1. 'Interoperability. A quick chat, a few war stories'. Carl Wilson, Open Planets Foundation.
A quick chat, a few war stories...
The OPF team
● Events (face-to-face/virtual)
● Training (staff development)
● Comms (web/email/social)
● Membership (engagement/value)
● Open preservation advocacy
● Operational management
● Infrastructure (host/test)
● Software stewardship (roadmap/maturity/packaging)
● Data corpora
Trying to help institutions create robust
digital preservation workflows by:
● sharing knowledge and best
● improving the quality of digital
Here to set the scene by:
● Defining interoperability
● Providing a personal perspective
● Covering a few engineering fundamentals
● Looking for an easier way
“is the ability of making systems and
organisations work together (inter-operate).”
straight from Wikipedia
Coined in IT and system services to allow for
But it’s not just an IT thing
A broader definition also considers:
● social issues
● political & organisational factors
that “impact system to system performance”
“Task of building coherent services for users
when the individual components are technically
difficult and managed by different
Who’s in the interoperability game?
● Information professionals
● Software engineers
● Staff who work on distributed projects
● People dealing with complexity
Sound like anyone we know?
So those of us lucky enough to work in IT for
Information professionals on collaborative projects
it’s welcome to interoperability X 3…..
But IT does it better...
The UK NHS National Program for IT
● biggest civilian IT project of its kind
● “mashups” for health records
● in theory 2003 - 2007 costing £2.4 billion
● in reality 2003 - 2011 costing £13 billion
● incidentally it didn’t work…….
Things to think about
● test early
● test often
Small is beautiful
● small parts can build beautiful robust wholes
● it’s one way of tackling complexity
● BUT the parts need testable, tested and reliable
Standards are your friends
It’s always worth checking if there’s a standard
available and in use but:
● there’s often more than one
● they’re not always easy to understand
● they’re hostage to commercial interests
Be clear in your intentions
Again simplicity / small size helps here
● Document clearly and publicly
● Keep it up to date
● If you’re not willing to explain how it works….
Test early and often
And test the parts together rather than as well as
Technologies to automate testing:
● online public continuous integration
Interoperability is the future
The Internet of Things is gathering pace
Information exchange happening between more
and more devices, more regularly.
REST, XML, Json
But it’s NEVER finished
Two aspects of interoperability:
● Syntactically interoperable
● Semantically interoperable
That’s to say speaking the same language is one
thing, understanding what the other is saying is
Why data is a special problem
In the cultural heritage sector
● shared data == shared problems
● legal issues (IPR) means we don’t share as we
Moving the mountain
Moving the large volume takes way too long….
● If I want to move > 5 TB for processing
● THEN move the results back…..
A better way…..
It’s now MUCH easier to move the application to
● It’s more secure
● virtual machines are approx. 300 MB
● Docker - application containers
So the oldest interoperability story?
● It might not be the oldest but it’s close,
● about old school integration issues.
● Something of a cautionary tale,
● it didn’t end well…..