Benefits and Applications

XML is good not only for transmission of data from server to browser, it is also ideal for passing data from application to application and from machine to machine. Since XML is database-neutral, it is likely that it will play a major role in connecting heterogeneous databases. Let's examine a number of different ways that XML can extend our Web solution set.

1. domain-specific vocabularies - XML is used to create new tags that all browsers with XML parsers can understand (e.g., mathematical, chemical, and music markup languages). XML vocabularies will provide a standard way for doctors to send prescriptions to pharmacists, for electronic transactions to be transmitted, for astronomical instruments to be described, etc. Microsoft describes horizontal-industry vocabularies (i.e., software distribution and e-commerce) and vertical-industry vocabularies (i.e., telecommunications and aerospace).

2. structured data - Any XML-aware application can extract exactly the info it needs (e.g., it can distinguish between instances of <AUTHOR>Paul McCartney</AUTHOR> and <BEATLES>Paul McCartney</BEATLES>).

3. XML complements HTML - XML and HTML can be combined, with each performing the portion of the task that it does best (e.g., XML data can be used to populate HTML forms or tables; see the Microsoft XML/XSL Viewer Demo, which requires Internet Explorer 4.x).

4. self-describing - No prior knowledge of sender application is necessary because the syntax of an XML document instance describes the relationships among the various elements (either explicitly via a DTD or implicitly by means of element context).

5. metadata - Data about data can be used to describe web sites, to describe a collection of related pages, or to "push" structured content channels to subscribers' desktops.

6. search engines - With XML content markup, queries are more likely to retrieve relevant files due to contextual information. Search engines could retrieve a specific portion of file; they also could be much faster if the added context eliminates numerous irrelevant matches.

7. distributed applications - In a pipeline of XML applications, each application can extract data elements it needs and pass either the entire dataset downstream, or perhaps pass filtered content to participating applications.

8. granular updates - Microsoft explains how with XML embedded within HTML pages, only the changed elements need to be downloaded; the client updates without refreshing the entire page (in some cases); data can be cached in the client.

9. user-selected view of data - XML enables client-side data manipulation; user selects one, several, or all records, sorts by different attributes, switches to graphical view, etc. without requesting data from the server each time.

10. user-specific view of data - Similarly, the same data can be presented differently, perhaps as a subset, depending on the viewer's role with respect to the data (e.g., accounting department sees more details than the purchaser).

11. middle-tier data servers - In a three-tier architecture, XML can be used in the middle tier to integrate data from various back-end databases. According to Microsoft: "Agents will be built to run on the middle tier to access multiple existing DBMSs and output XML."

12. display of data is independent of structure - XSL style sheets will describe how to render the same data on different devices (monitors, printers, palm pilots, WebTV).

13. round-trip HTML from Microsoft - In December 1997, Microsoft announced that the next version of Microsoft Office will use XML to represent the various Office-specific formats. Users will be able to save Office documents as HTML, modify them in the HTML domain, and then import them back into Office due to XML as the intermediate format.

14. numerous possible XML applications include:
    • intelligent agents - content personalization via smart pull/push (possibly with a date-stamped XML repository)
    • structured records (purchase order)
    • object with methods and data (Java, and potentially JavaScript)
    • meta-content about your web site (improves searches)
    • query results
    • graphical user interface of an application
    • persistent storage format (e.g., ODBMS-powered XML repository [see POET])
    • interactive parts catalogs
    • electronic service manuals
    • online process/procedures documentation
    • EDI (electronic data interchange) - mapping data between purchasing and inventory departments of same or different companies