spread easily. Another option is
to let the new definiton shadow the old one, but only for the database
in question. This, however, would imply that the new definition didn't
automatically become available for new databases on that instance of
JabRef.


Morten