[OWW-SC] Biblio extension to support arXiv.org e-prints

[Julius B. Lucks]

Hey Reshma,

Good questions.
>
> 1) Do documents in arXiv.org have DOI's?
>
No they don't.  All arXiv documents live on arXiv.org so I imagine  
there has been no need to register DOI's with them (since they are  
guaranteed not to move around.) (I don't know the official reason why  
though - something I should inquire about.)

> 2) Are digital objects that have DOI's required to have a certain set
> of information associated with them?  If yes, is there a way to access
> that information (like you can for ISBNs and PubMed ID's)?

I'm pretty sure that DOI's are one way to implement URI's (Universal  
Resource Identifiers), and as such are only used to locate  
documents.  There is I am sure some metadata associated with the DOI  
to let resolvers locate it, but I don't think the DOI specification  
has anything about actual content.

>
> If yes to all of the above, would it make more sense to extend Biblio
> to support references by DOI's rather just arXiv.org documents ...
> i.e. go after the general solution.
>

The problem is that once the biblio extension locates a resource (via  
a DOI, pubmed ID, arXiv ID, ISBN, etc.) it still has to parse out  
information about that resource (title, author, etc.).  Given that  
DOI's can point to many different types of resources (many different  
journals with different formats, etc.), there would still have to be  
domain-specific code in the extension to digest these different types  
of information.  Basically until all publishers agree on a common  
format, this problem will always persist.  This is at least to the  
best of my knowledge, but someone probably knows more than me on this  
issue.

> Of course, I don't think that Genbank entries have DOI's so we'd still
> need a specific solution for those.
>

You can think of arXiv as pubmed, with a few differences.  With both  
systems, you can identify a resource via a unique id, and there are  
APIs to perform basic searches, and to retrieve related articles.   
The key difference is that with arXiv, the full text is ALWAYS  
available, while with pubmed not.  In this aspect, the arXiv is much  
like pubmed central. (And in this case the similarity is not  
accidental since the arXiv had a lot to do with the creation of the  
latter.)

Cheers,

Julius

> -Reshma
>
> On 10/10/07, Julius B. Lucks <julius at younglucks.com> wrote:
>> Hey Bill,
>>
>> Do you maintain the biblio extension?  As I think I mentioned  
>> before, I am
>> writing an API that gives full programmatic access to all of the e- 
>> prints on
>> arXiv.org.  While this has primarily been used as a physics/math/ 
>> computer
>> science venue, there is a new, but growing quantitative biology (q- 
>> bio)
>> section.  In fact, Nature was recently inspired by the arXiv.org e- 
>> print
>> model to start a similar e-print service themselves.
>>
>> In any case, I am certainly an OWW user that links to arXiv e- 
>> prints all the
>> time, and I suspect there will be more and more as time goes on.   
>> The API is
>> simple in design, and I actually used the NCBI APIs as a starting  
>> point, so
>> there should be some similarity.  The results are returned in Atom  
>> 1.0 with
>> some extensions for arXiv-specific metadata.
>>
>> How easy/hard is it to add support for arXiv e-prints in the biblio
>> extension?  I imagine instead of putting the pmid, you put the  
>> arxivid which
>> looks something like q-bio/0703028 or just 0708.2038 (for the new  
>> style
>> id's).  The api is in very active development, and we plan to do  
>> an alpha
>> release soon, hopefully within the next couple of weeks.  I can  
>> send more
>> details and documentation as well.
>>
>> Cheers,
>>
>> Julius
>>
>>
>> --------------------------------------------------------------------- 
>> ------------------
>> Please Reply to My Permanent Address: julius at younglucks.com
>> http://www.openwetware.org/wiki/User:Julius_B._Lucks
>> --------------------------------------------------------------------- 
>> -------------------
>>
>>
>>
>>
>> On Oct 9, 2007, at 6:55 PM, Bill F wrote:
>> Accession Numbers and NCBI Lookup for Biblio Extension to OpenWetWare
>>
>> I'm looking into adding GenBank Accession Number lookup, a feature  
>> which has
>> been mentioned in the past. To do this, I need to know what the  
>> set of
>> information from GenBank would be that people would like to see  
>> available.
>>
>> This is a cop of the the current comment on the Software/Features  
>> page:
>> (http://openwetware.org/wiki/OpenWetWare:Software/Feature_requests
>> )
>>
>> Initially, this might be just an extension to Biblio to allow  
>> referencing by
>> accession number. Eventually, this could pull in info from databases.
>> See also
>> http://blogs.nature.com/wp/nascent/2007/07/scientific_blogging_p
>> NCBI has an example of the full record they maintain:
>>
>>
>> http://www.ncbi.nlm.nih.gov/Sitemap/samplerecord.html
>>
>> Would a pointer to a similar page for a specific accession number  
>> be what's
>> needed?
>>
>> The actual accession number for the example record is:
>>
>> "U49845".
>>
>> There's an additional version number on he next line:
>>
>> "U49845.1 GI:1293613".
>>
>>  Does the version number generally get used in citations or is the  
>> Accession
>> Number adequate?
>>
>> Using the Accession number, I can point the Biblio reference to  
>> the NCBI
>> page (similar to the example) for that particular number. Is the  
>> version
>> number also needed?
>>
>>
>> Thanks.
>>
>> Bill Flanagan
>> OpenWetWare.org
>>
>>
>>
>> _______________________________________________
>> OpenWetWare Steering Committee Mailing List
>> sc at openwetware.org
>> http://mailman.mit.edu/mailman/listinfo/oww-sc
>>
>> _______________________________________________
>> OpenWetWare Steering Committee Mailing List
>> sc at openwetware.org
>> http://mailman.mit.edu/mailman/listinfo/oww-sc
>>
>>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/oww-sc/attachments/20071010/1bba3aa7/attachment.htm