Installation

Almost all of the Encyclopedia consists of HTML documents, so installation on most systems should be trivial - just use your favorite browser to open index.htm in the top directory. However, several parts of the Encyclopedia require special attention.

CGI scripts

The interactive parts of the Encyclopedia use CGI scripts; you'll need to browse the Encyclopedia via a web server (I recommend Apache) to take advantage of them. All of the Encyclopedia's CGI scripts are written using Perl, so you need a working copy of Perl (version 5) from www.perl.org. The standard location for Perl is /usr/bin/perl. The web server must be configured to execute the CGI script when it's URL is requested. Newer versions of Apache should be configured by uncommenting the line in httpd.conf that reads:

         AddHandler cgi-script .cgi

If you have an older version of Apache that lacks this command, make sure a line such as the following appears in srm.conf:

         AddType application/x-httpd-cgi .cgi

You may also need to enable CGI execution on a per-directory basis in access.conf or httpd.conf, by setting the ExecCGI option. This can be done system-wide by adding the option to root Directory section, or by adding the option on a per-directory basis, like this (note that you must use the full UNIX path, not just the URL path):

         <Directory /home/httpd/html/CIE>
         Options Indexes ExecCGI
         </Directory>

I strongly suggest you run the webserver using an untrusted account, and regard the scripts with some suspicion. Although I try not to distribute anything that would create a security hole, I also can't take responsibility for security on other sites. If you're concerned about CGI security, or have difficulty understanding these directions, consult with a web server expert.

Search Engine

The search engine deserves special attention, since in addition to the CGI script, the Isearch program must be installed for it to work. In addition, the index files, which can be quite large, must be maintained and updated. Therefore, I have provided the option to allow other sites to relay their searches to www.freesoft.org; all other web pages, including the search results, are served from your site. I suggest you consider this option, since it is much simpler than maintaining your own search engine. Of course, you must have a network connection to www.freesoft.org for your searches to work.

On the other hand, you may decide not to depend on www.freesoft.org to run your searches, because of speed, intermittent network connectivity or other reasons. In this case, you'll have to setup your own search engine. Currently, this can only be done on a UNIX workstation with a web server.

You chose one configuration or the other by picking which version of search.htm you'll use.

Option A: Use www.freesoft.org.

Make search.htm a symbolic link to searchR.htm (R for remote). That's it. If your system doesn't support symbolic links (i.e, Windows), just make search.htm a copy of searchR.htm.

Option B: Setup your own search engine.

Make search.htm a symbolic link to searchL.htm (L for local).

Next, download and install Freesoft Isearch. Don't use the standard version of Isearch from www.etymon.com, since the freesoft version has been modified to support the Encyclopedia. You also need the md5sum program, which is now standard on most UNIX distributions.

Create the CIE/cgi-bin/results directory and make it writable by the webserver. This is where search results will be cached.

Check to see if the CIE/cgi-bin/index directory exists. If not, use rsync to fetch the index modules CIEindex and/or RFCindex. Don't be too concerned about getting a version of the indices that exactly matches the encyclopedia itself; the next step brings everything up to date. If you don't download the indices, reindex.pl will rebuild them from scratch, though it's probably faster to download them.

cd CIE/cgi-bin and run reindex.pl. It may take quite a while; especially if the index files are missing or out of date. Any error messages should be a cause for concern. Once the script has run successfully, the index files are up-to-date. The script should be run again if you update the encyclopedia or the RFCs.