12. For facts and for meaning
Domain experts and enthusiasts know more than we do
And have more resources to devote to understanding of specific resources
Our value is in connections and curation
And facilitating new knowledge by others
We don’t “own” the information ecosystem
We must do what it takes to scale up to current and future
volumes of data
We will be increasingly reliant on others
13. For users:
To mine usage data to provide more relevant and authoritative discovery
To start from the most relevant information and provide easy means for quick
expansion
To coherently display conflicting information
For cataloguers:
To flag dead ends for review and action
To mine and show candidate connections for review
To highlight potentially conflicting information
To hide complexity (URIs, etc)
We will be increasingly reliant on systems
14. The data and systems are all in the cloud
Library-based discovery less important but likely still around for a
while
Several ways systems can navigate the graph
Crawling
Dereferencing
Query federation
(See http://linkeddatabook.com/editions/1.0/#htoc84)
So how will Linked Data systems work?
15. Creativity
Judgment
Quick analysis
Ability to see patterns
Good technical instincts
Trust
What skills are needed in the future for cataloguers?
Connecting lots of different sources of data
Creating new connections
Adding value by bringing together
This is what the bibliographic universe of the future will look like
Not just bibliographic information but everything anyone would want to know
Why shouldn’t the library catalogue be an encyclopedia?
Graph thinking
These are astonishing numbers
Need maintenance/caretaking
But also a different way of thinking
Can’t manually manage
Have to let the ecosystem (technology) keep things going, only intervene at key points
Be pragmatic, can’t lovingly curate as we did
This volume helps us get over Not Invented Here
Metadata creation and maintenance as gardening
Lots of variety
Needs upkeep
But never perfect
Things are introduced, grow, and die
Gardener as an overseer not a creator
Sun, soil, rain all occur out of our control – the system does a lot to maintain itself
We just poke at it from time to time
We will do some metadata creation, however
Prepare the soil (design/build systems)
Take seeds from elsewhere (do research, create new data – think author research, not necessarily subject analysis)
Give them some TLC and get them into the ground
They grow, change, interact with the environment and give rise to new things
Relationships (Sequel, influenced by, etc)
What we imagine is factual info about a new publication coming from the source, so we don’t have to create it
And copy cataloguing actually working – not lots of different copies everywhere
For rare, unique, archival – research is more important (higher level skill).
Context of the resource more important than factual information
Finding other sources – more about this in a minute
Stop:
Transcribing
Subject analysis on mainstream works you haven’t read
Tweaking records to make them look nice or meet local standards
Start:
Focusing more on rare/unique material with little information available
Filling in gaps where no information exists
Getting enough information out there to allow experts to do their job
Find data out there
Evaluate it
Massage it to make connections, make additional connections in the graph
Be part of the LD world
Change in thinking – not OUR data and THEIRS
Not bringing in vs pushing out data
You’re creating hybrids
Merging things to make new things that others will continue to build on
Not “is this good enough” but “how can this be most useful?”
You’re enhancing the graph
Eg sameAs for both classes and properties
Find and connect good sources of topical/analytical data
Find good vocabularies and integrate them into library managed and external vocabularies
And flagging/processing/marking importance of data for use in library run discovery
Poke a bit to make things more machine readable (Eg make a template to parse some data)
Cataloguers as part of LD world, not just library world
Needs maintenance
NOT perfect, thoughJust keep the ecosystem healthy
Stop:
Tweaking records for consistency (systems will take care of this)
Making local copies of records
Complaining about data quality
Probably stop:
Performing searches for common typos (fewer strings, remember!)
Delegate:
- Maintaining holdings
Do:
Research people to determine if they’re the same (don’t call it authority work)
Follow up on system or colleague suggested leads
Address truly incorrect data, and leave a trail
We are NOT the gatekeepers of all good metadata
role of expert is changing
- And in many ways has not been as strong as we imagine
- Think of the bitter disagreements in academia
Importance of crowdsourcing
MPLP – get it out there
Esp rare and unique materials
We must give up this idea of libraries as gatekeepers – that’s an outdated model that doesn’t work with info flood
A la Kathleen Fitzpatrick
Instead rely on in time filters
Volume of data makes this necessary
For users:
Wisdom of the crowd
Don’t have to predetermine what they care about
For catalogers:
Lots of system intelligence suggesting things that need human expertise
Don’t have to interact directly with the data model for all thingss
Distinction between local and external data minimized or gone
No local copies of records that we edit directly
To the degree libraries run discovery systems will need design and effective use of data
But we can’t write all this intelligence ourselves
Will utilize software written elsewhere
Patterns:
Crawling -> local index (all automated – doesn’t get out of sync easily!)
Dereferencing – on the fly go grab a URI to learn about it; can be slow
Query federation – complex queries to predetermined data sources
This is going to get better
Reprofessionalization of cataloging
NOT transcribing
Soft skills
Quick thinking
Know what the system can do for you and where it needs help
Value others’ expertise