WWW Technical Guidelines

Contents:

• Introduction
• General Restrictions
  • Server Software
  • System Administration
• General Guidelines
  • Images
  • File Sizes
  • File Names
  • Directory Structure

I. Introduction

Besides planning the content of your World Wide Web pages, the way you organize that information in the file system of the serving computer system can make a big difference in your ability to maintain the information over time. In general you should try to have the file system "parallel" the organization of the information on your "index" pages; i.e., major categories should each have their corresponding information found in a subdirectory (subfolder) of the same (or as close as reasonable) name. This way related information will all be in a common directory and should it become necessary to move that information at a later date, only one link needs to be changed in the index page.

II. General Restrictions

A. Server Software

The specific server software used on the computer system you're using may place certain restrictions on how you organize certain categories of web-related information. For example, it may be necessary to place all cgi support programs in one directory (note: make sure that such a directory is not in any of your public web directories) or all context sensitive maps may have to be placed in one directory.

B. System Administration

As a guest on a shared computer facility, you are likely to encounter specific restrictions enforced by that facility's administrator. Such restrictions are usually placed on you in order to protect other parts of the system's operation and to make system administration manageable. Examples of such restrictions include specific file naming conventions for your index file(s), restrictions on the location of coordinate files for context-sensitive maps, and file naming conventions on form and map files.

Any CGI scripts that you wish to use should be reviewed by the system administrator before they are put into production. These scripts can pose system security risks, and this needs to be checked for beforehand.

III. General Guidelines

A. Images

Image files should be kept in the same directories as the associated html file that references that image. An exception occurs in the case of an image file that is used in several different html files that are in different directories (e.g., a frequently used or global image). In those cases the image should be placed in a common image directory at the top of the public html file area, so that they are readily available to all html files below the top (root) directory. (Placing commonly used global images in such a central location will also reduce the amount of work the browser has to do each time such an image is used.)

When selecting images to be included in your web material, "smaller is better," particularly for material on opening or home pages. If large images are desired, it is strongly recommended that a "thumbnail" version of the large image be the default image displayed. This thumbnail image can then reference the full image, so that the user can click on the thumbnail image to obtain a copy of the full image.

To help the viewers of your web pages, you should provide some information on the size of an image that will be transferred to the user's browser for links that are associated with image files. This way a user connected to the web via a slow modem has an idea how long a particular linked image will take to load.

B. File Sizes

Text files should be limited in size, since users frequently have to wait for all the information to be transferred to their (browsing software on their) computer before they can work with the information. And the transfer time can cause significant user delays, especially if the user is connected to the web over a slow modem. A suggested rule-of-thumb for the maximum size of a text html file is 20,000 characters (20k). Should a file become larger than that, it should be broken into two or more smaller files linked in the appropriate ways.

C. File Names

Users of DOS/Windows-based computer systems deal with filenames and directory names that are all in upper case letters. However in html information, the case used in link references to html file names is case sensitive. File and directory names should be carefully chosen so as to convey useful information yet not be so long as to cause an absolute reference to the material involved to be too long a string in the referencing web material.

D. Directory Structure

An example directory for a complete set of web materials is shown in the figure below. The HOME_DIR represents the root of a particular set of html material and may be a few levels below the root of the host system. Within that HOME_DIR the subdirectories should be the major supporting parts for this web material; e.g., the cgi-bin directory contains the CGI scripts used, the image_map directory has the coordinate files for context-sensitive maps, and the public_html directory contains the actual html material that will be viewed by users' web browsers.

The root of public html area (public_html in the example) should just contain your index (or home) html file, the directories for each major category in your index file, and any supporting html and image files. Each subdirectory of your public html directory should try to follow the same philosophy, i.e., an index or main menu file plus supporting files and subdirectories for each major category in that area. This way if a subdirectory has to be moved to another area of the computer file system (or even to a different computer system), very few links need to be changed in the html material that refers to the material in this subdirectory.

Two important related guidelines regarding the use of links within html documents:

1. Use relative addressing for links wherever possible. For example, A HREF="../mystuff.html" is much better than A HREF="http://www.mysystem.edu/general/particular/mystuff.html"
2. Structure your material so that links within a html document never go more than one level down in the file tree from the present location in the file tree.