HTML 4.0 Sourcebook
(Publisher: John Wiley & Sons, Inc.)
Author(s): Ian S. Graham
ISBN: 0471257249
Publication Date: 04/01/98

This is not a trivial exercise, since the filtering process is complicated and must take into account the structure of the entire document collection. Thus, to date, use of this approach is limited to sites where the documents are largely delivered from a database. A good example of this approach can be found at the Electronic Books Technology Inc. site (EBT was, however, recently purchased by Inso Inc., so this URL may soon change):

www.ebt.com/

These documents are entirely served from an underlying database of non-HTML documents—filtering programs dynamically convert the database entries (coded in markup language defined, like HTML, using SGML) into HTML and automatically configure the presentation dependent on many variables, including the path being explored by the reader.

Designing and Managing a Web Site

Large Web sites are usually designed and maintained by more than one person. As a site becomes large, maintenance and development become major issues, so it is useful to have a model for management that allows for distribution of management as the site grows and for a well-defined delineation of responsibilities amongst authors, editors, and managers.

Site Organization: The Webmaster

A Web site should have an overall Web manager, or webmaster, who organizes the home page and the main navigational pages. This manager is responsible for ensuring valid links to the top pages of the collection, but cannot be responsible for the internal consistency or details of those underlying collections. This makes for a reasonable division of responsibilities and obligations, giving the individual project groups full control over their own document collections, while leaving the overall management of the site in the hands of a site manager.

A site manager can prepare a collection of HTML document templates for use by groups developing their own resources. This template can contain a generic graphical logo for the site, some document templates that reflect a generic page or collection design and, finally, a list of URLs that link to the main navigational pages maintained by the site manager. Each project group can integrate these links into their own project pages to create links back to the site main home page. They can also use the logos and page templates to preserve, if they wish (or are mandated) to do so, a common look and feel for the entire site.

Individual Project Directories

Each major document subcollection should be placed in its own directory or subdirectory, distinct from the directory housing the Web site home page and main organizational pages. Users responsible for maintaining each collection can then be given permission to create and modify files in their particular directory, but nowhere else. This isolates the projects from each other and ensures that no user can accidentally modify another group’s pages.

Test and Production Directories

In most cases, it is best to give each group two directories—one for the finished and publicly available material, the other as a development or test area. In this way, the development team can develop and modify documents in the test directory, without worrying about affecting the documents being seen by visitors to the site. Once the new material has been fully developed and tested, it can be formally “published”—that is, collectively copied into the public directory, making it available to the general public.

There are several Web development packages—for example, Microsoft FrontPage, Adobe SiteMill, SoftQuad HotMeTaL, and Silicon Graphics WebMagic—that support this type of development and “publication,” while several HTTP servers integrate database management of the document collection with the server. Many of the newer intranet⁴ software suites also include this type of functionality. You can expect more tools such as these to become available in the near future.

---

⁴Intranet: an Internet network designed to work internally within a company or organization and provide a suite of network-based services, such as e-mail, document management, and workflow monitoring.

---

Revision Control and Document Management

In many cases, a single collection will be designed collaboratively by many different users. In this case, it is often convenient to use revision control or document management software to manage the collaborative development process. Such software can archive the changes made to a collection and can also ensure that only a single person can modify a given document at a time. To some degree, the tools mentioned in the previous section support this functionality.

There are also many standalone revision and document management packages available on every platform (most were originally designed for software development projects, for which similar issues are important) that can be adapted to Web applications. Some, such as rcs (Revision Control System) and sccs (Source Code Control System), often come bundled with UNIX systems and were designed to manage large software development projects. If you are developing a large project involving many document developers, you should definitely consider one of these tools. Sites such as Yahoo! have links to information about other similar packages. If you are interested in this option, you should search Yahoo! (or other) sites, using search strings such as “revision control” or “document management.”

Document Management Tips

In the absence (or even in the presence) of document management software, you will want to use the following useful rules when designing and maintaining a large document collection:

•  Keep your documents well organized on disk, using subdirectories to appropriately organize the material by topic or workgroup.

•  Use HTML comments (<!-- comment -->), META elements, or some other mechanism, to record modifications to the documents. Document management systems often maintain a secondary database for these data. HTML comments and META elements are described in Chapter 6.

•  Use partial URLs to reference documents on the same server. This makes the collection transportable, as relative references are correct regardless of the absolute location of the collection.

•  Preserve the original high-resolution image files used to create page graphics, logos, and buttons. Then, at a later date, you can easily create new document graphics from these original data.