RE: [xml-dev] Build Rich Complexity from a Small Set ofWell-Defined Mark

["Cox, Bruce" <Bruce.Cox@USPTO.GOV>]

Seems to me that markup is intended to facilitate machine processing.  I take that as axiomatic.

When I was studying symbolic logic (many years ago, and not very successfully), I was impressed by those individuals in the class who could find the path from a statement or two to the provided conclusion, using the few operations and primitives at hand.  Once they had done their magic, I could quite easily understand the sequence of statements, but was completely baffled by the process of pulling them seemingly out of the air.  Often, there was more than one path from the premises to the conclusion.

There are many instances in nature of a few things combining into astounding complexities.  Chemistry, astrophysics, biology, all offer exceptional examples.  In the case of XML, the primitives and operations allowed on them are significantly constrained by the need for machine processing.  There are additional constraints added by the need for not leaping too far beyond what is already well known.  And there is the need to design something that can actually be implemented in existing environments.  What is a customer willing to pay or pay for?

Layers upon layers of added constraints are what make things complex, in my judgment.  Nature is by nature simple, even when astounding us.  The added constraints we impose when building structures for machine processing are a measure of a machine's inability to tolerate variation or to finesse conflicts.  If machines were alive ... .  Well, they aren't, so we have to think of just about everything in advance.  Hence, the constraints; hence, the complexity.

Bruce B Cox
Director, Policy and Standards Division, OCIO
571-272-9004

-----Original Message-----
From: cbullard@hiwaay.net [mailto:cbullard@hiwaay.net] 
Sent: 2011 August 24, Wednesday 12:16
To: rjelliffe
Cc: Costello, Roger L.; xml-dev@lists.xml.org
Subject: Re: [xml-dev] Build Rich Complexity from a Small Set of Well-Defined Markup Combinators

Seems like we've been here before:

1.  Complexity of the expression (an instance of the language) 2.  Complexity of the expressible (the space of all possible expressions) 3.  Fitness for the expressed (given some goal, is the expression fit, that is, will it be understood in the most efficient way)

Of these, the controls over the evolution of the language affect what is expressible and all too often that interferes with the fitness for the expressed:  a common problem of type definitions where too many document types are captured under a single root.  IOW, there are co-occurring requirements from the controlling dimensions that create strong consensus but weak control.

Over time, increased expressibility weakens the practice.  This weakens production and evolution (Where evolution is learning applied to practice).

There are two basic approaches both of which are used at different phases of learning by doing:

1.  Vikings: 12 stout men and a boat.
2.  Next generation Vikings:  12 stout men, a boat, and a map.

The number of wise men, priests and kings who must touch the map  
before the voyage approximate the number of wrinkles in the map.   
Excessively wrinkled maps become dangerous and if over annotated, almost useless.

len


Quoting rjelliffe <rjelliffe@allette.com.au>:

> On Wed, 24 Aug 2011 10:16:23 -0400, "Costello, Roger L."  
> <costello@mitre.org> wrote:
>
>> Interestingly, XML Schemas is considered to be a complicated 
>> language. Perhaps 7 markup combinators are too many in a markup 
>> language?
>
> I'd say what people call complexity comes from two things:
>
>  1) application power: is it easy to express the kinds of thing I am 
> interested in expressing?
>
>  2) how memorable is it? does my normal routine reinforce my memory of 
> its details or let them fade?
>
>  3) how many caveats or special cases pr multiple ways to do the same 
> thing does it have?
>
> The first relates to the problems being addressed and the scientific 
> limits to the technology.
>
> The second relates to people's tasks and the amount the schema 
> language re-uses concepts/syntax/vocabulary/models from elsewhere:
> if you use DTDs everyday, XML Schemas will not be so complex; if you 
> use DTDs every day, Schematron will not be so complex.
>
> XML Schemas rather suffers from low bangs per buck in the first case 
> (many flowing out of the requirement to act in a streaming fashion for 
> grammars), low connection with XML processing tasks in the second 
> (depending on whether you do schemas all day or not), and is riddled 
> with caveats and alternatives in the third case.
>
>> When creating an XML markup language consider following this approach:
>>
>>    1.  Create a small set of well-defined markup combinators. From 
>> experience with XML Schemas, 7 or less markup combinators might be 
>> adequate.
>>    2.  Create well-defined mechanisms for combining the markup 
>> combinators. Again, from experience with XML Schemas, 4 or less 
>> combining mechanisms might be adequate.
>
> Lets take three things that contribute to XML Schema 1.0's perceived
> complexity: @elementFormDefault, xsd:complexContent, and compex type 
> extension by suffixation only.  Where do they fit in your scheme?
>
> Cheers
> Rick Jelliffe
>
>
> ______________________________________________________________________
> _
>
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS to 
> support XML implementation and development. To minimize spam in the 
> archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
> subscribe: xml-dev-subscribe@lists.xml.org List archive: 
> http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>
>