Help Information on Autoindexing
The Corporate "Auto-Directory" system allows the user to refresh the directory
of HTML files in one area with data indexed according to a few different
views. This can be done using a hot link on the index itself, or manually
from the Unix command line (as done by the administrator).
Make sure that the compiled makeindex wrapper in /ftp/web/bin/ is being
run. Do this by giving the path /ftp/web/bin/xxxx before /ftp/pub/bin in the
.cshrc/.tschrc .
There is a help file on
Message Posting available.
There is a help file on the various switches and options for the
makeindex command
available.
The details of the resulting directory in an area are controlled by a file
called help_on_index_format_file.html
that is found in the
same directory as this help file.
The goal will be to allow users to transfer HTML files from their own machine
to this area on the web server, then to hit the "Update" button to rebuild
the index. This implies publishing Web documents using only the browser
and FTP. This is the goal of many first-level publishing systems being
introduced on the web.
Compromises
The world is full of them and so is this system.
- There are only two fixed views of the data today (there are more
options available in a separate file in each directory - used the
"Choose a Sub-List" option).
- Update effectively does nothing if the date of the previously
generated list is newer than all files in the area. This implies lower
machine load if people hit the "Update" selection often, but it is
possible to mix up the system if files are moved without changing the
date (not a problem with FTP - only a problem with Telnet).
- The HTML data format requires a specific format for the header.
We will loosen this with time, but at present it requires a <title>
line, and generally prefers (we may insist in the future) on a <meta
line following the title (within HTML comments) that gives the full
long name of the file (using lower-case letters/numbers and the "_" character).
- The directory structure requires that areas have either subdirectories
and one index.html file, or else be the end nodes with just HTML files (plus
an img.d gif/jpg directory and maybe another for data in original MSWord or
PowerPoint format). This structure makes it difficult to add data in an
ad-hock fashion - fortunately.
What We Do Not Aim to Accomplish
Our web site is not a powerful database system that allows SQL or other complex
queries to be run on the data. The desire to use complex criteria to select
files is obvious, the implementation and support is not. Hence we stay to a
fairly simple web-based list of files in the area, plus various simple
search tools.
fetterley@houston.sns.slb.com