1. What’s next for the web?
Christine JM. Connors
TriviumRLG LLC
November 20, 2009
2. What is the Semantic Web?
✤ The result of a collection of open technology
standards for data modeling and markup that use
existing transport protocols for exchanging
information which can be processed by both humans
and machines.
✤ Also known as Web 3.0 and Linked Data (more or
less)
3. Semantic Web Layer Cake
Key components; time left to influence - publish your use cases
http://www.w3.org/2007/03/layercake.png
33
4. How is it different?
✤ Embraces existing web technologies
✤ HTML, XML, multimedia, Natural Language
Processing et al
✤ Adds technologies that allow a computerized agent to
perform on a user’s behalf
✤ Formalizes business logic in a non-proprietary format
for better exchange of data and data portability
5. The Myths of the Semantic Web
✤ There will be a handful of “killer apps” to replace Web2.0 giants
✤ The current web will be replaced by the semantic web
✤ Everything will have to be tagged
✤ It will be expensive to migrate everything
✤ We will experience instant gratification
✤ It’s hard to get started
4
6. Text
Predicate
Subject Object
Two views of the semantic web
Machine learning, natural language processing, artificial intelligence and linked data
Images frim Wikipedia
7. The Continuum
Thesaurus
Ambiguity Control
Folksonomy Synonym Ring Synonym Control
Hierarchical Relationships
Personalized Labels Synonym Associative Relationships
Control Scope Note
(Equivalency) (BT, NT, RT, USE, SeeAlso)
Less Complexity More
Taxonomy Ontology
List Ambiguity Control Ambiguity Control
Ambiguity Synonym Control Synonym Control
Control Hierarchical Relationships Hierarchical Relationships
(BT, NT) Associative Relationships
Classes
Properties
Localization
Annotation
Reasoning
“NOT”
See NISO Z39.19-2005
8. The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
9. The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
10. Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
11. Power
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
12. Power
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
13. Power
List
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
14. Power
Synonym Ring
List
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
15. Taxonomy
Power
Synonym Ring
List
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
16. Thesaurus
Taxonomy
Power
Synonym Ring
List
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
17. Ontology
Thesaurus
Taxonomy
Power
Synonym Ring
List
Folksonomy
Complexity
The Continuum
We are building more complex and powerful data architectures; all types are available for
use on the semantic web
18. NT
England
Britain BT
NT
NT BT
BT Wales
Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
19. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
20. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
21. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
22. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
23. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
United NT Northern
Kingdom BT Ireland
24. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
Kingdom BT Ireland
25. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
26. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
27. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
28. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
English
29. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
English
30. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
31. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
Parliament
32. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
Parliament
33. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
legislature
Parliament
34. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
legislature
pound sterling
Parliament
35. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
legislature
pound sterling
Parliament
36. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
currency
legislature
pound sterling
Parliament
37. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
currency
legislature London
pound sterling
Parliament
38. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
currency
legislature London
pound sterling
Parliament
39. NT
England
Britain BT
God and my right
NT
NT BT
BT Wales
motto Great
Britain NT
NT
BT Scotland
BT
flag
United NT Northern
God Save the Queen Kingdom BT Ireland
anthem
official
English language
capital
currency
legislature London
pound sterling
Parliament
48. Circle size indicates the # of triples in the dataset
Circle Size Triple Count
Very large > 1B
Large 1B-10M
Medium 10M-500k
Small 500k-10k
Very small <10k
Arrow direction indicates that a given dataset
contains concepts from the indicated dataset
Arrow thickness indicates the # of shared triples
Arrow Thickness Triple Count
Thick >100k
Medium 100k-1k
Thin <1k As of March 2008
50. Sindice
Index to linked data: books, people, places, news, statistics, events, business, music ...
http://sindice.com/map
51. Capabilities
✤ Business development - market analysis, use cases
✤ Technical development - servers, apps, web
✤ Information architects
✤ Information scientists - define, organize, link
✤ User interface and interaction designers - user studies, structural
design
52. Why do clients care?
✤ Reduce, reuse, recycle
✤ Use less storage space
✤ Purpose content for multiple delivery channels and
services
✤ Mashup content
✤ Increase revenue and customer satisfaction
✤ Improved findability
✤ Brand Management
53. Why do clients care?
✤ Reduce time to action
✤ More finely tuned analytical capabilities
✤ Competitive intelligence
✤ Reduce risk
✤ More granular legal and business models for more
precise compliance certification
✤ Comply with regulations
✤ e.g. XBRL (adopting semantic capabilities)
Semantic Web is the internet&#x2019;s equivalent of the Green Building movement: reduce, recycle, reuse. (Re-use, re-mix, = Mashup)
Tim Berners-Lee and others frequently remind us that the idea is to put data IN the web, as well as ON the web.
Sites today are destinations. Putting your data in the web means that it is a reference point.
I happen to be most familiar with the W3C standards - this is the layer cake -- the tehcnologies they see as necessary compeonents.There are a variety of formats to use - find the one that suits your requirements. The data model is more important than the serialization.
Right now the majority of the data on the web is locked up in applications and markup languages that jumble the format, the style, delivery mechanism and the content all together. The semantic web is a group of OPEN standards that provide the common format for describing data so that data from different sources can easily be combined and integrated.
need to work on trust and security
can be one way
myths -
security - open
We are not looking for a Google killer; there is a difference between documents on the web and data on the web. Google is a leader in providing access to documents and will likely remain so for some time. They will not be replaced, but there is opportunity for a &#x201C;Google for data&#x201D; to emerge. The user goals are different, and so the inputs and methods of analysis and retrieval will be different.
The semantic web will co-exist with the current web; per TimBLs blog, there will be markets for both raw data and mashed up data
Tagging everything does not scale. Everything old is new again &#x2013; entity extraction and Natural Language processing tools will (are having) a renaissance. There are critical technical and editorial choices to make when employing those tools, but no, everything does NOT need to be tagged.
$$$ - not really. Franz recently revealed they had converted 10B triples using Amazon&#x2019;s EC2 service for just 2 days for only $192. Many semantic web technologies are being built by people passionate about their work, and they are making it open source. Enterprise level applications will want the security and stability of tested, supported systems that require investment &#x2013; as well as the smart consultants to go along with it &#x2013; but you can GET STARTED with open source tools while you make your case.
It&#x2019;s not hard to get started, and now we&#x2019;re going to show you some simple things you can do. The best and most consistent advice I&#x2019;ve received since becoming interested in the semantic web is this: take baby steps. Solve one discrete problem at a time. Don&#x2019;t try to read the OWL spec and jump in with an OWL Full representation of your knowledge domain &#x2013; you&#x2019;ll drive yourself crazy. Work the model of your domain in small chunks, learn about how to make things disjoint when you have a need for it. Learn about domains and ranges when they come up. Don&#x2019;t worry about first and second order logic until you&#x2019;ve advanced to the point where it INTERESTS you and you NEED it.
As I was thinking about how to begin this presentation, I mused over some ideas at home. It is human nature to reuse, to mash-up data. In early childhood we use the same tune to carry the lyrics for Baa, Baa Black Sheep, Twinkle Twinkle Little Star and the ABC song. Authors and playwrights are inspired by earlier myths - Shakespeare may have been inspired by Pyramus and Thisbe or a handful of other stories when he wrote Romeo & Juliet. West Side Story is another adaptation. Baz Luhrman, the film director, took a stab at it with Leonardo DiCaprio as Romeo, and then went on to produce Moulin Rouge, one of the most ambitious mashups of songs and stories seen in the film industry of late, combining snippets and full songs from David Bowie, Bono, Madonna, Elton John, Fatboy Slim, Rufus Wainwright, Labelle, Nirvana, Nat King Cole and many many more.
There is a computationally complex view of the web that involves Boolean logic, Bayesian algorithms, syntax, pattern recognition, neural networks and more. There is another view that is concerned about meaning, categorization, classification and relationships. This view tends to require more human power. Neither is particularly practical &#x2013; one requires heavy-duty processing and lots of monitoring. The other requires a great deal of handcrafting and maintaining. Using the best of each world will get you further in the long run. There are brilliant minds working in the artificial intelligence space, and we make great use of those tools in our own processing platform, but that&#x2019;s not what we&#x2019;re going to focus on today.
http://en.wikipedia.org/wiki/File:Artificial_neural_network.svg
http://en.wikipedia.org/wiki/File:Xbarst1.jpg
http://en.wikipedia.org/wiki/Naive_Bayes_classifier
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
Taxonomies - are they dead? No, but most of you are NOT building a taxonomy. You&#x2019;re building light-weight ontologies. You may not realize it because your architecture and tools require that you put pieces of it in different places. We talk about breaking down silos, and then go and silo the components of our data models.
There is NO ONE RIGHT WAY to build any of these. They are an ART and a SCIENCE. The IA, UX, UI, etc - all human-computer interaction models for your system are important inputs to the design.
A list can be a pick list, an index, an authority file
Ambiguity Control
Christine Connors vs. Christine Conners :(
List of food
We recently had a long holiday weekend, usually highlighted by barbecues, so let&#x2019;s start with a list of food: hot dogs, hamburgers, buns, mustard, mayo, ketchup, onions, pickles, chips, salad, cookies, etc etc
-----------
A synonym ring is what we think Roget&#x2019;s Thesaurus is.
Synonym Control (Equivalence Relationships)
Ketchup or Catsup
----------
Hierarchical Relationships
Is A, Part of type relationships
Where would you put the poor tomato?
Tomato - vegetable? Fruit? Both? It&#x2019;s part of ketchup, should it be linked to ketchup under condiments?
Mono-hierarchical vs. poly-hierarchical
------------
Associative Relationships - See Also
Salt and Pepper - Spice? Condiment? Or would it be helpful to tell the user who is looking at Spices to also review Condiments? (or, do it for them -- Steve Krug&#x2019;s Don&#x2019;t Make Me Thnk)
See NISO Z39.19-2005
BT = Broader Term
NT = Narrower Term
RT = Related Term (&#x201C;See also&#x201D;)
SN = Scope Note
UF = Used For
USE = &#x201C;See&#x201D; (Refers reader from variant term to vocabulary term.)
------------
Get to define your own relationship types!
Localization
Annotation
Reasoning
&#x201C;NOT&#x201D;
Ontology 101 by Natalya Foy and Deb McGuinnes
Semantic Web for the Workind Ontologist by Dean Allemang and James Hendler
----------------------------
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
I can do my BT/NT stuff, but I can also create classes and properties that i need for my own application
I can also say NO - this object is NOT part of a certain class or have a certain property. In this diagram, I could easily add a branch for the British Isles, define it as having England, Wales, Scotland, Northern Ireland, Ireland, the Isle of Man, the Channel Islands etc, but be able to state specifically that Ireland is not part of the United Kingdom, so as not to create confusion for a machine processing the concept base. The power of no.
I an also now take this graph (presuming I&#x2019;d encoded it properly) and link it to other graphs that define the UK if I choose.
one URI can be embedded everywhere vs. a web page which is maintained by one creator (entity)
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
So where does one get some of this data? As there are many data sets available, and I&#x2019;d like to show you the growth as visualized in these graphs from Richard Cyganiak and Chris Bizer.
Not all of these are completely FREE. I promise you that IEEE and ACM are NOT giving away their research papers.
Here is a directory of data.
CONTRIBUTE TO THE CLOUD of Data
Tom Tague mentioned all of the free data sets - well, many are free to use, many are not; some are hidden behind firewalls.
There are also many schema available - try SchemaWeb, for example.
But even the free ones cost someone money to build - do your part to contribute what you can, even if it&#x2019;s simply managing your own profiles.
Open world vs closed world
And the list goes on&#x2026;