Semantic interoperability is often an afterthought. QSi is proposing a radical shift in the way we currently view the nature and relationship between Information, Language, and Data. In the process, semantic interoperability is an emergent characteristic of data management.
2. How I Think About Information
Atomic / Universal: Information artefacts (data, forms, documents, and diagrams)
are all made of the same elements. These elements and their patterns of
association are found everywhere in the digital universe.
Geometric / Organic: The only difference between a database record and a
semantically equivalent document is its ‘shape’. Placing limits on the geometry of
raw information supports an infinite number of diverse expressions.
Taxonomic / Associative: There are a fixed number of orthogonal element classes
and a similarly small number of independent association types in any given
business information environment. From these, any informational representation
may be constructed.
Semantic / Digital : An information artefact can have the same ‘shape’ with
different meaning, or a different ‘shape’ and the same meaning. While context may
affect interpretation, it does not affect semantics.
3. How I Think About Information
Information is the ‘stuff’ of the Universe
“It from bit symbolizes the idea that every item of the physical world
has at bottom...an immaterial source and explanation...that all things
physical are information-theoretic in origin and that this is a
participatory universe.” John Archibald Wheeler
4. Which Problems Are We Solving Here?
‘Multiples of multiples’ is the name given to the challenges faced by
an increasing number of enterprises. The number of factors in
successful integration of information assets is growing and as a
consequence so are the number of combinations of those factors.
Here is a list of the variables enterprises are trying to manage:
• Multiple sources
• Multiple formats
• Multiple languages
• Multiple audiences
• Multiple jurisdictions
• Multiple threats
• Multiple meanings
• Multiple ‘shiny-new-things-that-will-solve-my-problems’
5. What Are We Proposing and Why?
The nature of and relationships between Information, Language, and Data are in
need of a major overhaul. The current lack of precise definitions is a crippling
factor in today’s data management environments. Ditto the lack of understanding
about how Information, Language and Data should work together.
Our proposition is to manage data according to first principles of human
communication. We will separate concerns around data into three areas:
1. Signal or ‘Shape’
2. Semantics
3. Significance
Managing data this way deliver semantic interoperability as a matter of course.
6. The Journey So Far…
Q6 is a protocol dedicated to the idea that there are a small set of
mutually exclusive classes of language elements and a
correspondingly small set of ways to connect them.
7. Language First Design Principles
• Inclusive. All terms and items are first class citizens.
• Context Free. All terms and items are identified and classified at face value.
• Constrained. ‘Assemblies’ of terms use a fixed number of association types.
• Precise. Specialization and variation rules are grounded in linguistic principles.
• Time-sensitive. Temporality is managed in both items and associations.
• Extensible. English is not the only way to express business concepts.
8. Requirements for a Language First Model
1. Technology / application agnostic
2. Domain independent
3. Language neutral
4. 100% ambiguity free, must manage:
a. Equivalents
b. Homonyms
c. Aliases
d. Temporality
In the “Think Big. Start Small.” philosophy, these are the ‘big’ ones:
9. Language First Elements
quant association
+properties +properties
A ‘quant’ is a reusable unit
of information of some
value to an organization.
In order to maintain
application independence,
each quant is classified
according to common usage.
The quant ‘add’, for
example, is a member of the
Task class.
Associations are bi-directional relationships between
two quants. To improve consistency, there are only six
types of association.
The default expression uses the present tense
predicate “is/has” but events that change the
temporal nature of an association may be expressed
accordingly: “was/had” or “will be/will have”.
In the “Think Big. Start Small.” philosophy, these are the ‘small’ things:
10. Q6 Language First Element: quant
quant types
abbreviation
acronym
code
definition
description
expansion
identifier
name
phrase
term
value
quant classes – aka facets
‘quant’: (kwänt)
An atomic* unit of
communication.
*’Atomic’ in the context of the model
means that it is self-contained and
reusable without alteration.
quant ~ quant type
~ quant class
~ digital object identifier
~ language
~ status
~ scope
~ has-equivalent
~ is-plural
~ create date
~ recent event
~ recent event date
~ source
~ contributor
11. Q6 Language First Element: Association
Association
Class
Examples
Q6 associations are designed to
be read in either direction.
Reading from left to right, we
use the ‘is’ predicate and leave
out the ‘has’. Reading right to
left we do the opposite.
This helps with translations and
with managing the time
variations in statements.
12. Data Centric Applications Using Q6
• Business Glossary
applications apply little
or no securityvalues
• Fact Registry
applications apply
medium to high Role-
Based security
protocols
• Analytics and
Business Intelligence
applications apply
high security protocols
13. OK… So What’s Next?
In the course of day-to-day business, no one is going to let
you muck about with live data. More importantly, as a
consultant if you are not delivering value to your client on a
regular and timely basis, you may as well go home.
• Relevance. Where does this approach ‘fit’ in an
enterprise data architecture?
• Value. How do we demonstrate utility in the shortest
possible time?
14. The Relevance Solution: Faceted Glossary
Business
Glossary
Metadata
Assignment
Information
Alignment
Business language is
‘harvested’ from structured
data and registered in the
business glossary
Registered business
glossary values are used as
metadata for documents
15. The Value Solution: MVP to FAIR*
Alphabet Vocabulary Topics Books Libraries
Elements Molecules Compounds Proteins Organisms
Glossary Facts Documents Tagged
Documents
Linked Data and
Documents
Findable. Accessible. Interoperable. Reusable.
16. Repeatable Process – Glossary to Analytics
Diagrams
Digital Twins
Analytics
Maps
Illustrations
Identify
Classify
Associate Tag
Register
Display
Visuals
Choose one or more
database tables and
assign identities to the
values in each column.
Classify and Register the
identified values (quants)
using Q6.
Connect individual quants
to other quants according
to the database schema.
Assign one or more
connected quants as
faceted metadata.
Embed faceted business
glossaries in visuals and
analytics
17. To learn more about how to get started with
Language First solutions please email me at
jogorman@qsi-x.com
Thank You