Standards in theory and practice:
Evolution, Committees and Revolutions

This is who I am...

Late 1700's-Early 1800's: Industrialization

  • Steam Power

  • Semaphore Communications

  • Rotary engines

  • Telegraph

  • Railroad

  • Telephone

Oh snap, this is hard.

Engineering Associations

Engineering Associations form because engineers think about common problems and they know about common concerns.  

They realize early: We need standards.
So they start to create them - local, regional, incompatible standards.

Technically great ideas that you can't get implemented aren't worth much...

Physical resource makers had veto power.

Companies begin to learn the value of internal standards as they scale - and slowly, external.

We need standards.

1903: 1500 buildings burned in Great Baltimore Fire
No standards around anything: Others couldn't help.
Over 600 variants of hose size/couplings.

Early National Standards

  • 1901: Engineering Standards Committee (UK) 

1918:  American Engineering Standards Committee

 

American Institute of Electrical Engineers, American Society of Mechanical Engineers, American Society of Civil Engineers, American Institute of Mining and Metallurgical Engineers, American Society for Testing Materials, U.S. Department of War, U.S. Department of Navy, U.S. Department of Commerce

Judicial rather than technical

1928:  s/American Engineering Standards Committee/American Standards Association

 

1947: International Organization for Standards (ISO)

1966: s/American Standards Association/United States of American Standards Institute

Benefits through certification

1969: s/United States of American Standards Institute/ANSI

Public comment and review

Our approach to standards changes and adapts
for a number of reasons...

This has all happened before...

Standards are:

  • Necessary at some level.

  • Hard.

  • Better at "last miles".

  • Trying to solve problems that are fluid and have to adapt when necessary.


 

Mid 1960's: New Problems

IBM System 360

Software Changes the Game:
Solving the problem at a higher level
 

  • Trading Data

  • Networking
  • Web / Scripting

EDI vs (S)GML

UNH+ME000021+IFTSTA:D:96A:UN:EAN002'

BGM+61E::9+95-455+9'

DTM+137:19970204:102'

NAD+FW+5422331123459::9'

NAD+CZ+5412345123453::9'

CNI+1+4215A'

STS+1+21'

RFF+DQ:5/2334'

DTM+334:199702011450:203'

NAD+DP+5411111123451::9'

CNI+2+5122C'

STS+1+31'

DTM+40E:19970207:102'

CNI+3+5145E'

STS+1+17+229'

DTM+334:19970115:102'

LOC+16E+:::H.MCUSTOMS AND EXCISE LONDON HEATHROW AIRPORT'

CNI+4+7655S'

STS+1+49'

DTM+334:19970128:102'

GID+1+1:CT'

PCI+30+354107380000001068'

STS+1+21'

DTM+334:19970128:102'

GID+2+1:CT'

PCI+33E'

GIN+BJ+354107380000001051'

STS+1+21'

DTM+334:19970128:102'

GID+3+1:09::9'

PCI+33E'

GIN+BJ+354123451234567892'

UNT+33+ME000021'

:h1.Chapter 1:  Introduction
   :p.GML supported hierarchical containers, such as
   :ol
   :li.Ordered lists (like this one),
   :li.Unordered lists, and
   :li.Definition lists
   :eol.
   as well as simple structures.
   :p.Markup minimization (later generalized and formalized in SGML),
   allowed the end-tags to be omitted for the "h1" and "p" elements.
  • Early success/govt/business driven
  • Hyper specific / Controlled:  No way it can support kittens

 

  • IBM Research, intellectually driven by small group - taken to standards later (SGML ISO 8879:1986)

  • Extremely general / separation of concerns:  Can totally support kittens

Standards relapse
a) Design by Committee, Process and Compromise 

vs
b) Isolation / Rough Consensus and Running Code
 

Once again, adoption matters - but notice the shift?

Makers are now pushing bits, not iron and steel.

HTTP / HTML

The Web

  • Measurably worse than many existing answers
  • Had a good mix and a seriously low barrier to entry
  • It was good enough to do useful things.

W3C
Great Progress,

Scripting

& Great Crisis.

 

Developers become the new veto power 
important innovators

High-level Observations

  • Technical superiority means little - ultimately there are many kinds of advantage that trump it.
  • Software is a great equalizer - #developer-power 
  • Standardization in the small is easier than in the large.
  • "Adjacent Possible" becomes possible and moves fast. 
  • Isolation/Competition has advantages - to larger changes frequently happen here.
  • We can figure out what's beneath via study.

From a distance: 
This looks more like a living system than a machine.

Recognize new realities and adapt the approach.


Traditional means won't work, it's time for standards to evolve again.

Living systems are actually amazing!

Step 1

New ideas as primitives.
Focus on excavating.
Tighten the feedback loop.
Prollyfill before native.

#extendthewebforward

Made with Slides.com