Article:		The XML Book Business
Subject:		XML is a disappointment
Date:		2003-11-20 02:14:59
From:		Michael Grazebrook
XML is a disappointment for me. At a business level, the need for a set of common standards for data interchange is vast and important. But why oh why was it done like this? Over the past couple of decades I've worked with many data formats, and XML / XSLT is one of the less readable ways to present data. It's really hard to provide layout which makes an XML document readable without transformation. This gets worse when dealing with XSLT for use with XHTML. The code and the content (XHTML tags) are hopelessly intertwined. Modularity - one of the foundation stones of development - is out the window as a result. And yet - the business need is powerful. XML Schema is indeed hard to understand (and I write as one who has written schema processors). But the schema for an industry - an agreed definition of the data standards for a domain - is something of immense power. With standards like MPEG-7 (multimedia), DIG-35 (cameras), FPML (financial products), SWIFT (financial back office) - the data architecture of an indistry is laid bare. And the data architecture of an industry lets me develop applications for that industry in a fraction of the time it would normally take. I can generate the database. Class libraries. Data input screens. Reports. All with a little manipulation of the schema. All knowing that there will be no problem with incompatibilities. With change reduced by the perfection of the initial analysis in the schema, shining gems of analysis crystallized in the hard form of a specification agreed by an industry. And THIS is what none of the books teach me.

Previous Message

Next Message

• XML is a disappointment
  2003-12-05 10:09:58 Jonathan Kajeckas [Reply]
  
  XML is a standard for a standard. XML represents the means for data exchange among parties who have agreed on a common data architecture, and many potential XML users, though they performs similar functions and would sorely like to exchange data (I'm thinking of government), do not yet have a common data architecture. To use XML then, such customers need to form a standards body, define a common data dictionary and schemas, and purchase mainstream tools which can validate input against their arbitrary schemas.

  
  There's a chicken and egg problem. Until those industry standards have been defined such that vendors can easily program for that market, vendors hold off. Individual players usually cannot define the standards themselves, and have little track record of cooperation. Scarce money also makes them timid; let someone else define the standard, then we'll use it as soon as a major vendor makes it easy for us.

  
  This might cause a lull in the book market. Developers already either have their XML books or look online, and beginners may be waiting for mainstream tools such as Office 2003 (in addition to the definition of standards). I'll bet that books on Dreamweaver outsell books on HTML these days. One downside to the idea of Office 2003 being the mainstream tool that popularizes the use of XML is that Office 2003 will probably have slow uptake because money is tight.