refdb-users — List for RefDB users

Re: [Refdb-users] FreeBSD: After several updates (PHP included) SRU/webinterface doesn't present search results in particular

[O. H. <oha...@ze...>]

On 06/08/10 15:37, Markus Hoenicka wrote:
> "Hartmann, O." <oha...@ma...> was heard to say:
>
>> Hello,
>> after a lot of updates on recent FreeBSD ports one of our RefDB servers
>> doesn't present search results anymore when using the web-based SRU
>> PHP-scripts. I was expecting search results listet in particular after
>> the header of presenting how many results have been found, but I only
>> get Matching Results number, not any reference to edit, print or
>> whatsoever.
>>
>
> Hi,
>
> it's been a while since I've been working on RefDB's web interfaces, but
> to the best of my knowledge it either isn't SRU, or it isn't PHP. RefDB
> ships a SRU interface which is implemented as a CGI module. The results
> are XML which are displayed by appropriate stylesheets. RefDB also ships
> a PHP-based web interface. However, this doesn't use SRU, but talks to
> the server via the command-line clients.
>
> In order to help, please let me know which of these web interfaces is
> affected by your upgrade. I have a fairly recent FreeBSD-8.0
> installation, so I could run a few tests. Also, did you try to increase
> the log level in all involved programs? This won't catch all problems,
> but might give a few additional hints.
>
> regards,
> Markus
>

Sorry for the noise.
The update-orgy on FreeBSD due to a gettext-update left several boxes 
with partially hidden problems. After a forced reinstallation of PHP5 
and all additional PHP-ports everything runs again like expected. I 
never figured out what caused the problem.

Regards,
Oliver

Re: [Refdb-users] FreeBSD: After several updates (PHP included) SRU/webinterface doesn't present search results in particular

[Markus H. <mar...@mh...>]

"Hartmann, O." <oha...@ma...> was heard to say:

> Hello,
> after a lot of updates on recent FreeBSD ports one of our RefDB servers
> doesn't present search results anymore when using the web-based SRU
> PHP-scripts. I was expecting search results listet in particular after
> the header of presenting how many results have been found, but I only
> get Matching Results number, not any reference to edit, print or whatsoever.
>

Hi,

it's been a while since I've been working on RefDB's web interfaces,  
but to the best of my knowledge it either isn't SRU, or it isn't PHP.  
RefDB ships a SRU interface which is implemented as a CGI module. The  
results are XML which are displayed by appropriate stylesheets. RefDB  
also ships a PHP-based web interface. However, this doesn't use SRU,  
but talks to the server via the command-line clients.

In order to help, please let me know which of these web interfaces is  
affected by your upgrade. I have a fairly recent FreeBSD-8.0  
installation, so I could run a few tests. Also, did you try to  
increase the log level in all involved programs? This won't catch all  
problems, but might give a few additional hints.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

[Refdb-users] FreeBSD: After several updates (PHP included) SRU/webinterface doesn't present search results in particular

[Hartmann, O. <oha...@ma...>]

Hello,
after a lot of updates on recent FreeBSD ports one of our RefDB servers 
doesn't present search results anymore when using the web-based SRU 
PHP-scripts. I was expecting search results listet in particular after 
the header of presenting how many results have been found, but I only 
get Matching Results number, not any reference to edit, print or whatsoever.

I guess since the main setup of the server hasn't change since the 
massive ports-update this problem can be solved easily - but I don't know.

Can anyone help?

Regards,
Oliver

Re: [Refdb-users] refdbib: multiple occurence of the same reference ends up in multiple entries in .bib file

[O. H. <oha...@ma...>]

On 02/22/10 11:33, Markus Hoenicka wrote:
> "O. Hartmann" <oha...@ze...> was heard to say:
>
>> Dear Sirs,
>> I have some trouble in creating my bibliography using a refdb-database.
>> My text is spread over several files, connected via \input{} statements
>> in a main head-file. Running latex/pdflatex on the head file, creates an
>> .aux files, which is used by refdbib as input. refdbib then retrieves
>> for each entry the appropriate reference from the SQL DB (postgresql 8.4
>> server on a FreeBSD 8.0-STABLE platform) - but there a multiple
>> occurences of the same reference, exactly as many as \cite't in the main
>> text. Is this a feature or a bug?
>
> This might be a documentation bug. If you use refdbib directly instead
> of the runbib script, you have to eliminate duplicates from the .aux
> file manually. The runbib script does this by passing the contents of
> the .aux file through sort and uniq. You can pipe the results directly
> into refdbib. Have a look at the runbib script as a guidance (search for
> the block of code starting with "### BibTeX"), or use that script
> instead of doing it manually. I'll add a note about this problem to the
> RefDB manual.
>
> regards,
> Markus
>
>

All right,
I got it. Thanks for the update and fast help, hope others would 
appreciate that hint in the doc as well.

Regards,
Oliver

Re: [Refdb-users] refdbib: multiple occurence of the same reference ends up in multiple entries in .bib file

[Markus H. <mar...@mh...>]

"O. Hartmann" <oha...@ze...> was heard to say:

> Dear Sirs,
> I have some trouble in creating my bibliography using a refdb-database.
> My text is spread over several files, connected via \input{} statements
> in a main head-file. Running latex/pdflatex on the head file, creates an
> .aux files, which is used by refdbib as input. refdbib then retrieves
> for each entry the appropriate reference from the SQL DB (postgresql 8.4
> server on a FreeBSD 8.0-STABLE platform) - but there a multiple
> occurences of the same reference, exactly as many as \cite't in the main
> text. Is this a feature or a bug?

This might be a documentation bug. If you use refdbib directly instead  
of the runbib script, you have to eliminate duplicates from the .aux  
file manually. The runbib script does this by passing the contents of  
the .aux file through sort and uniq. You can pipe the results directly  
into refdbib. Have a look at the runbib script as a guidance (search  
for the block of code starting with "### BibTeX"), or use that script  
instead of doing it manually. I'll add a note about this problem to  
the RefDB manual.

regards,
Markus


-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

[Refdb-users] refdbib: multiple occurence of the same reference ends up in multiple entries in .bib file

[O. H. <oha...@ze...>]

Dear Sirs,
I have some trouble in creating my bibliography using a refdb-database. 
My text is spread over several files, connected via \input{} statements 
in a main head-file. Running latex/pdflatex on the head file, creates an 
.aux files, which is used by refdbib as input. refdbib then retrieves 
for each entry the appropriate reference from the SQL DB (postgresql 8.4 
server on a FreeBSD 8.0-STABLE platform) - but there a multiple 
occurences of the same reference, exactly as many as \cite't in the main 
text. Is this a feature or a bug?

Regards,

Oliver

[Refdb-users] Possible risx data loss

[Markus H. <mar...@mh...>]

Hi all,

I'd like to inform everyone that there is a possible loss of data if
you use risx to export references. The problem was reported by Torsten
Bronger in this bug tracker item:

https://sourceforge.net/tracker/?func=detail&atid=385991&aid=2945806&group_id=26091

Symptoms:

CONF (conference proceedings) datasets contain up to 4 titles: article
title, proceedings title, series title, and conference name. The risx
DTD up to 1.2.0 was not able to represent the latter. Thus, if you add
a CONF dataset in RIS format, the database will contain all four
titles. If you export such datasets as risx, the conference name will
be lost. If you export them as RIS, all titles will be retained.

Affected:

All versions of RefDB which support risx up to 0.9.9. The problem has
been fixed in the current svn revision (716) and will not show up in
all future releases.

Workaround:

Use RIS to back up or export your data. Remember that the datasets are
correctly stored in your database. Unless you re-create a database
from a risx dump, your data will be safe.

I apologize for any inconveniences.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

Re: [Refdb-users] multlingual titles in refdb

[Christian W. <cwi...@gm...>]

On 2010-01-13 0:13, O'Donnell, Dan wrote:
>
> Given how unsupported this is at the moment regardless of software, I 
> wonder if a kluge might not be the way to go no matter what system you 
> use, Christian? I.e. code the data with a wiki-style token that you 
> could later use to extract a xml:lang attribute from. e.g. +fr+Les 
> temps perdus; +en+Remembrance of things past
>
Hi Dan,

Thanks for your suggestion.  I was just thinking that maybe not all is 
lost.  Working around the problem certainly sounds less work than 
building from scratch...

My idea was to use a kludge involving the ID to achieve something 
similar, eg have T51n2076 as a base ID for the original text and then 
use T51n2076-en for an English translation of the title, 
T51n2076-ja-Latn for the romanized title and so on.  But I would have to 
see if using subfields like you suggest might be a better solution.  One 
other requirement is being able to pick and combine the various 
languages, but this is a formating issue.  But as Markus pointed out, I 
guess the retrieval will be most tricky to fit in.
All the best,

Christian

-- 
  Christian Wittern
  Institute for Research in Humanities, Kyoto University
  47 Higashiogura-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8265, JAPAN

Re: [Refdb-users] multlingual titles in refdb

[O'Donnell, D. <dan...@ul...>]

Given how unsupported this is at the moment regardless of software, I wonder if a kluge might not be the way to go no matter what system you use, Christian? I.e. code the data with a wiki-style token that you could later use to extract a xml:lang attribute from. e.g. +fr+Les temps perdus; +en+Remembrance of things past

It might be much easier to build the language tokens into an existing bibliographic database than to build a bibliographic database from the ground up. How does MARC handle this? 

Daniel Paul O'Donnell, PhD
Associate Professor,
Department of English,
University of Lethbridge
Lethbridge AB T1K 3M4
Canada

Chair, Text Encoding Initiative (http://www.tei-c.org/)
Director, Digital Medievalist Project (http://www.digitalmedievalist.org/)



-----Original Message-----
From: Christian Wittern [mailto:cwi...@gm...]
Sent: Tue 2010-01-12 5:14
To: ref...@li...
Subject: Re: [Refdb-users] multlingual titles in refdb
 
Dear Markus,

Thanks for your answer.  What a pity, RefDB would otherwise quiet nicely 
fit the bill.  Unfortunately, I am not aware of other software that 
could do this.  MODS maybe could be shoehorned to support this, but last 
time I checked it did not have xml:lang in the schema.  TEI could do 
this, so I might have to go with a custom made XML database (eXist) or 
some such:-(

All the best,

Christian

On 2010-01-12 19:15, Markus Hoenicka wrote:
> Christian Wittern<cwi...@gm...>  was heard to say:
>
>    
>> Hi there,
>>
>> For a project here, I would like to adopt refdb for bibliographic data
>> management.  However, one of the requirements is to be able to have
>> multilingual titles, that is titles repeated for different languages.  As an
>> example, I might have a title of a Japanese paper in original script
>> (Kanji+Kana), Roman Transcription, translation into English, translation
>> into German etc.
>> I wonder how refdb could deal with this.  It appears that at the moment all
>> fields are language-agnostic, right?  Would it be possible to introduce
>> language at some level?
>>      
> Hi,
>
> RefDB cannot deal with this at the moment. Current practice in
> biomedical sciences is to provide the native title, followed by the
> translated title in square brackets. I'm well aware that this is not
> sufficient for the humanities.
>
> Adding language support to article and publication titles requires a
> couple of internal changes. Currently the title is part of the main
> reference entry, except for periodical titles which are already stored
> in a separate table. At this time RefDB only supports a fixed number
> of periodical synonyms. In order to support titles in different
> languages, we'd have to change this to a key-value kind of storage,
> with multiple keys indicating the type and language of a title. This
> is basically doable, but it will require quite a bit of coding time
> (which I'm particularly short of at the moment). Also, it will require
> extensions to the query language as you'd have to be able to query
> specific translations of a title to make full use of these changes,
> something like ":TX[en]:~whatever" to query the English version of a
> title. Adding language support to other fields (which ones would be
> required?) would follow the same pattern and would add quite a bit of
> complexity to the whole thing.
>
> To sum it up, RefDB could be modified to suit your needs, but this is
> nothing that I could do on a rainy sunday afternoon.
>
> regards,
> Markus
>
>
>
>
>
>    


-- 
  Christian Wittern
  Institute for Research in Humanities, Kyoto University
  47 Higashiogura-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8265, JAPAN


------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_______________________________________________
Refdb-users mailing list
Ref...@li...
https://lists.sourceforge.net/lists/listinfo/refdb-users

Re: [Refdb-users] multlingual titles in refdb

[Christian W. <cwi...@gm...>]

Dear Markus,

Thanks for your answer.  What a pity, RefDB would otherwise quiet nicely 
fit the bill.  Unfortunately, I am not aware of other software that 
could do this.  MODS maybe could be shoehorned to support this, but last 
time I checked it did not have xml:lang in the schema.  TEI could do 
this, so I might have to go with a custom made XML database (eXist) or 
some such:-(

All the best,

Christian

On 2010-01-12 19:15, Markus Hoenicka wrote:
> Christian Wittern<cwi...@gm...>  was heard to say:
>
>    
>> Hi there,
>>
>> For a project here, I would like to adopt refdb for bibliographic data
>> management.  However, one of the requirements is to be able to have
>> multilingual titles, that is titles repeated for different languages.  As an
>> example, I might have a title of a Japanese paper in original script
>> (Kanji+Kana), Roman Transcription, translation into English, translation
>> into German etc.
>> I wonder how refdb could deal with this.  It appears that at the moment all
>> fields are language-agnostic, right?  Would it be possible to introduce
>> language at some level?
>>      
> Hi,
>
> RefDB cannot deal with this at the moment. Current practice in
> biomedical sciences is to provide the native title, followed by the
> translated title in square brackets. I'm well aware that this is not
> sufficient for the humanities.
>
> Adding language support to article and publication titles requires a
> couple of internal changes. Currently the title is part of the main
> reference entry, except for periodical titles which are already stored
> in a separate table. At this time RefDB only supports a fixed number
> of periodical synonyms. In order to support titles in different
> languages, we'd have to change this to a key-value kind of storage,
> with multiple keys indicating the type and language of a title. This
> is basically doable, but it will require quite a bit of coding time
> (which I'm particularly short of at the moment). Also, it will require
> extensions to the query language as you'd have to be able to query
> specific translations of a title to make full use of these changes,
> something like ":TX[en]:~whatever" to query the English version of a
> title. Adding language support to other fields (which ones would be
> required?) would follow the same pattern and would add quite a bit of
> complexity to the whole thing.
>
> To sum it up, RefDB could be modified to suit your needs, but this is
> nothing that I could do on a rainy sunday afternoon.
>
> regards,
> Markus
>
>
>
>
>
>    


-- 
  Christian Wittern
  Institute for Research in Humanities, Kyoto University
  47 Higashiogura-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8265, JAPAN

Re: [Refdb-users] multlingual titles in refdb

[Markus H. <mar...@mh...>]

Christian Wittern <cwi...@gm...> was heard to say:

> Hi there,
>
> For a project here, I would like to adopt refdb for bibliographic data
> management.  However, one of the requirements is to be able to have
> multilingual titles, that is titles repeated for different languages.  As an
> example, I might have a title of a Japanese paper in original script
> (Kanji+Kana), Roman Transcription, translation into English, translation
> into German etc.
> I wonder how refdb could deal with this.  It appears that at the moment all
> fields are language-agnostic, right?  Would it be possible to introduce
> language at some level?

Hi,

RefDB cannot deal with this at the moment. Current practice in  
biomedical sciences is to provide the native title, followed by the  
translated title in square brackets. I'm well aware that this is not  
sufficient for the humanities.

Adding language support to article and publication titles requires a  
couple of internal changes. Currently the title is part of the main  
reference entry, except for periodical titles which are already stored  
in a separate table. At this time RefDB only supports a fixed number  
of periodical synonyms. In order to support titles in different  
languages, we'd have to change this to a key-value kind of storage,  
with multiple keys indicating the type and language of a title. This  
is basically doable, but it will require quite a bit of coding time  
(which I'm particularly short of at the moment). Also, it will require  
extensions to the query language as you'd have to be able to query  
specific translations of a title to make full use of these changes,  
something like ":TX[en]:~whatever" to query the English version of a  
title. Adding language support to other fields (which ones would be  
required?) would follow the same pattern and would add quite a bit of  
complexity to the whole thing.

To sum it up, RefDB could be modified to suit your needs, but this is  
nothing that I could do on a rainy sunday afternoon.

regards,
Markus





-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

[Refdb-users] multlingual titles in refdb

[Christian W. <cwi...@gm...>]

Hi there,

For a project here, I would like to adopt refdb for bibliographic data
management.  However, one of the requirements is to be able to have
multilingual titles, that is titles repeated for different languages.  As an
example, I might have a title of a Japanese paper in original script
(Kanji+Kana), Roman Transcription, translation into English, translation
into German etc.
I wonder how refdb could deal with this.  It appears that at the moment all
fields are language-agnostic, right?  Would it be possible to introduce
language at some level?

best,  Christian

-- 
 Christian Wittern
 Institute for Research in Humanities, Kyoto University
 47 Higashiogura-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8265, JAPAN

Re: [Refdb-users] RefDB on 64bit Linux?

[Markus H. <mar...@mh...>]

Quoting Torsten Bronger <br...@ph...>:

> It works smoothly on 64bit Ubuntu 9.04 and 9.10.
>

Thanks for the quick confirmation!

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

Re: [Refdb-users] RefDB on 64bit Linux?

[Torsten B. <br...@ph...>]

Hallöchen!

Markus Hoenicka writes:

> has anyone tried running RefDB on a 64bit Linux system?

It works smoothly on 64bit Ubuntu 9.04 and 9.10.

Tschö,
Torsten.

-- 
Torsten Bronger, aquisgrana, europa vetus
                   Jabber ID: tor...@ja...
                                  or http://bronger-jmp.appspot.com

[Refdb-users] RefDB on 64bit Linux?

[Markus H. <mar...@mh...>]

Hi,

has anyone tried running RefDB on a 64bit Linux system? I've received
a request of a user who tried to run RefDB on a 64bit SuSE
system. refdbd reproducibly crashes in a call to iconv_open:

#0  0x00007fd6dbd6cc7b in ?? () from /lib64/libc.so.6
#1  0x00007fd6dbd6f626 in ?? () from /lib64/libc.so.6
#2  0x00007fd6dbd71111 in malloc () from /lib64/libc.so.6
#3  0x00007fd6dbd16030 in ?? () from /lib64/libc.so.6
#4  0x00007fd6dbd1595e in iconv_open () from /lib64/libc.so.6
#5  0x000000000041d675 in preprocess_citekey_copy (string=0x6e7730
"otazu:2001", n_maxlen=255) at risdb.c:4261

I'd like to know if this might be a libc bug, or if this is just a
symptom of RefDB not being properly ported to 64bit environments.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

Re: [Refdb-users] Regadrs installatinf Refdb

[Markus H. <mar...@mh...>]

Quoting kamal kannan <kkw...@gm...>:

> Hello friends,
>                        I am trying to install Refdb in Ubuntu (Linux), but I
> couldn't complete the installation. If there is any procedure available for
> the installation in ubuntu kindly inform to me.
>

Hi there,

it would be extremely helpful if you could describe in more detail  
where your installation attempt got stuck. Could you please provide a  
list of commands which you ran so we can see at which stage you had  
problems, as well as the output of the command(s) which failed.

regards,
Markus


-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

[Refdb-users] Regadrs installatinf Refdb

[kamal k. <kkw...@gm...>]

Hello friends,
                       I am trying to install Refdb in Ubuntu (Linux), but I
couldn't complete the installation. If there is any procedure available for
the installation in ubuntu kindly inform to me.

                                                        Thanks in advace

-- 
whatever happens life will goes until we live


by


kamalakkannan @ kishor

Re: [Refdb-users] author tag of RIS-dataset

[Markus H. <mar...@mh...>]

Quoting Stefan Schlee <ste...@ya...>:

> If I am correctly interpreting you email you seem to be somewhat  
> dissatisfied with RIS as the foundation/starting point of the data  
> model of refDB.  Are you planning to change the data model on which  
> refDB is based in the future?
>

Well, the rbib article is sort of a blueprint of a future data model.  
The thing is, while it would serve most needs in medicine and in life  
sciences, it is likely still not sufficient for the humanities. They'd  
like to have a more flexible approach to bibliographic data. Traces of  
this discussion can be found on the web. However, as I currently lack  
the time to do a complete rewrite of the data storage, I'm not  
pressing a decision here.

> Besides that, your blog-entries and my test driving of refDB have  
> stirred a lot of questions, remarks, objections, feature requests  
> etc..  But before I can discuss these topics in a productive way I  
> have to base my arguments on more research into refDB and its  
> documentation.  So I would like to postpone the discussion until  
> soon.  Is this kind of discussion better shifted to the devel list?
>

Feel free to use the devel list to discuss the implementation in a  
more technical sense (data storage, client-server protocol, data  
formats and so on). I reckon that feature requests and such are better  
served by the user list.

regards,
Markus

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

Re: [Refdb-users] author tag of RIS-dataset

[Stefan S. <ste...@ya...>]

Hallo!

Your answer helped me a lot.  The two blog entries you have referred to and the link to the "Rbib"-paper have been very helpful.  Not only have I learned about your motivation behind the data model and handling of RIS data.  They have also been informative in a broader sense too, namely in their good exposition of the standard taxonomy applied by library science.

If I am correctly interpreting you email you seem to be somewhat dissatisfied with RIS as the foundation/starting point of the data model of refDB.  Are you planning to change the data model on which refDB is based in the future?

Besides that, your blog-entries and my test driving of refDB have stirred a lot of questions, remarks, objections, feature requests etc..  But before I can discuss these topics in a productive way I have to base my arguments on more research into refDB and its documentation.  So I would like to postpone the discussion until soon.  Is this kind of discussion better shifted to the devel list?

For the time being I think I know enough to (ab)use refDB to meet my requirements (although, further questions not ruled out :)

Kind regards

Stefan


      

[Refdb-users] author tag of RIS-datasets

[Markus H. <mar...@mh...>]

Stefan Schlee writes:
 > Hello!
 > 
 > I have a question concerning the author tag of RIS-datasets.  
 > 
 > I have added the example RIS-datasets shipped with the refDB software which is contained in "testrefs.utf8.ris" file.  This file contains one dataset of type BOOK, with an A1 author tag of name "Smith,K.M.".  When I enter the "getrefs :A1:=Smith,K.M" command I get the result ":0 retrieved:0 failed".  I get the same result with the search expression ":AU:=Smith,K.M.".  When I use the search expression ":AX:=Smith,K.M." the correct dataset is returned, though; the same is the case when I use the expression ":A2:=Smith,K.M.".
 > 
 > As I understand, in the case of RIS-type CHAP the book-author/editor would be entered with the A2 tag, but why is the book author of a RIS-type BOOK dataset "mapped" from A1 to A2?
 > 

In retrospect I have to admit that the usage of A1/A2/A3 both in RIS and in the
RefDB query language has not been the wisest of choices. RIS uses a
somewhat historical and in many ways broken system to map titles and
author names to the corresponding field tags. Feel free to peruse the
following blog entries for some background:

http://www.mhoenicka.de/system-cgi/blog/index.php?itemid=515&catid=37

http://www.mhoenicka.de/system-cgi/blog/index.php?itemid=567&catid=37

Your confusion stems from the fact that RIS uses A1/A2/A3 to denote
the "primary", "secondary", and "tertiary" author of a
publication. Primary is, in the order analytical-monographic-series,
the first "relevant" level in RIS speak. This results in the obvious
problem which you mentioned above: If a book is a monograph, the
author is tagged as A1. As soon as the book contains chapters, the book
"author" (who is an editor now) is tagged as A2.

RefDB tries to be a bit more sane here. No matter what RIS uses, RefDB
uses A1 to retrieve authors at the analytical level (journal article
author, book chapter author, abstract author), A2 to retrieve
authors/editors at the monographic level (book author/editor, thesis
author), and A3 to retrieve authors/editors at the series level (book
series editor). AX retrieves authors from any level. The mapping of
reference data to the various levels are explained in the graphs in
this article:

http://refdb.sourceforge.net/rbib.html

hope this helps
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

[Refdb-users] author tag of RIS-datasets

[Stefan S. <ste...@ya...>]

Hello!

I have a question concerning the author tag of RIS-datasets.  

I have added the example RIS-datasets shipped with the refDB software which is contained in "testrefs.utf8.ris" file.  This file contains one dataset of type BOOK, with an A1 author tag of name "Smith,K.M.".  When I enter the "getrefs :A1:=Smith,K.M" command I get the result ":0 retrieved:0 failed".  I get the same result with the search expression ":AU:=Smith,K.M.".  When I use the search expression ":AX:=Smith,K.M." the correct dataset is returned, though; the same is the case when I use the expression ":A2:=Smith,K.M.".

As I understand, in the case of RIS-type CHAP the book-author/editor would be entered with the A2 tag, but why is the book author of a RIS-type BOOK dataset "mapped" from A1 to A2?

Kind regards

Stefan


      

[Refdb-users] Deleting links with sending xnote XML?

[Torsten B. <br...@ph...>]

Hallöchen!

Can one *delete* links from an extended note to references by
sending xnote XML datasets?

Tschö,
Torsten.

-- 
Torsten Bronger, aquisgrana, europa vetus
                   Jabber ID: tor...@ja...
                                  or http://bronger-jmp.appspot.com

Re: [Refdb-users] Build Debian bin pkg from Ubunt scr pkg

[Markus H. <mar...@mh...>]

Stefan Schlee writes:
 > Only having installed a package "libdbd-sqlite" instead of "libdbd-sqlite3" was one of the reasons why the server could not start when I first tried to install the package.  But installing "libdbd-sqlite3" only solved the problem partially - the server still did not start.  The reason for this was that the "refdbd"-server does not find an "/etc/refdb/refdbdrc" file, because the "/etc/refdb" only contains "*.example" files until I _manually_ run the "refdbd-init" script.
 > 

The configuration file is necessary only if the built-in defaults of
refdbd do not work on the target system. As mentioned in a previous
mail, refdbd tries to load the mysql driver by default for historical
reasons. This is likely the reason why refdbd bails out in the absence
of a configuration file if only sqlite or sqlite3 are available. As
most packaging systems assume sqlite or sqlite3 as the (cheapest)
default driver, I'll change refdbd to load the sqlite3 driver by
default. This will likely allow the server to start during package
installation.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38

Re: [Refdb-users] Build Debian bin pkg from Ubunt scr pkg

[Stefan S. <ste...@ya...>]

Hallo!

My last posting is so terse that it nearly qualifies to be wrong.  Sorry for that, I will be verbose again (thanks for your patience ;)

When I freshly install the "refdb" package, either the first time or after a purge of the package the server exits prematurely after the installation is completed.  So far I was right.  But this statement/answer is not relevant for Markus, to whom's posting I have replied, because the server code apparently works as he writes (only checking for the presence of a db-driver) and the reason why the server exits is a packaging issue.

Torsten Bronger writes:

> I thought it fails on Stefan's system because it doesn't find
> libdbd-sqlite3.  This was the reason why I suggested to swap
> libdbd-sqlite3 and libdbd-sqlite in the "Depends:" and "Recommends":
> fields of the package.  Which libdbd's must be installed for a
> refdbd which doesn't find any configuration files?

Only having installed a package "libdbd-sqlite" instead of "libdbd-sqlite3" was one of the reasons why the server could not start when I first tried to install the package.  But installing "libdbd-sqlite3" only solved the problem partially - the server still did not start.  The reason for this was that the "refdbd"-server does not find an "/etc/refdb/refdbdrc" file, because the "/etc/refdb" only contains "*.example" files until I _manually_ run the "refdbd-init" script.

And now for the somewhat intricate part of the story.  As Torsten writes:

> For me, it works, probably because my system has everything
> installed anyway.

Although it did not work for me, this still may be true for Torsten, because:

1. When you have a running installation and your remove (not purge !) the package and reinstall the package, you will have a running server at the end of the installation.

2. When you install the package for the first time or reinstall it after purging the package (not removing !) it may seem that the "refdbd"-server runs after the end of the installation, because the installation scripts do not issue an error message - although the server has exited prematurely.  If you run the "refdb-init" script immediately after the installation in this case, the script will create a copy of the appropriate "refdbdrc" in the "/etc/refdb" directory, initialise the "refdb" database and start the "refdbd" server.

In both cases everything works fine or seems to work fine.

ad case 1.

When you have a running installation this is so because you manually executed the "refdb-init" script _after_ the installation had completed.  If you remove the package ("dpkg -r refdb") all the files that have been installed into the "/etc/refdb" directory _during_ the Debian installation process will be removed.  The "refdbdrc" that was copied into the "/etc/refdb" directory though will _not_ be removed.  The reason for this is that Debian tries hard to respect the work of the administrator.  If the package-removal software recognises any changes to the original configuration files or any added files in the configuration directory it will leave these files untouched during a package removal.  This means that if you reinstall the package the installation process will find "the old refdbdrc" in the "/etc/refdb" directory and will happily use this configuration file to start the server.

ad case 2.

The reason why the installation process (sometimes) does not issue an error message although the "refdbd" server exits prematurely is probably due to a timing issue.  The "/etc/init.d/refdb" script contains a variable STARTTIME.  Setting this variable to a (small) positive value serves two purposes: It shall prohibit false negative and false positive indications immediately after the server is started.  The false positive indication in our case may have its reason in the fact that STATTIME is not set at all.  This means that the server exits after the "server-probe" has been performed.   Hence the installation scripts issue no error, although the server is not running after the installation has been completed.

ad "purging the refdb-package".

If you try to purge the "refdb"-package ("dpkg -P refdb"), which is much more radical than removing and will eradicate (nearly ?) every trace of the package from the system, you will get some error messages. I assume that this is because the rm-scripts of the package do not currently work perfectly.  Because purging is very valuable if you are testing package installation and removing the remaining traces by hand gets tedious with the time I have written a short script that may be of some help for you.  I have enclosed this script at the end of the posting.  Be careful if you use it: It will remove your carefully crafted configuration file and it may remove your reference database if you are not careful (at least in the "sqlite(3)" case).

Kind regards

Stefan

8<- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

#!/bin/bash

exit_on_err () {
    echo $1
    exit 1
}

dpkg -P refdb 2> /dev/null

echo "Package refdb purged partially ..."

# Remove configuratin files
#
rm -f /etc/refdb/*

rmdir /etc/refdb || exit_on_err "Error: Unable to remove configuration files in /etc. Purging of package refdb probably not completed."

rm -f /root/.refdbarc || exit_on_err "Error: Unable to remove rc files in /root. Purging of package refdb not completed"

echo "Configuration files removed"

# Remove pid files
#
rm -f /var/run/refdb/*

rmdir /var/run/refdb || exit_on_err "Error: Unable to remove pid files. Purging of package refdb probably not completed."

echo "Pid files removed"

# Remove log files
#
rm -f /var/log/refdb/*

rmdir /var/log/refdb || exit_on_err "Error: Unable to remove log files of refdb server. Purging of package refdb probably not completed."

rm -f /var/log/refdba.log || exit_on_err "Error: Unable to remove log files of refdba. Purging of package refdb probably not completed."

echo "Log files removed"

# Remove the databases
#
if [ -e /var/lib/refdb ]
then
    echo "Do you want to remove the databases too? (y/n)"

    while true ; do
	read confirm
	[ `echo ${confirm} | grep "y\|n"` ] && break
	echo -ne "\a"
    done

    if [ "${confirm}" = "y" ]
    then
	rm -f /var/lib/refdb/db/*
	rmdir /var/lib/refdb/db
	rmdir /var/lib/refdb
	echo "All databases removed" || exit_on_err "Error: Unable to remove databases. Purging of package refdb probably not completed."
    fi
fi



      

Re: [Refdb-users] Build Debian bin pkg from Ubunt scr pkg

[Markus H. <mar...@mh...>]

Torsten Bronger writes:
 > I thought it fails on Stefan's system because it doesn't find
 > libdbd-sqlite3.  This was the reason why I suggested to swap
 > libdbd-sqlite3 and libdbd-sqlite in the "Depends:" and "Recommends":
 > fields of the package.  Which libdbd's must be installed for a
 > refdbd which doesn't find any configuration files?
 > 

The default seems to be mysql, as this was the only supported database
engine years ago. This should probably be changed to sqlite3 in order
to simplify installation.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38