refdb-users

[Refdb-users] Screenshots of new PHP web interface

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

Hi,

finally something to drool over :-) As suggested earlier today, here
are some screenshots of the current svn version of the PHP web
interface (most of what you see here is also in pre6, sans a few
bugfixes):

http://refdb.sourceforge.net/preview/

Comments and suggestions are very welcome.

regards,
Markus

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

Re: [Refdb-users] Screenshots of new PHP web interface

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

Quoting St=E9phane T=E9letch=E9a <ste...@jo...>:

> On the main search page, i would like to see something closer to a
> citation match (author, journal, year, page number, the later is
> missing).
>

So we'd have to add a volume and a start page search box. Maybe we can =20
even rearrange the boxes a little to make journal article searching =20
more intuitive.

> I'm not usre if you index it or not, but getting a DOi field would be
> very nice since those are more and more common those days.

DOIs are stored in 0.9.9, so a DOI search box is easy to add as well.

>
> I reaaly need to go back to packaging in order to provide the latest
> version (there are some splits for perl module iirc).
>

There is also a SRU interface in 0.9.9 which I'd love to see packaged :-)

regards,
Markus

--=20
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] Screenshots of new PHP web interface

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

On 10/23/07, Markus Hoenicka <mar...@mh...> wrote:

> Comments and suggestions are very welcome.

One comment is the "simple search" is not simple. Compare to, say,
Google, or for reference management, Zotero, where there is a single
field. I would think ideally "simple search" would be integrated into
a main table view, and perhaps do updating using Ajax.

Bruce

Re: [Refdb-users] Screenshots of new PHP web interface

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

Bruce D'Arcus writes:
 > One comment is the "simple search" is not simple. Compare to, say,
 > Google, or for reference management, Zotero, where there is a single=

 > field. I would think ideally "simple search" would be integrated int=
o
 > a main table view, and perhaps do updating using Ajax.
 >=20

and

St=E9phane T=E9letch=E9a writes:
 > On the main search page, i would like to see something closer to a=20=

 > citation match (author, journal, year, page number, the later is mis=
sing).

 > I'm not usre if you index it or not, but getting a DOi field would b=
e=20
 > very nice since those are more and more common those days.

I've revamped the main search page to include these suggestions. I
hope the page isn't too cluttered, but now there are four search
sections:

- Simple search: type in a word or a phrase, and RefDB returns matches
  in any field.

- Search by unique identifier: retrieve a single dataset by citation
  key or DOI

- Simple field-based search: same as "Simple search" of the previous
  version, plus fields for volume, issue, startpage

- Advanced query: same as before

An updated screenshot is right here:

http://refdb.sourceforge.net/preview/php-search.png


regards,
Markus

--=20
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] Screenshots of new PHP web interface

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

Quoting St=E9phane T=E9letch=E9a <ste...@jo...>:

>> - Simple search: type in a word or a phrase, and RefDB returns matches
>>   in any field.
>
> I'm not sure if this one is that usable or not, do you provide a '=E0 la
> google' search field ?
>

That's why I'd appreciate to get some feedback of users actually =20
trying out the interface. With a few hundred sample references, typing =20
in a keyword, a name, or a year returns a useful subset (usually =20
around 10 or less, unless the search term is "and" or "a") of the =20
references. I can't comment on the quality of the search though, =20
because I assume everyone's expectations are different here.

And no, there is no google-esque logic behind the search. It just =20
matches the search string in any field. Remember that Google tries to =20
find needles in the haystack of a gazillion unstructured documents. A =20
reference management software uses structured data on purpose, and =20
anything more elaborate than the simple search should involve =20
field-based search strategies (which the interface also offers).

>> - Search by unique identifier: retrieve a single dataset by citation
>>   key or DOI
>
> If possible, i'd like to see DOi first since this is more and more used
>   and is more cross-discplines thant other identifiers (like pubmedID
> for instance).

I'm not religious about the order, so I can change it anytime. But not =20
all entries have a DOI (many journals provide electronic versions only =20
for the past 10 years or so), or are suitable to have one =20
(unpublished, in press, emails, you name it). OTOH every entry must =20
have a citation key. I reckon its main purpose is to look up =20
references that you've cited in your own documents, and this is =20
usually done by citation key.

>
> I'll do my best to package it soon :-)

Much appreciated!

Markus

--=20
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] Screenshots of new PHP web interface

[<ste...@jo...>]

Markus Hoenicka a écrit :
> Quoting Stéphane Téletchéa <ste...@jo...>:
> 
>>> - Simple search: type in a word or a phrase, and RefDB returns matches
>>>   in any field.
>> I'm not sure if this one is that usable or not, do you provide a 'à la
>> google' search field ?
>>
> 
> That's why I'd appreciate to get some feedback of users actually  
> trying out the interface. With a few hundred sample references, typing  
> in a keyword, a name, or a year returns a useful subset (usually  
> around 10 or less, unless the search term is "and" or "a") of the  
> references. I can't comment on the quality of the search though,  
> because I assume everyone's expectations are different here.
> 
> And no, there is no google-esque logic behind the search. It just  
> matches the search string in any field. Remember that Google tries to  
> find needles in the haystack of a gazillion unstructured documents. A  
> reference management software uses structured data on purpose, and  
> anything more elaborate than the simple search should involve  
> field-based search strategies (which the interface also offers).

Well, i was thinking about a match of the term in every part of the 
reference (not the title only, or so), linknig it to google was not a 
good comparison. Btw, you answered the question :-)

>>> - Search by unique identifier: retrieve a single dataset by citation
>>>   key or DOI
>> If possible, i'd like to see DOi first since this is more and more used
>>   and is more cross-discplines thant other identifiers (like pubmedID
>> for instance).
> 
> I'm not religious about the order, so I can change it anytime. But not  
> all entries have a DOI (many journals provide electronic versions only  
> for the past 10 years or so), or are suitable to have one  
> (unpublished, in press, emails, you name it). OTOH every entry must  
> have a citation key. I reckon its main purpose is to look up  
> references that you've cited in your own documents, and this is  
> usually done by citation key.

"Oups", i should have read it with this is mind, of course citation 
_from the database_ is more relevant, i thinked of external citations 
... I see now your perspective, you're right in that case, DOI is second :-)

>> I'll do my best to package it soon :-)
> 
> Much appreciated!

This weekend, presumably tonight of monday nigth at worst.

Cheers,
Stéphane
-- 
Stéphane Téletchéa, PhD.                  http://www.steletch.org
Unité Mathématique Informatique et Génome http://migale.jouy.inra.fr/mig
INRA, Domaine de Vilvert                  Tél : (33) 134 652 891
78352 Jouy-en-Josas cedex, France         Fax : (33) 134 652 901

Re: [Refdb-users] Screenshots of new PHP web interface

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

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

> One comment is the "simple search" is not simple. Compare to, say,
> Google, or for reference management, Zotero, where there is a single
> field. I would think ideally "simple search" would be integrated into
> a main table view, and perhaps do updating using Ajax.
>

I haven't found a word on the zotero web page how their simple search  
works. Do you have any insight? If it is just about searching a string  
in all relevant fields this could be added for sure.

regards,
Markus


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

Re: [Refdb-users] Screenshots of new PHP web interface

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

On 10/24/07, Markus Hoenicka <mar...@mh...> wrote:
> Quoting Bruce D'Arcus <bda...@gm...>:
>
> > One comment is the "simple search" is not simple. Compare to, say,
> > Google, or for reference management, Zotero, where there is a single
> > field. I would think ideally "simple search" would be integrated into
> > a main table view, and perhaps do updating using Ajax.
>
> I haven't found a word on the zotero web page how their simple search
> works. Do you have any insight? If it is just about searching a string
> in all relevant fields ...

Yes, that's the case. The included fields include user notes, for example.

Bruce

Re: [Refdb-users] Screenshots of new PHP web interface

[<ste...@jo...>]

Markus Hoenicka a écrit :
> Bruce D'Arcus writes:
>  > One comment is the "simple search" is not simple. Compare to, say,
>  > Google, or for reference management, Zotero, where there is a single
>  > field. I would think ideally "simple search" would be integrated into
>  > a main table view, and perhaps do updating using Ajax.
>  > 
> 
> and
> 
> Stéphane Téletchéa writes:
>  > On the main search page, i would like to see something closer to a 
>  > citation match (author, journal, year, page number, the later is missing).
> 
>  > I'm not usre if you index it or not, but getting a DOi field would be 
>  > very nice since those are more and more common those days.
> 
> I've revamped the main search page to include these suggestions. I
> hope the page isn't too cluttered, but now there are four search
> sections:
> 
> - Simple search: type in a word or a phrase, and RefDB returns matches
>   in any field.

I'm not sure if this one is that usable or not, do you provide a 'à la 
google' search field ?

> - Search by unique identifier: retrieve a single dataset by citation
>   key or DOI

If possible, i'd like to see DOi first since this is more and more used 
  and is more cross-discplines thant other identifiers (like pubmedID 
for instance).

> - Simple field-based search: same as "Simple search" of the previous
>   version, plus fields for volume, issue, startpage

I think DOI and this one are sufficient for most of the queries, but it 
depends on other PoV :-)

> - Advanced query: same as before

I think you could keep this one and drop the simple search.


> An updated screenshot is right here:
> 
> http://refdb.sourceforge.net/preview/php-search.png
> 

I'll do my best to package it soon :-)
Stéphane

-- 
Stéphane Téletchéa, PhD.                  http://www.steletch.org
Unité Mathématique Informatique et Génome http://migale.jouy.inra.fr/mig
INRA, Domaine de Vilvert                  Tél : (33) 134 652 891
78352 Jouy-en-Josas cedex, France         Fax : (33) 134 652 901