The intelligence of machines and the branch of computer science which aims to create it

Artificial Intelligence Journal

Subscribe to Artificial Intelligence Journal: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Artificial Intelligence Journal: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Artificial Intelligence Authors: Corey Roth, Pat Romanski, Liz McMillan, Yeshim Deniz, Kevin Benedict

Related Topics: Artificial Intelligence Journal, XML Magazine, Wireless Technology Magazine, PHP Developer's Journal

XML: RSS Feed Item

Re: Should Subject Matter Experts Determine XML Data Implement

> Nonetheless, my question remains: what is the connection between a Data
> Specification and a derived XML Schema design?

The connection is ensuring the equivalent semantics are applied
regardless of the representation.

I receive a data specifications often in spreadsheet form. These
describe the logical data items, their relationships, and their
associated business meaning.

In most large organisations there are many business applications and
very often these use their own representation of business data. The
critical point is that there is fidelity between the mapping between
business definition and each technical implementation.

We also operate a cannonical data model which is actually the first
thing we map data specifications to. Again fidelity is crucial. We can
express this cannonical model in a number of forms, but most often we
use XML schema. There are issues of course since there is not always a
1:1 relationship between XML schema and the type systems and storage
formats of applications. That means that there is not always a
loss-less transformation or exchange. In my experience getting a 100%
fir is pretty hard and not always essential.

So the connection is the business semantics. So long as these can be
maintained, the rest is just syntax.

In practice this means that business data analysts and IT
designers/implementers need to share a common understanding and
vocabulary for describing data. In reality it usually means a lot of
to'ing and fro'ing between the two.


2008/10/4 Costello, Roger L. <[email protected]>:
> Hi Dave,
>> This is just echoing what was supposed to happen with AI
>> systems in the '80's. Pick the brains of the SMEs and dump
>> it into Expert Systems. It didn't work then. Why should
>> it be any different now?
> I'm not talking about creating Expert Systems. In fact, I'm not talking
> about "application design" at all. I'm talking about "data design."
> I'm talking about picking the brains of the people who truly understand
> the *data* and the *data relationships*, and capturing their knowledge
> in a Data Specification. Then, from that Data Specification an XML
> Schema is implemented by an XML Schema designer.
> My question is about the connection between a Data Specification and
> the derived XML Schema: should there be a 1:1 correspondence? If not,
> how much divergence is acceptable?
> For example, if an expert on Books says that the data which identifies
> a book is its ISBN value, is it acceptable for an XML Schema designer
> to instead use a GUID? If a Book expert says that the printer/producer
> of a book is called its "Publisher," is it acceptable for an XML Schema
> designer to instead call it "Pub"?  And so forth.
> Hi Tommie,
>> Many organizations have found that an iterative
>> process, in which technology experts consult with SMEs and
>> revise the specification, and the design, many times during
>> development.
> Good points. I did not mean to imply that the development of a Data
> Specification is a one-time thing. As you point out, a Data
> Specification may be revised iteratively as deeper insights are
> obtained.
> Nonetheless, my question remains: what is the connection between a Data
> Specification and a derived XML Schema design?
> /Roger
> _______________________________________________________________________
> 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:
> Or unsubscribe: [email protected]
> subscribe: [email protected]
> List archive:
> List Guidelines:

Read the original blog entry...