Naming files and directories that contain Web documents
- A Web document's location within the directory structure is
a crucial aspect of managing information on the Web.
- Directories on the UNIX server have two important and distinct
- A directory is a device for grouping many files together --
we say that a bunch of files are "in directory x".
- Directories are also in other directories as "sub-directories"
so that the directory structure looks like a tree.
- The grouping aspect of directories affords
- grouping files with a common security profile
- limiting the total number of files in a directory
- grouping files by content, so all files in the directory are
about one subject
- grouping files by other "invisible" characteristics
such as information source or creation method
- Mapping directories on the Web server to other directories
or locations on the LAN, the mainframe is tricky business.
- We name many directories to end with "_data" or just "data"; this
kind of name means
- All the files in the directory "can be recreated"
or "are regularly recreated" and "have the same
- The files are not backed up by IRM
- The files not on shown on our automated "what's new"
- file naming conventions or rules
- It's nice to put a file named 'what4.txt' in a directory to
explain what that directory is intended to contain.
- Standard file extensions include:
- .html - HyperText Markup Language
- .pl - Perl scripts
- .cgi - usually Perl scripts that are executed by HTML forms
- .tmp - temporary files that can be re-created easily and thus
could be deleted with impunity
- .gif - Graphics Interchange File pictures
- .jpg - graphics files
- .sh - shell scripts - UNIX commands
- .index - files that contain pointers to other files; small
files that are meant to be read by a .cgi script; some are written
by a Perl program while others are constructed by hand.
- .bak - back-up files
- .awk - an Awk program
- It is ideal if every directory on the WWW server contains
a file named 'index.html' for the following reasons:
- The server will return 'index.html' if a user only provides
the directory in the URL
- IRM's Web site outliner will
automatically show the title from 'index.html' files in
the IRM Table of Contents page.