1 . 3 Web Maintenance
As the Web of HTML documents grows, maintenance of links in the documents becomes increasingly difficult. It is frustrating to select a link only to have the browser return an error message that the document doesn't exist.
New tools are helping manage and maintain Web sites. The Webtest tool suite from EIT(5) is a freely available utility. It contains a Verify Web Link tool. It starts from a URL, traverses outward, subject to a searching profile; and reports the results.
Results of Link Verification test with EIT utility
As the Web matures, vendors are catching up to the demand for Web site management products. One product by Adobe is called SiteMill. SiteMill is a WYSIWYG site manager. It provides users with drag and drop controls and tools to manage links, resource usage, and error handling.
SiteMill's external URL reference list and error controls
SiteMill's visually oriented tools help track down references to external URLs and locate dangling links. In the Error windows a user can drag the correct file to the missing icon; all references in the site will be updated.
Another product in this new line of Web management software is Interleaf's CyberLeaf. This system is not an authoring tool, instead it incorporates Web pages authored with whatever tool you like. Integration with the entire enterprise is another feature authoring systems are starting to support. Interleaf uses the term "Web Lifecycle" to describe the process of updating and maintaining a Web. Web authoring systems are introducing templates coupled with tools to help set up the Web site. These are similar in concept to Microsoft Wizards, which lead people through the creation of complex documents. Interleaf's long history of document processing and management systems, primarily for large organizations, is clearly evident here.
Template usage and link management dialogs from Interleaf's CyberLeaf
Web browsers are applications that run on the user's client machine. The client operating system and particular configuration of the client software and networking all play a role in the operation and behavior of the application, the browser. The availability of ancillary applications and properly configured system-wide protocols contribute to the final document's portability or lack thereof.
One important issue associated with wide distribution of HTML documents results from the Web browser's loose coupling with various applications commonly known as helper applications. Web browsers sometimes launch helper applications when the user encounters an image file(6). The particular application launched is dependent on the data's particular MIME type (see Section 1 10 6 MIME ); it is often dependent on the extension used for the file name as well. If, for example, the HTML document points to a JPEG formatted image, the client machine must have an application capable of displaying JPEG images and the Web browser must be configured to launch that application upon links to JPEG images. This same scenario applies to sound and video files.
Naming links is another issue related to system dependencies. There is a trade-off, when authoring, in how to name the link. Using absolute URLs (Uniform Resource Locator) is more reliable but much more painful when you have to relocate the Web documents to another directory structure or Web server. If you know that your documents will be moving, you or the authors should be careful to use only links with relative address names. Doing this will make it easy to move the documents to other locations on the same server.
This becomes important if you think you may want to encapsulate the Web for CD-ROM distribution, an increasingly popular option. Webs of documents can be distributed on CDs with the portion that must be updated obtained from the on-line Web when needed. In this way, the entire hierarchy of HTML files can be moved as a unit without concern for renaming file path names inside the documents. In addition, the relative names often must only be names in directories down from the current location. This is a security feature of the server program.
Of course, after you author your Web pages you must have them placed onto a Web server. Thousands of companies now seem to be willing to host Web pages. They offer virtually any type of service you can imagine, albeit at a price. One particularly intriguing approach offered by AccessAbility Internet Services(7) is the concept of a selfservice Web site. They provide the Web server and host, but you, the author, can do all the maintenance and updates through a controlled process. It's kind of a selfserve copy shop for the 90s.
Self service Web site administration at AccessAbility
