Electronic Information Commerce   Table of contents   Indexes   XML: What HTML Wanted to Be!
 
 

Implementing the Proper Standard

 
Chris   Ziener
  National Accounts Manager
  Information Architects
6989 S. Jordan Road
Suite 5
Denver   Colorado  USA  80112
Phone: +1 303 766 1336
Fax: +1 303 699 8331
Email: chris@sgml.com Web: iai.sgml.com
 
Biographical notice:
 
Chris Ziener
 
Chris is an SGML Programmer/Analyst and the National Accounts Manager for Information Architects. He is the Webmaster for SGML University and <TAG> The SGML Newsletter as well as an instructor for SGML University. Chris has been involved as a Project Manager for Information Architects in several large projects and is proficient in a variety of programming languages including OmniMark and C.
 
ABSTRACT:
Implement
 XML  
 
The emergence of XML has increased the awareness of the benefits of a structured information system. The issue now: "Which is the proper standard for my company to implement?"
 
 

Overview
 
The early adapters have already begun implementing XML management systems that will have their companies poised to exploit the emerging XML distribution methods. Before the mainstream XML market becomes defined and people begin groping for solutions that address their XML needs, they need to know the possible pit-falls of such a process. More importantly they need to know how to utilize the increased Web functionality and data modeling power that XML provides.
 
The weathered steps to SGML implementation can still be mirrored when developing a system based on XML technology, but the full benefits of such a system are yet to be realized. What dividends can an early adapter hope to receive? The answer is many. By utilizing this technology in a manner for which it is best suited, as a data modeling language, some benefits can be immediately realized while at the same time providing added value today by investing in structured data.
 
Since I originally came up with this topic in January, I have been afforded the opportunity to talk to software developers who are using XML syntax for the basis for their business models. I have heard talks with leading industry analysts that espouse the benefits of transferring information using XML across the Web in order to increase commerce and facilitate trade. I have spoken with groups that are currently pondering the choice between XML and SGML for their corporate document modeling language.
 
The first step in answering the question of which standard to implement is to define your goals. Once you have decided on what you would like to focus on, either documents or data, then you have your answer; SGML for long term none-proprietary storage, re-use and re-purposing of information, XML for system and media neutral data transfer.
 
Although XML shares what we call the "SGML Philosophy" it is being touted and used as the best way to handle and transfer modular chunks of information. This is a fast, lightweight method for passing along structure and intelligence with your data. Unlike a basic data transfer where information is delimited by tabs or commas for transport from one source to another, XML ships the meta-information with its core data. We have worked with two different companies that have immediately seen the benefits associated with using XML as the underlying technology for data streaming.
 
 
 

Typical SGML Integration
 
Information Architects, Inc. is the leading integrator of SGML technologies. We have worked with clients in manufacturing and service environments, including companies in the automotive, aerospace, robotics and semiconductor industries as well as legal, medical, government, software and academic publishers, to evaluate their existing systems and implement an appropriate system.
 
The most successful implementations are well defined in scope, have reasonable expectations and comfortable timetables. The majority of projects do not have the luxury of working within this perfect scenario. However there are steps to ensure that ever effort is made to address the needs of the client and produce favorable results.
 
The overall outline of a full implementation is:
  1. Initial Information Gathering and Process Review
  2. Documented Results of Information Gathering and Implementation Strategies
  3. Development of an Overall Publishing Process Strategy
  4. Prioritization of Particular Strategies and Selection of an Initial Implementation
  5. System Design for the Initial Implementation
  6. Creation, Installation and Testing of the Initial Implementation
  7. Assessment of Results of Initial Implementation
  8. Revisions to Initial Implementation
  9. Development of Full Scale Implementation Plan
  10. Full Scale Implementation
 
None of these steps ever mentioned SGML or documents, so we use the same general blueprint for XML implementations as well.
 
 

Transfer of XML Data to an SGML Document Creation System
 
The first company we spoke with had a need to increase the automation in their document creation process. Most of the pieces in the larger system were already in place; our job was to handle the document creation, not to deal with the authoring which was mainly static text that was generated by the results of statistical analysis. The first order of business, once the high level architecture had been designed and agreed upon, was to extract information from an existing data-repository and send it to the document processing portion of the system. For this portion, we recommended XML for all the reasons that I touched on earlier. We gave them our requirements and in return they pass the text to us, in an orderly and usable fashion.
 
I am not suggesting that before XML this type of data transfer was impossible. XML however does provide a vehicle for passing this information that is agile - not tied down to one specific process. The programs and processes that we create to accept and manipulate this data-stream are simpler, sturdier and less expensive. If the flow of data for some unforeseen reason were to ever change, the system would not crash. Whereas with normal data streams, if there is a delimiter missing or a portion of data is out of order, normal processing systems are unable to cope.
 
 

Data Brokering with XML and CORBA
 
A second company hired us to work with them to develop a complex data brokering system for one of their clients. They required assistance in determining the positioning of a structured information meta-language and the corresponding meta-language tools for their high level architecture before they could proceed with creation of this system.
 
Of the many databases that will be integrated into this system, at least one consists of delimited information marked-up according to the proposed standards set forth by the HL7 medical initiative. XML was the frontrunner for the transfer and deliver of textual pieces of information.
 
Our initial assessment confirmed that XML was the appropriate technology for this process. We then worked with the client and determined that XML would be used for all data transfer throughout the data brokering system. We developed a series of XML DTDs for the various types of data being transferred and helped with the selection and integration of an XML parser.
 
 

SGML and XML
 
SGML and XML can be used in exactly the same way for exactly the same purposes. The differences between the two for the most part are the way each is perceived; SGML as a highly structured and expensive way to handle corporate information, while XML is a glitzy new initiative for displaying and transferring information over the Web.
Electronic Information Commerce   Table of contents   Indexes   XML: What HTML Wanted to Be!