Slides from ELAG2013 in Ghent on NTNU University Library's approach to deploying content to web using semantic technologies, OSS and workflow methodologies.
Making Future-proof Library Content for the Web: Metadata-driven Workflows and Doing Things the “Right” Way
1. Making future-proof library content for the Web
Metadata-driven workflows & doing things the “right” way
Tuesday, June 4, 2013
2. About us
Tuesday, June 4, 2013
NTNU UB
Gunnerus special collections
Data, since 2009
Extremists?->ODC PDDL, CC-BY-SA, moving towards RDF as a sole format
Disagree with the trend towards discovery
Reject ideas of working around legacy crap
Personal journey -> from scripter to coder, architect…planner
3. The big idea
Like,
dude,we
totally
need a new
webpage
“
”
Tuesday, June 4, 2013
depression, they mean a webpage
The average library manager is aware of their IT shortcomings
We need a new way of getting data and assets to users
Process of asset production, ingestion, documentation, preservation, storage and provision
5. From webapp to Web
You try
finding
smartass
a suitable image
Tuesday, June 4, 2013
Web scale? No, the only webscale thing is the Web
Web means via HTTP and standard Web tech…not weird library shit
Being a part of the Web is more important than anything else
Do what serious Web companies do
Consume data to provide data
10. Challenges and issues
Here
Tuesday, June 4, 2013
Status quo: IT policy
We need to revise everything (IT plan from early 2000s)
Where we are vs. where we need to be
Partners
Architectural choices
11. From metadata to data-driven
“Hi there!”
Tuesday, June 4, 2013
What we’re doing right now
Adopting linked data changed the way we looked at metadata
Much more at the centre of the process
Workflows are more important than publishing data
Data is very important
Scripting removes 2/3 of the workload, data drives the scripts
Killing holy cows…quality of data and image quality…
We can do a lot…
15. Documents, data, search & discovery
right here,
something,
really
stinks
“
”
Tuesday, June 4, 2013
The problem with discovery: it’s not Web, it’s just on the web…sort of
Search: One page of many millions of pages
Come to us via your preferred route
Add links to enrich
Provide content
16. The right tools for the job
“We’re going to need a bigger hammer…”
Tuesday, June 4, 2013
Documentation at every stage, code, processing, etc.
Technology choices
Nothing wrong with being custom…
Scripting IS on UNIX…
You’re saddled with legacy crap —WinXP? There is a solution
17. It grows
Yeah…
it’s
this
way…
“
”
Tuesday, June 4, 2013
See from own experience, eg face detection for img cataloguing
Provide solutions for real problems, people come back
Acceptance? In current climate?
…on offer from commercial providers is the same old stuff
18. Extending to the institutional level
Well,
this
is
nice
“
”
Tuesday, June 4, 2013
No reason to not extend this thinking to every level
PDF/A…
Partners with content or DIY
Slow and uphill struggle
Better than the alternative
19. Takeaways
Om nom
nom
Tuesday, June 4, 2013
Talk
Work towards the goal of being of the Web
Provide data in the formats for the Web
Consume and use the same data
ELAG2013 -> I see common movement, concensus Sven Schlarb, Joachim Neubert, Niklas