Thanks Dave--that makes things clearer. The point about JISC is a moot one
in terms of the current SCONUL/JISC shared services project ....and wider.
It would be interesting to see what publishers would make of a clause in
NESLi licence that required them to expose their data for harvesting (and
maybe even requiring certain standards?). But where would we ask for it to
go?--Google, SerSols. ExLibris? There is no equivalent HE 'platform' right
now and building one would be expensive and complex wouldn't it? Maybe it's
something that could be done with international cooperation?
Ken
Ken Chad Consulting Ltd
Tel +44 (0)7788 727 845. Email: [log in to unmask]
www.kenchadconsulting.com
Skype: kenchadconsulting Twitter: @KenChad
-----Original Message-----
From: Open source software in libraries [mailto:[log in to unmask]] On
Behalf Of David Pattern
Sent: 12 February 2011 07:34
To: [log in to unmask]
Subject: Re: [LIS-OSS] Getting from "here" to "there"? (Re: Welcome!)
Hi Ken
Being on the Summon Advisory Board, I can confirm that Sers Sols aren't just
smply slurping in data from publishers -- a lot of effort has gone into
normalising and de-duping the content, and also in tweaking the references
to try and standardise them. Also, I think they started gathering the data
for well over a year before they actually announced Summon, so that they'd
have a critical mass on day one.
To include Summon content in VuFind, you need to have purchased the full
Summon product (even if you don't plan to use the Summon interface). I
don't think the cost (to Sers Sols) of delivering the Summon interface is
high (it's a lightweight Ruby interface built on top of the Summon API), so
I can't see them offering API-only access (which is what VuFind uses) for a
much lower price.
I think one of the other issues is a lack of standards for publishers to use
to expose article data -- I think Sers Sols has to handle everything from
simple CSV files to complex XML.
Is there a role for JISC, NESLi2, etc in negotiating access to the data
(either for use by subscribing instituions or for ingesting into a national
HE search interface)? Like most thing, we'd need a critical mass before
there was enough content for it to be genuinely useful.
regards
Dave Pattern
University of Huddersfield
________________________________________
From: Open source software in libraries [[log in to unmask]] On Behalf
Of Ken Chad [[log in to unmask]]
Sent: 12 February 2011 01:54
To: [log in to unmask]
Subject: Re: [LIS-OSS] Getting from "here" to "there"? (Re: Welcome!)
Just a few thoughts --
I think Ian has hit a key nail on the head about the indexes for paid
(licensed) content. From what I've seen and heard its seems relatively easy
to harvest the LMS catalogue, repository and even other *local* data sources
into VuFind et al. Issues of data normalisation are not trivial but at least
the *amount* of data is relatively small. BTW JISC delivered a report of IR
and LMS interoperability a while back (2009) that touched on some of these
issues (OCRIS report see
http://www.jisc.ac.uk/publications/reports/2009/scholcommslinksfinalreport.a
spx ). It discusses some of the data problems
So companies like ExLibris and Serial Solutions have put a lot of effort
--negotiated with publishers etc --to get *huge* indexes of licensed
content. The indexes are hosted centrally and made available for local
discovery (I means so a student can get to the stuff their library has
licensed) . It seems like they typically cover about 85% of the licensed
content of a library. This data/index is not available as a free or open
dataset. Hosting, updating, organising etc such a dataset is certainly not
easy or cheap. One developer I spoke to said that it makes Amazon look
trivial (he was ex Amazon)
Of course Google indexes a lot of this content too. Google Scholar (which
itself is one of these 'vertical search or discovery services) can be
linked to a resolver so students can be linked from Google to the
'appropriate' copy (i.e. the full text licensed by their library). If that
worked well enough there would probably be little demand for VuFind or
Summon or Primo etc. If (when?) it worked well enough it would meet Ian's
demand for an approach where 'users don't want and often don't bother to
search in multiple places to find information, so it is our job to bring it
together and make it easier for them'
BASE (from Bielefeld university--' one of the world's most voluminous search
engines' ) does a pretty good job a indexing global (mostly open access)
repository data. http://base.ub.uni-bielefeld.de/en/index.php
In summary, unlike bib data for printed books, I haven't yet seen a free or
'open data' approach to indexes of licensed or even open access (mostly
journal article) content . The MESUR project put a ton of this data together
for the purpose of getting *usage* data but that project ended in 2008. So
even though I think I've heard (or did I imagine this?) about people using
or planning to use VuFind or similar OSS approaches to *search* these
indexes they still got to pay (SerSols etc) for the indexes . Is that right?
Ken
Ken Chad Consulting Ltd
Tel +44 (0)7788 727 845. Email: [log in to unmask]
www.kenchadconsulting.com
Skype: kenchadconsulting Twitter: @KenChad
Open Library Systems Specifications: http://libtechrfp.wikispaces.com
|
