refdb-users

[Refdb-users] Making RefDB cooperating with Zotero

[Wei-Wei G. <ww...@gm...>]

Hi Markus,

I searched the mailing list and found there are some discussions about Zotero years ago. The Zotero becomes very
popular recently, but, due to my experience, it has a major flaw. The Zotero has limited ability in managing it 
database, since its only an extension of Firefox. RefDB is powerful in managing reference database. If RefDB could 
managing the database of Zotero directly, which is a sqlite3 database, it will boost RefDB a lot.

It's just my immature thinking. Hope it's not unreality and bother you much.

Best wishes,
Wei-Wei

Re: [Refdb-users] Making RefDB cooperating with Zotero

[Bruce D'A. <bda...@gm...>]

On Sat, Aug 22, 2009 at 9:41 PM, Wei-Wei Guo<ww...@gm...> wrote:

> I searched the mailing list and found there are some discussions about Zotero years ago. The Zotero becomes very
> popular recently, but, due to my experience, it has a major flaw. The Zotero has limited ability in managing it
> database, since its only an extension of Firefox.

What do you mean by "managing [the] database"? Why is this a "major
flaw"? For whom?

> RefDB is powerful in managing reference database. If RefDB could
> managing the database of Zotero directly, which is a sqlite3 database, it will boost RefDB a lot.

I think the ability to easily move data around between systems is
probably the better approach.

Bruce

Re: [Refdb-users] Making RefDB cooperating with Zotero

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

Quoting Bruce D'Arcus <bda...@gm...>:

> On Sat, Aug 22, 2009 at 9:41 PM, Wei-Wei Guo<ww...@gm...> wrote:
>
>> RefDB is powerful in managing reference database. If RefDB could
>> managing the database of Zotero directly, which is a sqlite3  
>> database, it will boost RefDB a lot.
>
> I think the ability to easily move data around between systems is
> probably the better approach.
>

I second that. I've had a look at Zotero's internal data model, which  
isn't really documented anywhere except in the SQL script which  
creates the database. At first glance there are so many differences to  
the data model currently used by RefDB that it would come close to  
rewriting the app in order to make it handle Zotero databases.  
Moreover, the Zotero developers discourage external access to the  
database as Zotero may end up in an undefined state if the database is  
fiddled with while the browser is open.

How good are Zotero's import and export capabilities?

regards,
Markus


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

Re: [Refdb-users] Making RefDB cooperating with Zotero

[Bruce D'A. <bda...@gm...>]

On Sun, Aug 23, 2009 at 11:28 AM, Markus
Hoenicka<mar...@mh...> wrote:
> Quoting Bruce D'Arcus <bda...@gm...>:
>
>> On Sat, Aug 22, 2009 at 9:41 PM, Wei-Wei Guo<ww...@gm...> wrote:
>>
>>> RefDB is powerful in managing reference database. If RefDB could
>>> managing the database of Zotero directly, which is a sqlite3 database, it
>>> will boost RefDB a lot.
>>
>> I think the ability to easily move data around between systems is
>> probably the better approach.
>>
>
> I second that. I've had a look at Zotero's internal data model, which isn't
> really documented anywhere except in the SQL script which creates the
> database. At first glance there are so many differences to the data model
> currently used by RefDB that it would come close to rewriting the app in
> order to make it handle Zotero databases. Moreover, the Zotero developers
> discourage external access to the database as Zotero may end up in an
> undefined state if the database is fiddled with while the browser is open.
>
> How good are Zotero's import and export capabilities?

Keep in mind that Zotero is now split between a client, and a server.

The client has supported good import/export of standard formats like
RIS, BibTeX, MODS for awhile. Their primary (e.g. lossless) format is
RDF, which is about to get an upgrade (to the BIBO vocab I've worked
with them and others on), and be accessible via a server API as well.

Upshot is that I would assume some kind of server API interop might be
worth exploring.

Bruce

Re: [Refdb-users] Making RefDB cooperating with Zotero

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

Hallöchen!

Bruce D'Arcus writes:

> [...]
>
> Keep in mind that Zotero is now split between a client, and a
> server.
>
> [...]
>
> Upshot is that I would assume some kind of server API interop
> might be worth exploring.

It could be implemented with URIs:

    addref zotero://login@localhost
    getref -o zotero://login@localhost :ID:>0

Of course, the Zotero server must meet the following requirements:

* serve in RIS format

* serve only datasets that have changed after a certain timestamp
  (namely the last synchronisation)

Zotero's IDs become citation keys in RefDB.

This also means that RefDB must introduce last-modified (and maybe
created) fields for its datasets.  This is something that I could
use for the Web frontend, too, for easier caching and for presenting
newly-added references on the front page.  The last-modified field
must be in libinfo because it's user-dependent.

All this is not very high on my wishlist for RefDB, though.  Manual
synching via RIS with Zotero now and then is fully sufficient for
me.

Tschö,
Torsten.

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

Re: [Refdb-users] Making RefDB cooperating with Zotero

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

Torsten Bronger writes:
 > It could be implemented with URIs:
 > 
 >     addref zotero://login@localhost
 >     getref -o zotero://login@localhost :ID:>0
 > 

What is required on RefDB's end to support this URI usage?

 > 
 > This also means that RefDB must introduce last-modified (and maybe
 > created) fields for its datasets.  This is something that I could
 > use for the Web frontend, too, for easier caching and for presenting
 > newly-added references on the front page.  The last-modified field
 > must be in libinfo because it's user-dependent.
 > 

RefDB has an xuser_date field in the t_xuser table. As per the RIS
spec, this date is used to denote when a reprint was requested (RP =
ON REQUEST), but I don't see any reason why this field shouldn't be
reused for other reprint states. This would require moderate code
changes. If that is not sufficient, I could add a "last fiddled with"
field for internal purposes.

regards,
Markus

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

Re: [Refdb-users] Making RefDB cooperating with Zotero

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

Hallöchen!

Markus Hoenicka writes:

> Torsten Bronger writes:
>
>> It could be implemented with URIs:
>> 
>>     addref zotero://login@localhost
>>     getref -o zotero://login@localhost :ID:>0
>
> What is required on RefDB's end to support this URI usage?

It must treat given files which start with "zotero://" specially.
If refdbc hits a "zotero://..." filename, it must call (new)
functions which talk to the Zotero server (through sockets or dbus,
I don't know what it offers).

All I wanted to show is a possibility to realise the semi-automatic
synchronisation with Zotero without adding completely new refdbc
commands.

>> This also means that RefDB must introduce last-modified (and
>> maybe created) fields for its datasets.  [...]
>
> RefDB has an xuser_date field in the t_xuser table. [...] If that
> is not sufficient, I could add a "last fiddled with" field for
> internal purposes.

I save it in Web-frontend data structures, so I don't need it in
RefDB anymore.  Add it only if you think it could be useful for
others.

Tschö,
Torsten.

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