On 5/15/03 9:47 AM, Nicolas Morin <nicolas.morin at scd.uhp-nancy.fr> wrote:
> How could Koha and MyLibrary come together? ... We have many oss building
> blocks : there's no way Koha is going to become something like Millenium, but
> MyLibrary + Koha + DSpace + OpenILL, etc... could soon provide something not
> too far from it. For that to happen they would have to have the ability to
> work together perfectly and also, I feel, to appear under one unifying
> "umbrella" (OSS4LIB?) that would make this "de-centralized" system visible and
> credible to librarians and library managers. What should we do to achieve
> these strategic goals? I don't know. Perfect interoperability between the
> various oss projects I guess. Maybe XML, SOAP and web services are the way
> forward? Maybe developpers from the various projects could come up with
> something like a "roadmap for interoperability"?
On 5/15/03 2:04 PM, Pat Eyler <pate at eylerfamily.org> wrote:
> I'm working on an idea in Koha which might fit well with what you're doing ...
>
> I've created an engine to extract data from Koha and syndicate it with RSS
> (0.91). It's pretty simple, and very extensible. You can look at it here:
> http://www.kohalabs.com/rss/ -- for small values of looking at it
Nicolas and Pat, I think your ideas are very exciting, and I hope you don't
mind me sharing my reply with wider audiences.
Like other people, I believe the present implementations of integrated
library systems are overly monolithic and represent "black boxes" when it
comes to interoperability. For a long time I have thought about "rethinking
the integrated library system." Now these ideas are being expressed in
forums such as OSS4Lib as well the smaller OCKHAM group:
http://ockham.library.emory.edu/
In fact, the OCKHAM group has submitted an NSF proposal to address just
these issues. Only time will tell whether or not the proposal is accepted.
IMHO, there are three technical ways to make things like Koha and MyLibrary
(as well as other library-related software) work better together:
1) common application program interfaces (APIs) - APIs center
around particular computer programming languages. There are
Perl APIs, Java APIs, PHP APIs, etc. This approach is easy to
implement if everybody were to agree on the programming
language. Not likely. This approach is also prone to operating
system dependences. Sort of limiting.
2) common file formats - The sharing of common file formats has
a proven track record. MARC records are an excellent example.
The use of one or more XML vocabularies (RSS/RDF, MARCXML, EAD,
etc.) represent the more modern and more flexible formats of
today. The sharing of common file formats is not dependent on a
network as well as opens the door for storing and archiving
data/information. Like the written word, common file formats
transcend space and time, but they are not necessarily
accessible in the here and now.
3) common network protocols - Common network protocols have
made the Web what it is today. None of what we do would be
possible without things like DNS, SMTP, FTP, and HTTP. Network
protocols are essentially sharing data via conversations
between computers taking place at a specific time and a
specific place. They are operating system independent and
immediate, but utterly dependent on a fragile, ever-present,
working Internet connection.
As Nicolas mentions above, network protocol techniques such as Web services
seem to offer a lot of potential. As Pat demonstrates RSS version 0.91 files
can be the content shared via a Web Service, such as a SOAP envelope. This
is very similar to the framework described in the OCKHAM NSF proposal. In
this proposal a light-weight reference model will be borrowed, modified
and/or articulated -- essentially an XML vocabulary -- that will be used as
the common language for sharing data/information between library services
and modules. These services will then publicize their availability on a
peer-to-peer network. Other peers (clients) will identify required services
by first querying the network. Peers will then exchange data/information via
some sort of Web Service and the XML vocabulary.
Many of the tools necessary to make these ideas a reality already exist such
as JXTA for the peer-to-peer networking and SOAP for the Web Service.
What we need to do is two things. First we need to enhance our software so
it can operate as a Web Service. While not necessarily trivial, this process
is well documented and supported by lot o' API's.
Second, and more importantly, we need to agree on a file format/XML
vocabulary to encode our data into. This is the hard part. We need to
borrow, modify, or articulate a data structure that will hold the
data/information we want to exchange. Based on my knowledge, I advocate
using RDF. RSS (RDF Site Summary) 1.0 is a particular implementation of RDF:
RDF Site Summary (RSS) is a lightweight multipurpose extensible
metadata description and syndication format. RSS is an XML
application, conforms to the W3C's RDF Specification and is
extensible via XML-namespace and/or RDF based modularization.
http://web.resource.org/rss/1.0/
The nicest things about RDF is it extensibility. By defining any number of
additional XML namespaces, it is possible to encode many types of
information into a single XML file/stream. For example, it is possible to
include Dublin Core, RSS, or a made-up namespace (such as a MyLibrary name
space as I have done) for the purposes of encoding the necessary
data/information. I've done this in a minor way as illustrated and briefly
described here:
http://dewey.library.nd.edu/mylibrary/rdf/
I think the things outlined above are entirely possible. Again, the hard
parts would be integrating a Web Service into our applications and agreeing
on an XML vocabulary. The Web Services are well documented and can be
implemented through existing program (PHP, Perl, Java, etc.) libraries. The
really hard part would be agreeing on the XML vocabulary. This agreement
will only come after lots o' discussion and communication.
Comments are desired, and what are our next steps?
--
Eric Lease Morgan
Head, Digital Access and Information Architecture Department
University Libraries of Notre Dame
(574) 631-8604