| About InfoWorld : Advertise : Subscribe : Contact Us : Awards : Events : Store |
|
||||
|
|
||||
|
Captured in XML By Jon Udell October 25, 2002 XML DOCUMENTS, encapsulated in SOAP messages, are the packets of the business Web. Standards efforts under way focus on how to create, transform, interpret, sign, and encrypt these packets as they flow among communicating applications and services. Because XML documents will often model the real business documents that support business processes, such as purchase orders, it's clear that people will need to be able to write them, too. Sadly, the tools that capture nearly all of our keystrokes -- e-mail, word processors, Web pages -- can't compose valid XML. Solving this problem is as critical as any challenge facing Web services today. Ideally every operating system would offer a standard XML editing component, embeddable in Web pages and GUI applications. Wired to a DTD (Document Type Definition) or XML Schema, this component would allow users to interactively create or modify valid instances of the DTD or schema.
Microsoft has for years distributed a component that points toward the type of universal XML editing solution we envision. It's called the DHTML edit control. If you run Internet Explorer and have seen a Word-like interface appear on a Web page in place of a plain text input area, you've used this control. It supports WYSIWYG editing of HTML, not XML, and the quality of the HTML that it produces is extremely poor. But the component is so useful that one company, Ektron (www.ektron.com), has made a business out of improving it. Ektron's eWebEditPro enables the Microsoft control to embed in Netscape as well as IE; wraps a JavaScript API around the control so that developers can add or subtract features to customize it for specific applications; and can reduce the awful HTML generated by the control to nice, clean XHTML (Extensible HTML). This format, which combines the familiarity of HTML with the mechanical regularity of XML, is a great way to simplify the management of semistructured and unstructured content. Ektron's eWebEditPro can be used as a stand-alone, licensed product to deploy from a Web server and relay rich content back to it. It's also the authoring tool at the heart of Ektron's CMS200, an inexpensive content management system that combines Web-based site management with eWebEditPro's WYSIWYG editing. The "content objects" that CMS200 manages, and the templates through which it renders them, are stored in a SQL database. Uniquely, the Web site fed by this content need not be driven entirely by CMS200. It integrates with existing ASP, ASP .Net, ColdFusion, and PHP Web sites, which can selectively and incrementally make use of managed content. It's even possible to source CMS200-managed content into static HTML pages. This flexible and developer-friendly architecture, useful for managing Web sites, points in the direction of a general-purpose system for putting users in control of collections of XML business documents. Separately, and not yet integrated with CMS200, Ektron offers eWebEditPro+XML. In this version of the edit control, XML content is displayed in nested frames (defined in an XML configuration file), and made available for structured editing. The XML content is associated with a DTD or schema whose constraints are expressed in the UI. The available choices for the "city" pick-list, for example, are controlled by an XML Schema simpleType that enumerates them. This is a tour de force that pushes the edit control far beyond its intended use. Predictably, the results are not always seamless. In particular, because the editor has no real-time awareness of document structure, it can't detect changes that invalidate the document. To do that, you have to ship the content in batch mode to MSXML 4.0 (the required parser) or write JavaScript code to enforce constraints, which defeats the purpose of using a schema. This is, nevertheless, an interesting here-and-now solution to the problem of collecting XML data from users. A general approach to this problem looms on the horizon. XForms is a W3C work-in-progress that aims to update that trusty workhorse, the XHTML form, in order to cleanly separate data from presentation. (Microsoft won't say whether XDocs is based on, or will support, XForms.) In the XForms architecture, the data is exposed to XML-style manipulation (DOM), transformation (XSLT), and search (XPath, XQuery), and the presentation can be controlled by CSS (Cascading Style Sheets). A co-chair of the XForms working group, Sebastian Schnitzenbaumer is CEO of a company that offers a transitional product that tracks this emerging standard. Mozquito Technologies' Web Access 2.0 makes XForms-like technology deliverable in today's browsers. An XForms form, for example, uses XML Schema data types to constrain the values permitted in a form. As rendered by Web Access 2.0, the form is defined in HTML, uses generated JavaScript to handle validation, and returns valid XML. Mozquito has also demonstrated a Flash-based XForms renderer, based on a suite of XML support technologies (XPath, CSS) it is developing in ActionScript for the Flash 6 player. We like the concepts that Ektron and Mozquito are developing. XML data capture is too important to be siloed within applications such as Word and Excel. It's a capability that needs to be readily accessible to developers and deployable everywhere.
RELATED SUBJECTS SPONSORED WHITE PAPERS
SPONSORED LINKS
|
|||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||