Using the comparisons drawn in last four articles, we can identify experiences from software development that can be applied to the creation and maintenance of hypertext documents.
Structural forms
One aspect of software development that can provide valuable insight into the problems of maintaining Web pages is the management of structures in large software systems. It is widely accepted that embedding structural information within source code makes large systems difficult to understand, build, and maintain.
However, the separation of such information from the code itself, even within the context of a single organization, has been difficult to achieve. Configuration management systems that enforce such separation are gaining greater support, as more tools become available. However, such referential transparency can only be achieved if some widely used programming practices (for example, the use of #include in C programs) are prohibited.
Currently in Web page development, link information is embedded within documents. Besides making document structures difficult to understand and maintain, this approach reduces the value of structural information for purposes other than author-led navigation. If link information were stored separately, it could better support generic links (for example, links from all occurrences of a particular string), metrics, visualization, impact analysis, and consistency/ integrity checking.
Existing configuration management tools usually manage links using derivatives of the Unix make utility. This approach, however, does not easily lend itself to the management of open systems nor to link management across multiple sites and organizations.
References
"Hypertext: The Next Maintenance Mountain" by Pearl Brereton, David Budgen and Geoff Hamilton
Go To Page: 1
