Tradeoffs in Future Proofing Notations

1
Tradeoffs in Future Proofing Notations

• Is seduction by the now inevitable?

Luke Church luke_at_church.name
2
Agenda

• Whats the problem?
• Secondary Notation capture
• Notational design pressures
• Notational space as a resource

3
Seduction By The Now

• Thomas Green, PPIG, 2006

4
Seduction By the Now

• Designer are in danger of being Seduced by the
  Now
• So be as vague as possible to allow future
  expansion
• And were all designers now
• Data hierarchies
• Metadata management

5
Simplified Example

• Trying to rate photos in folder
• 5 -gt Best
• 1 -gt Junk
• 0 -gt Unrated
• But what happens when you want to merge folders
  for a year?

6
Secondary Notation Capture

• Colours are a secondary notation
• No system semantics, but user defined semantics
• Someone else sends you some photos they coloured
• The secondary notation has been captured into a
  user/domain/group specific primary domain
• It can never be reused as a global primary

7
Pressures on Notations
8
Attempt 1 Vagueness

• Decreases Premature Commitment ?
• Infamous for problems with
• Definition and implementation ?
• Interoperation ?
• Across organisation
• Across time
• Blamed for many software failures
• Instancing problem ?
• Cant instantiate a vaguery, like visual
  programming, we must have an instance

9
Attempt 2 Secondary Notation

• Unsystematic vagueness is harmful, so lets be
  systematic
• Secondary Notation
• Mark sections of the notation space as having
  semantics that are not interpreted by the system
• Variable names
• Folder names
• Comments
• Indentation

10
But Secondary Notation Capture

• Secondary notation is captured in another domain
• User Capture Captures for own definition
• Red photo tags mean X to you
• Social Capture Captures within a group of people
• You agree with colleges that red tags mean X to
  all of you
• Power tokens
• Capture is often implicit
• Secondary notation can never be reacquired
• -gt Secondary notation will not be enough

11
Attempt 3 Extensibility

• Secondary notation tends to get captured, so
  formalise extensibility
• E.g. Border Gateway Protocol
• Used to control routers on a network
• Core standard extensions
• gtExtensions are ill defined
• gtNo complete implementations
• gtThe implementation is the standard
• Why did we standardise in the first place?

12
Attempt 4 Precision

• Premature Commitment
• We dont know the problem were trying to solve
• Brittle to change
• Consistency problems
• Typical cycle
• Deploy with precision
• Technical innovation
• Heavily overload small sections of notation space

13
E.g. Network Address Translation

• IPv4 Machine IP a 32 bit number
• Weve run out -gt IPv6 128 bit - revolution
• NAT -gt Gateway that holds temporary connections
  using port numbers - evolution

14
Notational Space as a Resource
15
Intro to Network Economics

• Network externalities Benefit of membership a
  (No. members)2
• E.g. telephones -gt exp. Growth curve
• Evolutionary change vs. Revolutionary change
• Notational space as a resource
• Allows us to explain observed trends
• Clearly consider space tradeoffs

16
Explanation of premature notation deployment

• Deploy a limited solution today
• Offer all of the notation space to buy users
• Exponential growth phase
• Removes competitor notations
• Understand what the problem really was
• Mature disfluency
• Use profit from growth to find areas of the
  notation space that can be exploited -gt
  Inconsistencies
• There are economic pressures assisting
  prematurely committing notations

17
Designing Future Proof Notations

• Decide Notation Space lt-gt Cost tradeoff
• Allocate space permanently to
• Primary Notation
• Secondary Notation
• Must give enough to get rapid adoption
• Reserved space
• Applications must not use, but you can release
  at a later date as primary or secondary
  (Evolutionary change)
• Revolutionary change when reserved space becomes
  scarce

18
Summary

• Seduction by the Now is real
• There is no perfect solution, only tradeoffs
• Secondary notations get captured into primary
  notations in other domains
• Notational space is a resource
• Economics has powerful influences on notation
  design

19
Questions?
20
Why has NAT done better than IPv6?

• IPv6 IPv4 with larger address space
• Evolution vs Revolution
• IPv6 implementations are expensive
• Notation Space as a resource
• IPv4 Installed base
• Undoing seduction by the now proved too expensive