A Standard for Site Organization
Nov 02, 1998 ::
Standards have been inarguably good for computers. The explosive growth of the Net, the omnipresence of the machines themselves, their very usefulness is almost entirely the result of standards. There just need to be more of them.
The Web is a prime example. At one end of the dataflow are standards for transporting bytes. At the other are standards for designing single pages. In between is a gaping, yawning chasm -- the empty gulf of site organization.
While transport mechanisms such as TCP and HTTP handle shuttling data from one point to another well enough, and HTML and XML offer structure and organization to a single page (at least theoretically), there exists no common, accepted standard for the organization of information on a site. Well-defined standards bookend a seething, uncontrollable mess. The most difficult task on the Web is actually finding anything.
It doesn't have to be that way. Just as convention suggests a Web server start its host name with the now ubiquitous "www," the other end of the URL is ripe for standardization.
A selection of well-chosen and well-named root-level directories, implemented across as many sites as possible, would go a long way toward easing the complexities of navigating unfamiliar sites, no matter how good (or bad) their fundamental design.
Of course, standards are almost impossible to implement after the fact. "web" is arguably a better hostname prefix than "www," yet only a few oddballs have configured their nameservers to recognize it. Likewise, it may be too late for wide-scale adoption of site organization standards.
- /about/, for instance, should always give basic information about a site, its purpose and its creators -- the what-the-hell-is-going-on overview. http://www.theobvious.com/about/ is an example.
- /help/ should always offer basic usage and navigation assistance. First-time visitors should feel comfortable with a site's interface after reading /help/.
- /archive/ should be an entryway into older data on a site, especially if that data is in the form of regular articles. If a site has a history, it should be accessible through /archive/.
- /new/ should offer the latest updates to a site, both what has changed and where it has changed, in the past day or week or month. Even if the front page is regularly updated, /new/ should offer the same -- or more detailed -- information.
- /search/ should offer the interface to a local search engine, if one is available.
- /contact/ is where detailed contact information should be kept. While /about/ will undoubtedly offer some of this data, /contact/ is explicitly for listing people and departments and how to get in touch with them, via e-mail or the post or the telephone.
- /download/ should be a gateway to any software available on a site, either links directly to the programs themselves or to a search engine or to the root of browsable tree.
- /legal/ should contain disclaimers, copyright notices and ownership information about a site and its contents.
- /text/ should be a low-bandwidth, no-graphics version of the site, if one is available.
- And finally, every site should provide a 404 page that lists each of these directories and their purpose. Misdirected users should be offered usable, helpful information: links to the site highlights, both the above standards and any local customizations.
However, there are several places where a small effort would have significant results:
The explosive growth of the Web has left us with millions of sites, billions of pages, and no logical system for finding anything on any of them. The Web has saved us from sterile sameness at the cost of almost utter chaos.
- The Apache Group, makers of the world's most popular Web server, could give a huge boost to this process by including the directories outlined above in their default configuration files. The Alias command would allow Webmasters to leave the current organization of their sites untouched, simply repointing the standard directory names to where their data actually lives.
- Netscape and Microsoft could do the same. Though less responsive than Apache, both companies have shown a slowly growing appreciation for grass-roots standards.
- But ultimately, the easiest way to further this standard is to encourage the public to come to expect it. While convincing opinion leaders of the standard's merits is certainly a first step, a more wide-ranging solution would be to include knowledge of the standard directories into the browsers themselves. Site navigation buttons could easily be added to any current browser, pre-configured to jump to /about/ or /help/ or any other directory. The first popular browser to make these buttons available would encourage millions of Webmasters to implement the standard, benefiting everyone on the Web, whether they use that browser or not.
It doesn't have to be that way. This rough standard provides a flexible middle ground, where creativity and uniqueness can peacefully co-exist with stability and coherence. This standard, or something like it, fills a gaping hole in the organization of the Web. It's needed. It's time.
Greg Knauss is an irregular contributor to Stating the Obvious. Sam Pratt is responsible for the original idea behind this article. Gregory Alkaitis-Carafelli, Alexis Massie, Steven Champeon and Rogers Cadenhead contributed their ideas and their time to its creation.
Other pieces about online publishing: